To solve the real-time through-the-wall imaging problem in the presence of wall ambiguities, an approach based on the least-squares support vector machines (LS-SVMs) is proposed. This technique converts the through-the-wall problem into the establishment and use of a mapping between the backscattered data and the target properties. The wall parameters and the propagation effects caused by the walls are both included in the mapping and can be regressed after the LS-SVM training process. The target properties are estimated using LS-SVM. Noiseless and noisy measurements are performed to demonstrate that the approach can provide comparable performance in terms of robustness and efficacy, as well as improved performance in terms of accuracy and convenience, in comparison with the approach based on the support vector machine (SVM). The influence of the radius of the target on the estimation problem is discussed, and the estimated results show that both the LS-SVM and the SVM have good performances in terms of generalization.

A method for estimating the parameters of the sinusoidal frequency modulation (SFM) signal is presented in this paper. Based on the modified discrete sinusoid frequency modulation transform (DSFMT), the SFM signal can be transformed into the DSFMT domain where it is energy-concentrated and then the parameters can be estimated by the global maximum. To search for the location of the global maximum with less computational load, particle swarm optimization is used in this paper. Then the algorithm is used in the synthetic aperture radar imaging with high frequency vibration of the platform, and the focus performance can be improved significantly. Simulation results demonstrate the effectiveness of the method proposed in this paper.

Airborne light detection and ranging (LiDAR) system calibration is a crucial procedure for ensuring the accuracy of point data. A common practice is to use conjugate planar patches to recover systematic parameters based on coplanar constraints and to use planes with different orientations to decrease the correlations between the systematic errors. When there are not sufficient planar patches and the configuration of planar patches is not optimal, it is difficult to guarantee the reliability of the estimated system parameters. Based on the analyses of the bore-sight angle effects, we find that not only the orientations but also the distribution of planar patches play an important role in the calibration procedure. We propose an improved method for bore-sight calibration based on the principles of symmetry of coordinate offsets and low correlations between bore-sight angles. Comparisons of the experimental results of bore-sight angle calibration suggest that the proposed configuration of conjugate planar patches can decrease the correlations between bore-sight angles and improve the reliability of calibration results. The optical results obtained from four gable-roof buildings are very close to the results calculated by the RiProcess software with a deviation of about 0.001 deg.

Computer-based reconstruction models can be used to approximate urban environments. These models are usually based on several mathematical approximations and the usage of different sensors, which implies dependency on many variables. The sensitivity analysis presented in this paper is used to weigh the relative importance of each uncertainty contributor into the calibration of a panoramic camera–LiDAR system. Both sensors are used for three-dimensional urban reconstruction. Simulated and experimental tests were conducted. For the simulated tests we analyze and compare the calibration parameters using the Monte Carlo and Latin hypercube sampling techniques. Sensitivity analysis for each variable involved into the calibration was computed by the Sobol method, which is based on the analysis of the variance breakdown, and the Fourier amplitude sensitivity test method, which is based on Fourier’s analysis. Sensitivity analysis is an essential tool in simulation modeling and for performing error propagation assessments.

Carbon Observing Satellite (Tan-Sat) is the first satellite of China designed to monitor column-averaged atmospheric carbon dioxide (X_CO2) by detecting gas absorption spectra of the solar shortwave infrared radiation reflected from the Earth’s surface and atmosphere. Two instruments are accommodated on Tan-Sat: the high resolution hyperspectral sensor for carbon observation grating spectrometer (HRHS-GS) and the cloud and aerosol polarimetric imager (CAPI). HRHS-GS will provide the space-based measurements of CO₂ on a scale and with the accuracy and precision to quantify terrestrial sources and sinks of CO₂. CAPI is used to identify the contamination by optically thick clouds and to minimize the impact of scattering by aerosol. These two instruments work together to collect global column CO₂ concentrations with correction for cloud and aerosol contamination. The instrument design of HRHS-GS is presented. Ocean reflectivity and the incident radiation of the instrument for transverse electric and transverse magnetic polarizations in glint mode are discussed. The changes to glint mode operation are described. The spectral characteristics of HRHS-GS were determined through the laser-based spectral calibration. The onboard spectral calibration method based on spectrum matching is introduced. The availability was verified, satisfying the onboard spectral calibration accuracy requirement of better than Δλ/10 (Δλ is spectral resolution).

The moderate resolution imaging spectroradiometer (MODIS) has 20 reflective solar bands (RSB), covering a spectral range from 0.41 to 2.2  μm, which are calibrated on-orbit using its onboard calibrators, which include a solar diffuser, a solar diffuser stability monitor, and a spectroradiometric calibration assembly. A space view (SV) port is used to provide a background reference and also facilitates near-monthly lunar observations through a spacecraft roll. In every scan, the Earth’s surface, SV, and onboard calibrators are viewed via a two-sided scan mirror, the reflectance of which depends on the angle of incidence (AOI) as well as the wavelength of the incident light. Response-versus-scan-angle (RVS) is defined as a dependence function of the scan mirror’s reflectance over AOI. An initial RVS for each RSB was measured prelaunch for both Terra and Aqua MODIS. Algorithms have been developed to track the on-orbit RVS variation using the measurements from the onboard calibrators, supplemented with the earth view (EV) trends from pseudoinvariant desert targets obtained at different AOI. Since the mission beginning, the MODIS characterization support team (MCST) has dedicated efforts in evaluating approaches of characterizing the on-orbit RVS. A majority of the approaches focused on fitting the data at each AOI over time and then deriving the relative change at different AOI. The current version of the on-orbit RVS algorithm, as implemented in the collection 6 (C6) level-1B (L1B), is also based on the above rationale. It utilizes the EV response trends from the pseudoinvariant Libyan desert targets to supplement the gain derived from the onboard calibrators. The primary limitation of this approach is the assumption of the temporal stability of these desert sites. Consequently, MCST developed an approach that derives the on-orbit RVS change using measurements from a single desert site, combined with the on-orbit lunar measurements. In addition, the EV and onboard responses are fit first as a function of AOI before fitting temporally in order to eliminate the dependence on the stability of the desert site. Comprehensive comparisons are performed with current C6 RVS results for both Terra and Aqua MODIS. Results demonstrate that this alternative method provides a supplemental means to monitor the on-orbit RVS for MODIS RSB.

This paper integrates decision tree–based data mining (DM) and object-based image analysis (OBIA) to provide a transferable model for the detailed characterization of urban land-cover classes using WorldView-2 (WV-2) satellite images. Many articles have been published on OBIA in recent years based on DM for different applications. However, less attention has been paid to the generation of a transferable model for characterizing detailed urban land cover features. Three subsets of WV-2 images were used in this paper to generate transferable OBIA rule-sets. Many features were explored by using a DM algorithm, which created the classification rules as a decision tree (DT) structure from the first study area. The developed DT algorithm was applied to object-based classifications in the first study area. After this process, we validated the capability and transferability of the classification rules into second and third subsets. Detailed ground truth samples were collected to assess the classification results. The first, second, and third study areas achieved 88%, 85%, and 85% overall accuracies, respectively. Results from the investigation indicate that DM was an efficient method to provide the optimal and transferable classification rules for OBIA, which accelerates the rule-sets creation stage in the OBIA classification domain.

Semisupervised feature selection methods can improve classification performance and enhance model comprehensibility with few labeled objects. However, most of the existing methods require graph construction beforehand, and the resulting heavy computational cost may bring about the failure to accurately capture the local geometry of data. To overcome the problem, adaptive semisupervised feature selection (ASFS) is proposed. In ASFS, the goodness of each feature is measured by linear objective functions based on loss functions and probability distribution matrices. By alternatively optimizing model parameters and automatically adjusting the probabilities of boundary objects, ASFS can measure the genuine characteristics of the data and then rank and select features. The experimental results attest to the effectiveness and practicality of the method in comparison with the latest and state-of-the-art methods on a Worldview II image and a Quickbird II image.

An estimation of signal parameters via a rotational invariance techniques–like (ESPRIT-like) algorithm is proposed to estimate the direction of arrival and direction of departure for bistatic multiple-input multiple-output (MIMO) radar. The properties of a noncircular signal and Euler’s formula are first exploited to establish a real-valued bistatic MIMO radar array data, which is composed of sine and cosine data. Then the receiving/transmitting selective matrices are constructed to obtain the receiving/transmitting rotational invariance factors. Since the rotational invariance factor is a cosine function, symmetrical mirror angle ambiguity may occur. Finally, a maximum likelihood function is used to avoid the estimation ambiguities. Compared with the existing ESPRIT, the proposed algorithm can save about 75% of computational load owing to the real-valued ESPRIT algorithm. Simulation results confirm the effectiveness of the ESPRIT-like algorithm.

Bilateral filter (BF) theory is applied to integrate spatial contextual information into the spectral domain for improving the accuracy of the support vector machine (SVM) classifier. The proposed classification framework is a two-stage process. First, an edge-preserved smoothing is carried out on a hyperspectral image (HSI). Then, the SVM multiclass classifier is applied on the smoothed HSI. One of the advantages of the BF-based implementation is that it considers the spatial as well as spectral closeness for smoothing the HSI. Therefore, the proposed method provides better smoothing in the homogeneous region and preserves the image details, which in turn improves the separability between the classes. The performance of the proposed method is tested using benchmark HSIs obtained from the airborne-visible-infrared-imaging-spectrometer (AVIRIS) and the reflective-optics-system-imaging-spectrometer (ROSIS) sensors. Experimental results demonstrate the effectiveness of the edge-preserved filtering in the classification of the HSI. Average accuracies (with 10% training samples) of the proposed classification framework are 99.04%, 98.11%, and 96.42% for AVIRIS–Salinas, ROSIS–Pavia University, and AVIRIS–Indian Pines images, respectively. Since the proposed method follows a combination of BF and the SVM formulations, it will be quite simple and practical to implement in real applications.

An extended nonlinear chirp scaling (NLCS) algorithm is proposed to process data of highly squinted, high-resolution, missile-borne synthetic aperture radar (SAR) diving with a constant acceleration. Due to the complex diving movement, the traditional signal model and focusing algorithm are no longer suited for missile-borne SAR signal processing. Therefore, an accurate range equation is presented, named as the equivalent hyperbolic range model (EHRM), which is more accurate and concise compared with the conventional fourth-order polynomial range equation. Based on the EHRM, a two-dimensional point target reference spectrum is derived, and an extended NLCS algorithm for missile-borne SAR image formation is developed. In the algorithm, a linear range walk correction is used to significantly remove the range-azimuth cross coupling, and an azimuth NLCS processing is adopted to solve the azimuth space variant focusing problem. Moreover, the operations of the proposed algorithm are carried out without any interpolation, thus having small computational loads. Finally, the simulation results and real-data processing results validate the proposed focusing algorithm.

The increase of the spatial resolution of remote-sensing sensors helps to capture the abundant details related to the semantics of surface objects. However, it is difficult for the popular object-oriented classification approaches to acquire higher level semantics from the high spatial resolution remote-sensing (HSR-RS) images, which is often referred to as the “semantic gap.” Instead of designing sophisticated operators, convolutional neural networks (CNNs), a typical deep learning method, can automatically discover intrinsic feature descriptors from a large number of input images to bridge the semantic gap. Due to the small data volume of the available HSR-RS scene datasets, which is far away from that of the natural scene datasets, there have been few reports of CNN approaches for HSR-RS image scene classifications. We propose a practical CNN architecture for HSR-RS scene classification, named the large patch convolutional neural network (LPCNN). The large patch sampling is used to generate hundreds of possible scene patches for the feature learning, and a global average pooling layer is used to replace the fully connected network as the classifier, which can greatly reduce the total parameters. The experiments confirm that the proposed LPCNN can learn effective local features to form an effective representation for different land-use scenes, and can achieve a performance that is comparable to the state-of-the-art on public HSR-RS scene datasets.

Remote sensing technology has applications in various knowledge domains, such as agriculture, meteorology, land use, environmental monitoring, military surveillance, and mineral exploration. The increasing advances in image acquisition techniques have allowed the generation of large volumes of data at high spectral resolution with several spectral bands representing images collected simultaneously. We propose and evaluate a supervised classification method composed of three stages. Initially, hyperspectral values and entropy information are employed by support vector machines to produce an initial classification. Then, the K-nearest neighbor technique searches for pixels with high probability of being correctly classified. Finally, minimum spanning forests are applied to these pixels to reclassify the image taking spatial restrictions into consideration. Experiments on several hyperspectral images are conducted to show the effectiveness of the proposed method.

Markov random field (MRF) model is an effective tool for polarimetric synthetic aperture radar (PolSAR) image classification. However, due to the lack of suitable contextual information in conventional MRF methods, there is usually a contradiction between edge preservation and region homogeneity in the classification result. To preserve edge details and obtain homogeneous regions simultaneously, an adaptive MRF framework is proposed based on a polarimetric sketch map. The polarimetric sketch map can provide the edge positions and edge directions in detail, which can guide the selection of neighborhood structures. Specifically, the polarimetric sketch map is extracted to partition a PolSAR image into structural and nonstructural parts, and then adaptive neighborhoods are learned for two parts. For structural areas, geometric weighted neighborhood structures are constructed to preserve image details. For nonstructural areas, the maximum homogeneous regions are obtained to improve the region homogeneity. Experiments are taken on both the simulated and real PolSAR data, and the experimental results illustrate that the proposed method can obtain better performance on both region homogeneity and edge preservation than the state-of-the-art methods.

As a complementary imaging technology, coincidence imaging radar (CIR) achieves high resolution for stationary or low-speed targets under the assumption of ignoring the influence of the original position mismatching. As to high-speed moving targets moving from the original imaging cell to other imaging cells during imaging, it is inaccurate to reconstruct the target using the previous imaging plane. We focus on the recovery problem for high-speed moving targets in the CIR system based on the intrapulse frequency random modulation signal in a single pulse. The effects induced by the motion on the imaging performance are analyzed. Because the basis matrix in the CIR imaging equation is determined by the unknown velocity parameter of the moving target, both the target images and basis matrix should be estimated jointly. We propose an adaptive joint parametric estimation recovery algorithm based on the Tikhonov regularization method to update the target velocity and basis matrix adaptively and recover the target images synchronously. Finally, the target velocity and target images are obtained in an iterative manner. Simulation results are presented to demonstrate the efficiency of the proposed algorithm.

In order to achieve high-precision three-dimensional (3-D) imaging with an airborne downward-looking linear-array 3-D synthetic aperture radar (LA-3D-SAR), a uniform virtual antenna array can be obtained by aperture synthesis of the cross-track sparse multiple-input-multiple-output array. However, the actual 3-D imaging quality is unavoidably degraded by array errors such as the multichannel amplitude-phase errors due to the nonideal antenna characteristics, and the virtual element position errors due to vibrations and motion measurement deviations. We investigate the effects of these errors on the forms and the degrees of image quality degradation and consider the use of corresponding calibration methods to eliminate the effects of errors. For the multichannel amplitude-phase errors, the target response is subject to an integrated sidelobe level increase introduced by the phase error, which can be calibrated based on external (parallel or point target) calibrators, as proposed in the paper. For the virtual element position errors, they mainly the result of contrast degradation and noise in the image along the cross-track direction and have little impact on the range and along-track directions. The imaging performance is more sensitive to the error component in the height direction as compared to other components, the precision requirement of which should be established as the calibration reference. A calibration method based on time-divided active calibrators is proposed to estimate and correct the virtual element position errors. Both numerical simulations and real data experiments have shown the validity of the analyses as well as the effectiveness of the proposed calibration methods.

As remote sensing image applications are often characterized with limited bandwidth and high-quality demands, higher coding performance of remote sensing images are desirable. The embedded block coding with optimal truncation (EBCOT) is the fundamental part of JPEG2000 image compression standard. However, EBCOT only considers correlation within a sub-band and utilizes a context template of eight spatially neighboring coefficients in prediction. The existing optimization methods in literature using the current context template prove little performance improvements. To address this problem, this paper presents a new mutual information (MI)-based context template selection and modeling method. By further considering the correlation across the sub-bands, the potential prediction coefficients, including neighbors, far neighbors, parent and parent neighbors, are comprehensively examined and selected in such a manner that achieves a nice trade-off between the MI-based correlation criterion and the prediction complexity. Based on the selected context template, a high-order prediction model, which jointly considers the weight and the significance state of each coefficient, is proposed. Experimental results show that the proposed algorithm consistently outperforms the benchmark JPEG2000 standard and state-of-the-art algorithms in term of coding efficiency at a competitive computational cost, which makes it desirable in real-time compression applications, especially for remote sensing images.

Accurate determination of the water depth is important for marine spatial planning, producing maritime charts for navigation, seabed morphology studies, and carrying out different activities in the coastal waters. Bathymetric data are lacking foremost in the shallow water regions as those areas are often inaccessible to the hydrographic ships carrying out echo sounding measurements. Remote sensing technology can be used as an alternative for shallow water bathymetry mapping. Varieties of empirical methods have been proposed for bathymetry retrieval, where the relationship between remotely sensed radiance of the water body and the water depth at sampled locations was established empirically. Two most widely used depth derivation methods, the linear band model proposed by Lyzenga (1978, 1985, 2006), and the log-transformed band ratio model proposed by Stumpf et al. (2003), were applied to the different preprocessing level airborne Hyspex hyperspectral images from the optically complex Baltic Sea area and evaluated for accuracy. Results showed that the Lyzenga linear band model outperformed the Stumpf log-transformed band ratio model. The best results were achieved with the atmospherically corrected images. The application of glint correction did not improve, but even reduced the accuracy of bathymetric maps.

To improve the effectiveness of thermal sharpening in mountainous regions, paying more attention to the laws of land surface energy balance, a thermal sharpening method based on the geographical statistical model (GSM) is proposed. Explanatory variables were selected from the processes of land surface energy budget and thermal infrared electromagnetic radiation transmission, then high spatial resolution (57 m) raster layers were generated for these variables through spatially simulating or using other raster data as proxies. Based on this, the local adaptation statistical relationship between brightness temperature (BT) and the explanatory variables, i.e., the GSM, was built at 1026-m resolution using the method of multivariate adaptive regression splines. Finally, the GSM was applied to the high-resolution (57-m) explanatory variables; thus, the high-resolution (57-m) BT image was obtained. This method produced a sharpening result with low error and good visual effect. The method can avoid the blind choice of explanatory variables and remove the dependence on synchronous imagery at visible and near-infrared bands. The influences of the explanatory variable combination, sampling method, and the residual error correction on sharpening results were analyzed deliberately, and their influence mechanisms are reported herein.

Coherent point drift (CPD) method is a powerful registration tool under the framework of the Gaussian mixture model (GMM). However, the global spatial structure of point sets is considered only without other forms of additional attribute information. The equivalent simplification of mixing parameters and the manual setting of the weight parameter in GMM make the CPD method less robust to outlier and have less flexibility. An adaptive CPD method is proposed to automatically determine the mixing parameters by embedding the local attribute information of features into the construction of GMM. In addition, the weight parameter is treated as an unknown parameter and automatically determined in the expectation-maximization algorithm. In image registration applications, the block-divided salient image disk extraction method is designed to detect sparse salient image features and local self-similarity is used as attribute information to describe the local neighborhood structure of each feature. The experimental results on optical images and remote sensing images show that the proposed method can significantly improve the matching performance.

The incidence angle is a crucial parameter for synthetic aperture radar (SAR). The local incidence angle differs greatly in mountainous areas, and it may lead to quite different backscatter properties for the same ground targets for different incidence angles. As a result, it has to be taken into consideration in image classification. This work demonstrates the importance of the local incidence angle and the necessity of considering this angle in SAR image classification in mountainous areas. A local incidence angle referenced method based on support vector machines is developed to perform classification. In this method, the datasets are divided into different zones according to local incidence angle, and the training and predicting are performed within these zones to eliminate the influence of the local incidence angles. The experiments are performed around the Dongkemadi Glacier in the central Qinghai-Tibetan plateau using two RADARSAT-2 polarimetric SAR images. This paper concentrates on the effect of the local incidence angles in SAR image classification. Compared to the method using only backscatter coefficients, the proposed method improves the overall classification accuracy by 6 to 8%. More important, it is verified that this method is more helpful for the ground cover types where the terrain changes sharply.

Remote sensing image denoising faces many challenges since a remote sensing image usually covers a wide area and thus contains complex contents. Using the patch-based statistical characteristics is a flexible method to improve the denoising performance. There are usually two kinds of statistical characteristics available: interior and exterior characteristics. Different statistical characteristics have their own strengths to restore specific image contents. Combining different statistical characteristics to use their strengths together may have the potential to improve denoising results. This work proposes a method combining statistical characteristics to adaptively select statistical characteristics for different image contents. The proposed approach is implemented through a new characteristics selection criterion learned over training data. Moreover, with the proposed combination method, this work develops a denoising algorithm for remote sensing images. Experimental results show that our method can make full use of the advantages of interior and exterior characteristics for different image contents and thus improve the denoising performance.

A computational method for suppressing clutter and generating clear microwave images of targets is proposed in this paper, which combines synthetic aperture radar (SAR) principles with recursive method and waveform design theory, and it is suitable for SAR for special applications. The nonlinear recursive model is introduced into the SAR operation principle, and the cubature Kalman filter algorithm is used to estimate target and clutter responses in each azimuth position based on their previous states, which are both assumed to be Gaussian distributions. NP criteria-based optimal waveforms are designed repeatedly as the sensor flies along its azimuth path and are used as the transmitting signals. A clutter suppression filter is then designed and added to suppress the clutter response while maintaining most of the target response. Thus, with fewer disturbances from the clutter response, we can generate the SAR image with traditional azimuth matched filters. Our simulations show that the clutter suppression filter significantly reduces the clutter response, and our algorithm greatly improves the SINR of the SAR image based on different clutter suppression filter parameters. As such, this algorithm may be preferable for special target imaging when prior information on the target is available.

A remote sensing image (RSI) fusion method based on multiscale morphological component analysis (m-MCA) is presented. Our contribution describes a new multiscale sparse image decomposition algorithm called m-MCA, which we apply to RSI fusion. Building on MCA, m-MCA combines curvelet transform bases and local discrete cosine transform bases to build a multiscale decomposition dictionary, and controls the entries of the dictionary to decompose the image into texture components and cartoon components with different scales. The effective scale texture component of high-resolution RSI and the cartoon component of multispectral RSI are selected to reconstruct the fusion image. Compared with state-of-the-art fusion methods, the proposed fusion method obtains higher spatial resolution and lower spectral distortion with reduced computation load in numerical experiments.

Vibrating targets generally induce sinusoidal micro-Doppler modulation in high resolution synthetic aperture radar (SAR). They could cause defocused and ghost results by conventional imaging algorithms. This paper proposes a method on vibrating target imaging in frequency-modulation continuous-wave (FMCW) SAR systems. The continuous motion of sensor platform during pulse time is considered in the signal model. Based on Bessel series expansion of the signal in the azimuth direction, the influence of platform motion on the azimuth frequency is eliminated after dechirp and deskew. In addition, the range walk is compensated in the two-dimensional frequency domain by Doppler keystone transform. Next, using range cell migration correction, the azimuth quadratic phase compensation and the range curvature correction are made in range-Doppler domain for the focus of paired echoes. The residual video phase of paired echoes is eliminated, and vibration parameters are estimated to compensate in the sinusoidal modulation phase. Then the deghosted image of vibrating targets can be obtained. The proposed method is applicable to multiple targets with various vibrating states due to no need of a priori knowledge of targets. Finally, simulations are carried out to validate the effectiveness of the method in FMCW-SAR imaging of vibrating targets.

Unmanned aerial vehicles (UAVs) have become used increasingly in earth surface observations, with a special interest put into automatic modes of environmental control and recognition of artificial objects. Fractal methods for image processing well detect the artificial objects in digital space images but were not applied previously to the UAV-produced imagery. Parameters of photography, on-board equipment, and image characteristics differ considerably for spacecrafts and UAVs. Therefore, methods that work properly with space images can produce different results for the UAVs. In this regard, testing the applicability of fractal methods for the UAV-produced images and determining the optimal range of parameters for these methods represent great interest. This research is dedicated to the solution of this problem. Specific features of the earth’s surface images produced with UAVs are described in the context of their interpretation and recognition. Fractal image processing methods for extracting artificial objects are described. The results of applying these methods to the UAV images are presented.

Status observations of roofing material degradation are constantly evolving due to urban feature heterogeneities. Although advanced classification techniques have been introduced to improve within-class impervious surface classifications, these techniques involve complex processing and high computation times. This study integrates field spectroscopy and satellite multispectral remote sensing data to generate degradation status maps of concrete and metal roofing materials. Field spectroscopy data were used as bases for selecting suitable bands for spectral index development because of the limited number of multispectral bands. Mapping methods for roof degradation status were established for metal and concrete roofing materials by developing the normalized difference concrete condition index (NDCCI) and the normalized difference metal condition index (NDMCI). Results indicate that the accuracies achieved using the spectral indices are higher than those obtained using supervised pixel-based classification. The NDCCI generated an accuracy of 84.44%, whereas the support vector machine (SVM) approach yielded an accuracy of 73.06%. The NDMCI obtained an accuracy of 94.17% compared with 62.5% for the SVM approach. These findings support the suitability of the developed spectral index methods for determining roof degradation statuses from satellite observations in heterogeneous urban environments.

Aiming to precisely estimate the velocity of high-speed targets for step frequency (SF) radar, a positive–positive–negative SF waveform consisting of two continuous positive SF pulse trains and a negative one is designed, and a velocity estimation method is proposed based on two-dimensional time-domain cross correlation (2-D TDCC). Making full use of the characteristics of the designed waveform, the coarse velocity estimation is achieved by 2-D TDCC of positive–positive SF pulse trains and then the Radon transform is applied to solve velocity ambiguity for high-speed targets. After velocity compensation for positive–negative SF pulse trains, the velocity residual is estimated precisely by 2-D TDCC. Simulation results show that the proposed method exhibits good performance for estimation accuracy, stability performance, computational complexity, and data rate by comparisons.

In CIELab color space, we propose a remote sensing image fusion technique based on nonsubsampled shearlet transform (NSST) and pulse coupled neural network (PCNN), which aim to improve the efficiency and performance of the remote sensing image fusion by combining the excellent properties of the two methods. First, panchromatic (PAN) and multispectral (MS) are transformed into CIELab color space to get different color components. Second, PAN and L component of MS are decomposed by the NSST to obtain corresponding the low-frequency coefficients and high-frequency coefficients. Third, the low-frequency coefficients are fused by intersecting cortical model (ICM); the high-frequency coefficients are divided into several sub-blocks to calculate the average gradient (AG), and the linking strength β of PCNN model is determined by the AG, so that the parameters β can be adaptively set according to the quality of the sub-block images, then the sub-blocks image are input into PCNN to get the oscillation frequency graph (OFG), the method can get the fused high-frequency coefficients according to the OFG. Finally, the fused L component is obtained by inverse NSST, and the fused RGB color image is obtained through inverse CIELab transform. The experimental results demonstrate that the proposed method provide better effect compared with other common methods.

With the gradual increase in the spatial and spectral resolution of hyperspectral images, the size of image data becomes larger and larger, and the complexity of processing algorithms is growing, which poses a big challenge to efficient massive hyperspectral image processing. Cloud computing technologies distribute computing tasks to a large number of computing resources for handling large data sets without the limitation of memory and computing resource of a single machine. This paper proposes a parallel pixel purity index (PPI) algorithm for unmixing massive hyperspectral images based on a MapReduce programming model for the first time in the literature. According to the characteristics of hyperspectral images, we describe the design principle of the algorithm, illustrate the main cloud unmixing processes of PPI, and analyze the time complexity of serial and parallel algorithms. Experimental results demonstrate that the parallel implementation of the PPI algorithm on the cloud can effectively process big hyperspectral data and accelerate the algorithm.

This paper presents the development of index to detect haze from moderate resolution imaging spectroradiometer remote sensing data. Detection of haze over a large area has always been a problem. This study focuses on Beijing, Tianjin, and Shijiazhuang cities in China. These cities have suffered the worst hazy weather in recent years. The spectral influence of haze on surface features was determined through analysis of the spectral variations of surface covers between hazy and haze-free days. A spectral index known as modified normalized difference haze index (m-NDHI) is developed that can be used to monitor haze distribution and intensity. Correlation analysis of the derived m-NDHI and previously developed NDHI with in situPM_2.5 (particulate matter with diameter <2.5  μm) data reveals that m-NDHI over water bodies has a coefficient of 0.7096, 0.5864, and 0.4857 and NDHI has coefficient of 0.5625, 0.5321, and 0.4618 with PM_2.5 for Beijing, Tianjin, and Shijiazhuang, respectively, in winter. Moreover, the correlation of m-NDHI with PM_2.5 is 0.4097, 0.8092, and 0.5546 during the spring, summer, and autumn, respectively, in Beijing. This developed index can be a much easier and more effective method to detect haze in large scales from remotely sensing data and characterize the situation of urban atmospheric pollution.

Hyperspectral imagery (HSI) has high spectral and spatial resolutions, which are essential for anomaly detection (AD). Many anomaly detectors assume that the spectrum signature of HSI pixels can be modeled with a Gaussian distribution, which is actually not accurate and often leads to many false alarms. Therefore, a sparsity model without any distribution hypothesis is usually employed. Dimensionality reduction (DR) as a preprocessing step for HSI is important. Principal component analysis as a conventional DR method is a linear projection and cannot exploit the nonlinear properties in hyperspectral data, whereas locally linear embedding (LLE) as a local, nonlinear manifold learning algorithm works well for DR of HSI. A modified algorithm of sparsity divergence index based on locally linear embedding (SDI-LLE) is thus proposed. First, kernel collaborative representation detection is adopted to calculate the sparse dictionary matrix of local reconstruction weights in LLE. Then, SDI is obtained both in the spectral and spatial domains, where spatial SDI is computed after DR by LLE. Finally, joint SDI, combining spectral SDI and spatial SDI, is computed, and the optimal SDI is performed for AD. Experimental results demonstrate that the proposed algorithm significantly improves the performance, when compared with its counterparts.

It is well known that rapid building damage assessment is necessary for postdisaster emergency relief and recovery. Based on an analysis of very high-resolution remote-sensing images, we propose an automatic building damage assessment framework for rainfall- or earthquake-induced landslide disasters. The framework consists of two parts that implement landslide detection and the damage classification of buildings, respectively. In this framework, an approach based on modified object-based sparse representation classification and morphological processing is used for automatic landslide detection. Moreover, we propose a building damage classification model, which is a classification strategy designed for affected buildings based on the spectral characteristics of the landslide disaster and the morphological characteristics of building damage. The effectiveness of the proposed framework was verified by applying it to remote-sensing images from Wenchuan County, China, in 2008, in the aftermath of an earthquake. It can be useful for decision makers, disaster management agencies, and scientific research organizations.

Accurate mapping of invasive species in a cost-effective way is the first step toward understanding and predicting the impact of their invasions. However, it is challenging in coastal wetlands due to confounding effects of biodiversity and tidal effects on spectral reflectance. The aim of this work is to describe a method to improve the accuracy of mapping an invasive plant (Spartina alterniflora), which is based on integration of pan-sharpening and classifier ensemble techniques. A framework was designed to achieve this goal. Five candidate image fusion algorithms, including principal component analysis fusion algorithm, modified intensity-hue-saturation fusion algorithm, wavelet-transform fusion algorithm, Ehlers fusion algorithm, and Gram–Schmidt fusion algorithm, were applied to pan-sharpening Landsat 8 operational land imager (OLI) imagery. We assessed the five fusion algorithms with respect to spectral and spatial fidelity using visual inspection and quantitative quality indicators. The optimal fused image was selected for subsequent analysis. Then, three classifiers, namely, maximum likelihood, artificial neural network, and support vector machine, were employed to preclassify the fused and raw OLI 30-m band images. Final object-based S. alterniflora maps were generated through classifier ensemble analysis of outcomes from the three classifiers. The results showed that the introduced method obtained high classification accuracy, with an overall accuracy of 90.96% and balanced misclassification errors between S. alterniflora and its coexistent species. We recommend future research to adopt the proposed method for monitoring long-term or multiseasonal changes in land coverage of invasive wetland plants.

This work aims to verify the applicability of models obtained using interferometric synthetic aperture radar (SAR) data for estimation of biophysical Eucalyptus saligna parameters [diameter of breast height (DBH), total height and volume], as a method of continuous forest inventory. In order to obtain different digital elevation models, and the interferometric height (Hint) to retrieve the tree heights, SAR surveying was carried out by an airborne interferometric SAR in two frequencies X and P bands. The study area, located in the Brazilian southeast region (S 22°53′22″/W 45°26′16″ and S 22°53′22″/W 45°26′16″), comprises 128.64 hectares of Eucalyptus saligna stands. The methodological procedures encompassed: forest inventory, topographic surveying, radar mapping, radar processing, and multivariable regression techniques to build Eucalyptus volume, DBH, and height models. The statistical regression pointed out Hint and interferometric coherence as the most important variables for the total height and DBH estimation; for the volume model, however, only the Hint variable was selected. The performance of the biophysical models from the second campaign, two years later (2006), were consistent and its results are very promising for updating annual inventories needed for managing Eucalyptus plantations.

The Guanabara Bay (GB) is an estuarine system in the metropolitan region of Rio de Janeiro (Brazil), with a surface area of ∼346  km² threatened by anthropogenic pressure. Remote sensing can provide frequent data for studies and monitoring of water quality parameters, such as chlorophyll-a concentration (Chl-a). Different combination of Medium Resolution Imaging Spectrometer (MERIS) remote sensing reflectance band ratios were used to estimate Chl-a. Standard algorithms such as Ocean Color 3-band, Ocean Color-4 band, fluorescence line height, and maximum chlorophyll index were also tested. The MERIS Chl-a estimates were statistically compared with a dataset of in situ Chl-a (2002 to 2012). Good correlations were obtained with the use of green, red, and near-infrared bands. The best performing algorithm was based on the red (665 nm) and green (560 nm) band ratio, named “RG3” algorithm (r²=0.71, chl-a=62,565 * x^1.6118). The RG3 was applied to a time series of MERIS images (2003- to 2012). The GB has a high temporal and spatial variability of Chl-a, with highest values found in the wet season (October to March) and in some of the most internal regions of the estuary. Lowest concentrations are found in the central circulation channel due to the flushing of ocean water masses promoted by pumping tide.

Until recently, Landsat technology has suffered from low signal-to-noise ratio (SNR) and comparatively poor radiometric resolution, which resulted in limited application for inland water and land use/cover mapping. The new generation of Landsat, the Landsat Data Continuity Mission carrying the Operational Land Imager (OLI), has improved SNR and high radiometric resolution. This study evaluated the utility of orthoimagery from OLI in comparison with the Advanced Land Imager (ALI) and hyperspectral Hyperion (after preprocessing) with respect to spectral profiling of classes, land use/cover classification, classification accuracy assessment, classifier selection, study area selection, and other applications. For each data source, the support vector machine (SVM) model outperformed the spectral angle mapper (SAM) classifier in terms of class discrimination accuracy (i.e., water, built-up area, mixed forest, shrub, and bare soil). Using the SVM classifier, Hyperion hyperspectral orthoimagery achieved higher overall accuracy than OLI and ALI. However, OLI outperformed both hyperspectral Hyperion and multispectral ALI using the SAM classifier, and with the SVM classifier outperformed ALI in terms of overall accuracy and individual classes. The results show that the new generation of Landsat achieved higher accuracies in mapping compared with the previous Landsat multispectral satellite series.

In remote sensing fusion, the spatial details of a panchromatic (PAN) image and the spectrum information of multispectral (MS) images will be transferred into fused images according to the characteristics of the human visual system. Thus, a remote sensing image fusion quality assessment called feature-based fourth-order correlation coefficient (FFOCC) is proposed. FFOCC is based on the feature-based coefficient concept. Spatial features related to spatial details of the PAN image and spectral features related to the spectrum information of MS images are first extracted from the fused image. Then, the fourth-order correlation coefficient between the spatial and spectral features is calculated and treated as the assessment result. FFOCC was then compared with existing widely used indices, such as Erreur Relative Globale Adimensionnelle de Synthese, and quality assessed with no reference. Results of the fusion and distortion experiments indicate that the FFOCC is consistent with subjective evaluation. FFOCC significantly outperforms the other indices in evaluating fusion images that are produced by different fusion methods and that are distorted in spatial and spectral features by blurring, adding noise, and changing intensity. All the findings indicate that the proposed method is an objective and effective quality assessment for remote sensing image fusion.

Automatic mapping of tree crown size (radius, diameter, or width) from remote sensing can provide a major benefit for practical and scientific purposes, but requires the development of accurate methods. This study presents an improved method for average tree crown diameter estimation at a forest plot level from high-resolution airborne data. The improved method consists of the combination of a window binarization procedure and a granulometric algorithm, and avoids the complicated crown delineation procedure that is currently used to estimate crown size. The systematic error in average crown diameter estimates is corrected with the improved method. The improved method is tested with coniferous, beech, and mixed-species forest plots based on airborne images of various spatial resolutions. The absolute (quantitative) accuracy of the improved crown diameter estimates is comparable or higher for both monospecies plots and mixed-species plots than the current methods. The ability of the improved method to produce good estimates for average crown diameters for monoculture and mixed species, to use remote sensing data of various spatial resolution and to operate in automatic mode promisingly suggests its applicability to a wide range of forest systems.

During the last decades, the Amazon rainforest underwent uncontrolled exploitation that modified its environmental variables. The current paper analyzes the spatiotemporal dynamics of the normalized difference vegetation index (NDVI), leaf area index (LAI), and surface albedo, and temperature in two different vegetation covers, preserved and deforested areas. We calculated the remote-sensing products using Landsat 5 TM images obtained during the dry season 1984, 1991, 2000, and 2011 of the central region of the State of Rondônia, Brazil. The results showed a reduction of vegetation indexes NDVI (∼0.70 in 1984 to ∼0.27 in 2011) and LAI (∼1.8 in 1984 to ∼0.3 in 2011), with an increase of surface albedo (0.12 in 1984 to 0.20 in 2011) and temperature (∼24°C in 1984 to 30°C in 2011) as the effect of the rainforest converted in grassland during the study period. No changes in any variables were observed in the protected area. Forest conversion into grassland resulted in a decrease of 69% in NDVI and 110% in LAI and a rise of 59% and 24% in albedo and surface temperature, respectively.

This article presents an analysis of the scattering measurements for an entire wheat growth cycle by ground-based scatterometers at a frequency of 5.3 GHz. Since wheat ears are related to wheat growth and yield, the radar backscatter of wheat was analyzed at two different periods, i.e., with and without wheat ears. Simultaneously, parameters such as wheat and soil characteristics as well as volume scattering and soil scattering were analyzed for the two periods during the entire growth cycle. Wheat ears have been demonstrated to have a great influence on radar backscatter; therefore, a modified version of water-cloud model used for retrieving biomass should consider the effect of wheat ears. This work presents two retrieval models based on the water-cloud model and adopts the advanced integral equation model to simulate the soil backscatter before the heading stage and the backscatter from the layer under wheat ears after the heading stage. The research results showed that the biomass retrieved from the advanced synthetic aperture radar (ASAR) images to agree well with the data measured in situ after setting the modified water-cloud model for the growth stages with ears. Furthermore, it was concluded that wheat ears should form an essential component of theoretical modeling as they influence the final yield.

We established a model for an airborne wind lidar. Numerical optimization algorithms should be used to solve this nonlinear model. We designed a Levenberg–Marquardt (L–M) algorithm and tested it with the modeled data. The retrieved velocity and the true velocity agree very well, and the adjusted R² is 0.99947. We have carried out an airborne coherent wind lidar experiment in January 2015, and we used the model and the L–M algorithm to process the wind lidar experiment data, and compared the retrieved results with the radiosonde wind profile. The consistency is very good, especially at an altitude above 1.8 km. We may speculate that when the atmosphere flows are not so dramatic, the lidar and the radiosonde measurements are strictly synchronous, it is possible to retrieve horizontal wind speeds and directions consistently with the radiosonde using our wind lidar model and the L–M algorithm.

Monitoring dam performance is a critical parameter in maintaining a safe dam. Safety concerns include seepage, internal erosion, and seismic issues in the case that the dam is located in high seismic hazard areas. Seismic considerations for dam safety among others includes the expected dam’s performance during seismic events. The scope of this research work concerns the capability to record potential deformation on the Mornos earth dam (central Greece) induced by major earthquake events that occurred in the broader area. For this purpose, a hybrid interferometry synthetic aperture radar (InSAR) method was applied using elements of conventional differential InSAR, short baseline interferometry approaches, and persistent scatterers interferometry. A time series of ascending and descending acquisitions of active microwave instrument/ERS-1 and 2 and advanced synthetic aperture radar/ENVISAT scenes covering the period from 1993 to 2010 were interferometrically combined. Five very strong seismic events with epicenters close to the dam, at the same period, were considered as potential sources of deformation. The lake’s water surface elevation was also considered as an additional factor of induced deformation. Results show a maximum deformation rate of ∼10  cm along the line of sight for the whole period. Although the observed deformation appears to be due to changes in water level following a particular pattern, this is interrupted over time, and these interruptions coincide in time with specific seismic events.

Soil moisture content (SMC) plays an important role in different environmental. In this study, four different soil moisture indices, namely, SOMID, SOMID-FS, SOMID-FT, and CSOMID-FT, were introduced. In this work, the following parameters were used to estimate SMC at a depth of 5 cm: (a) the distance of pixels from the origin in the scatter-plot of near-infrared (NIR) and red bands (SNIR-R), (b) the fraction of soil cover in each pixel, and (c) the land surface temperature. It was concluded that the CSOMID-FT was the most accurate index for estimation of SMC (RMSE=0.045, R=0.92). This index divides the SNIR-R into three separate regions based on the pixels’ normalized difference vegetation index (NDVI) values and assigns a specific regression equation to each region. The results showed that as the NDVI values increase, the accuracy of the proposed indices decreases. Furthermore, the SOMID-FT and CSOMID-FT were used to estimate SMC at five different depths of 5, 10, 20, 50, and 100 cm. It was concluded that the satellite-estimated SMC was highly correlated with the field-measured data at 5-cm soil depth.

We explored a one-class classifier, the support vector data description (SVDD), using the Suomi National Polar-Orbiting Partnership Satellite–Visible Infrared Imaging Radiometer Suite and normalized difference vegetation index to map the urban extent, which was tested in the Beijing and Tianjin city group area. The urban edge-pixels were selected as training samples for SVDD based on a profile-based sampling method combining nighttime light value histograms. The results showed that the overall accuracy of SVDD was similar to the support vector machine (SVM) model. However, kappa coefficients of SVDD for highly developed cities were superior to SVM, as producer and user accuracies of SVDD were almost equal to show high agreement of urban and nonurban areas. For metropolitan areas, such as Beijing and Tianjin, the urban extent generated by SVDD is closer to the reference data. The R² between the quantity of SVDD-estimated urban extent and population, 0.86, was higher than that obtained from SVM, 0.76, indicating that the estimated urban extent from the SVDD is more efficient for understanding the population development. The SVDD was further applied for three other representative metropolitans in China: Shanghai, Guangzhou, and Shenzhen to validate the SVDD’s performance, and similar results were achieved. The success of the SVDD-based urban extent extraction improves our ability to map urban extent at regional and national scales.

An input subset including average radar reflectivity (Z_ave) and its standard deviation (SD) is proposed to improve radar estimates of rainfall based on a radial basis function (RBF) neural network. The RBF derives a relationship from a historical input subset, called a training dataset, consisting of radar measurements such as reflectivity (Z) aloft and associated rainfall observation (R) on the ground. The unknown rainfall rate can then be predicted over the derived relationship with known radar measurements. The selection of the input subset has a significant impact on the prediction performance. This study simplified the selection of input subsets and studied its improvement in rainfall estimation. The proposed subset includes: (1) the Z_ave of the observed Z within a given distance from the ground observation to represent the intensity of a storm system and (2) the SD of the observed Z to describe the spatial variability. Using three historical rainfall events in 1999 near Darwin, Australia, the performance evaluation is conducted using three approaches: an empirical Z−R relation, RBF with Z, and RBF with Z_ave and SD. The results showed that the RBF with both Z_ave and SD achieved better rainfall estimations than the RBF using only Z. Two performance measures were used: (1) the Pearson correlation coefficient improved from 0.15 to 0.58 and (2) the average root-mean-square error decreased from 14.14 mm to 11.43 mm. The proposed model and findings can be used for further applications involving the use of neural networks for radar estimates of rainfall.

Jinan, located in the South of the North China Plain, is an area where underground water has been exploited excessively. However, land deformation surveys only focus on the small district obtained by GPS and Leveling. Here, we use interferometric synthetic aperture radar (InSAR) time-series of ASAR data to resolve land subsidence in the entire Jinan region. In our research, we get 20 interferograms with a temporal threshold of 700 days and spatial-baseline threshold of 300 m from 14 ASAR satellite images on a descending orbit, and then get the surface displacement using Small Baseline InSAR (SBAS D-InSAR) retrained with a periodic model. Meanwhile, the accuracy of our work is proved by the results of GPS measurements. Finally, several settlement funnels are observed with extreme values of −20  cm, and their generation is related to massive groundwater extraction.

Coordinated triple Doppler wind lidars (DWLs) were employed during the Mountain Terrain Atmospheric Modeling and Observations (MATERHORN) program field campaign to observe turbulent winds in the mountain terrain atmospheric boundary layer (ABL). The feasibility of observing large turbulent eddies was investigated by pointing three DWL at an intersecting probe volume adjoining a sonic anemometer mounted on the top of a meteorological tower. The time series and spectra of the sonic anemometer measurement were compared with the lidars. The lidar radial velocities closely followed those of the sonic anemometer, both in time and in the low frequency spectral domain, suggesting that the DWL technique is suitable for observing large turbulent eddies in the ABL. In addition, coordinated scanning triple DWL were used to directly measure the three-dimensional wind vectors, thus circumventing the assumptions required in using single or dual lidar deployments for full velocity measurements. The scanning triple lidar results were in satisfactory agreement with data from tower-based sonic anemometers. Notwithstanding, because of the difficulty of obtaining temporal and spatial synchronizations of the three lidars, the data were scant since a large amount of data had to be rejected in postprocessing. This difficulty is surmountable in the future by employing a robust control system for coordinated scanning.

Soil surface temperature (T_s) is an important indicator of global temperature change and a key input parameter for retrieving land surface variables using remote sensing techniques. Due to the masking in the thermal infrared band and the scattering in the microwave band of snow, the temperature of soil surfaces covered by snow is difficult to infer from remote sensing data. We attempted to estimate T_s under snow cover using brightness temperature data from the special sensor microwave/imager. T_s under snow cover was underestimated due to the strong scattering effect of snow on upward soil microwave emissions at 37 GHz. The underestimated portion of T_s is related to snow properties, such as depth, grain size, and moisture. Based on the microwave emission model of layered snowpacks, the simulated results revealed a linear relationship between the underestimated T_s and the brightness temperature difference (TBD) at 19 and 37 GHz. When TBDs at 19 and 37 GHz were introduced to the T_s estimation method, accuracy improved, i.e., the root mean square error and bias of the estimated T_s decreased greatly, especially for dry snow. This improvement allows T_s estimation of snow-covered surfaces from 37 GHz microwave brightness temperature.

Five healthy Haloxylon ammodendron trees were randomly selected to monitor the photochemical reflectance index (PRI) and its relation to light use efficiency (LUE) in shaded and sunlit assimilating branches. Linear regression was performed where the results revealed that LUE decreased with seasonal courses with vivid midday depression regardless of light variation within canopies. PRI and LUE decoupled in early spring when the foliage was deeply downregulated, with high correlation (R²=0.8, p<0.001) during most parts of the year. Measured and estimated LUE were highly correlated for the July to September dataset (R²=0.73, p<0.001) compared to May to June datasets (R²=0.35, p<0.001). The estimated LUE for young assimilating branches could be obtained with a high RMSE of 0.01, compared to mature branches with an RMSE of 0.003. The measured LUE values were ∼20% higher than estimated LUE during the period from May to June, and values were slightly lower at the end of the growing season. This was attributed to differences in the physiological and biochemical mechanisms controlling the seasonal dynamics of PRI. Therefore, PRI is the best estimator of LUE when age-related changes and canopy light partitioning are considered in assimilating organs. Combined use of PRI with pigments could improve the remote evaluation of LUE.

Monitoring agricultural areas is still a very challenging task. Various models and methodologies have been developed for monitoring the agricultural areas with satellite images, but their practical applicability is limited due to the complexity in processing and dependence on a priori information. Therefore, in this paper, an attempt has been made to investigate the utility of the Kanade–Lucas–Tomasi (KLT) tracker, which is generally useful for tracking objects in video images, for monitoring agricultural areas. The KLT tracker was proposed to deal with the problem of image registration, but the use of the KLT tracker in satellite images for land cover monitoring is rarely reported. Advanced Land Observing Satellite Phased Array type L-band Synthetic Aperture Radar (ALOS PALSAR) data has been used to identify and track the agricultural areas. The tracked pixels were compared with the agriculture pixels obtained from a decision tree algorithm and both results are closely matched. An image differencing change detection technique has been applied after KLT tracker implementation to observe the “change” and “no change” pixels in agricultural areas. It is observed that two kinds of changes are being detected. The areas where agriculture was not there earlier, but now is present, the changes are called positive changes. In the areas where agriculture was present earlier, but now is not present, those changes are referred to as negative changes. Unchanged areas retrieved from both the images are labeled as “no change” pixels. The novelty of the proposed algorithm is that it uses a simplified version of the KLT tracker to efficiently select and track the agriculture features on the basis of their spatial information and does not require a priori information every time.

Super-resolution (SR) as a postprocessing technique is quite useful in enhancing the spatial resolution of hyperspectral (HS) images without affecting its spectral resolution. We present an approach to increase the spatial resolution of HS images by making use of sparse representation and Gabor prior. The low-resolution HS observations consisting of large number of bands are represented as a linear combination of a small number of basis images using principal component analysis (PCA), and the significant components are used in our work. We first obtain initial estimates of SR on this reduced dimension by using compressive sensing-based method. Since SR is an ill-posed problem, the final solution is obtained by using a regularization framework. The novelty of our approach lies in: (1) estimation of optimal point spread function in the form of decimation matrix, and (2) using a new prior called “Gabor prior” to super-resolve the significant PCA components. Experiments are conducted on two different HS datasets namely, 31-band natural HS image set collected under controlled laboratory environment and a set of 224-band real HS images collected by airborne visible/infrared imaging spectrometer remote sensing sensor. Visual inspections and quantitative comparison confirm that our method enhances spatial information without introducing significant spectral distortion. Our conclusions include: (1) incorporate the sensor characteristics in the form of estimated decimation matrix for SR, and (2) preserve various frequencies in super-resolved image by making use of Gabor prior.

This paper aims to evaluate the contribution of multitemporal polarimetric synthetic aperture radar (SAR) data for winter wheat and rapeseed crops parameters [height, leaf area index, and dry biomass (DB)] estimation, during their whole vegetation cycles in comparison to backscattering coefficients and optical data. Angular sensitivities and dynamics of polarimetric indicators were also analyzed following the growth stages of these two common crop types using, in total, 14 radar images (Radarsat-2), 16 optical images (Formosat-2, Spot-4/5), and numerous ground data. The results of this study show the importance of correcting the angular effect on SAR signals especially for copolarized signals and polarimetric indicators associated to single-bounce scattering mechanisms. The analysis of the temporal dynamic of polarimetric indicators has shown their high potential to detect crop growth changes. Moreover, this study shows the high interest of using SAR parameters (backscattering coefficients and polarimetric indicators) for crop parameters estimation during the whole vegetation cycle instead of optical vegetation index. They particularly revealed their high potential for rapeseed height and DB monitoring [i.e., Shannon entropy polarimetry (r²=0.70) and radar vegetation index (r²=0.80), respectively].

The contribution of the integration of optical and polarimetric synthetic aperture radar (PolSAR) data to accurate land-cover classification was investigated. For this purpose, an object-oriented classification methodology that consisted of polarimetric decomposition, hybrid feature selection, and a support vector machine (SVM) was proposed. A RADARSAT-2 Fine Quad-Pol image and an HJ-1A CCD2 multispectral image were used as data sources. First, polarimetric decomposition was implemented for the RADARSAT-2 image. Sixty-one polarimetric parameters were extracted using different polarimetric decomposition methods and then merged with the main diagonal elements (T₁₁, T₂₂, T₃₃) of the coherency matrix to form a multichannel image with 64 layers. Second, the HJ-1A and the multichannel images were divided into numerous image objects by implementing multiresolution segmentation. Third, 1104 features were extracted from the HJ-1A and the multichannel images for each image object. Fourth, the hybrid feature selection method that combined the ReliefF filter approach and the genetic algorithm (GA) wrapper approach (ReliefF–GA) was used. Finally, land-cover classification was performed by an SVM classifier on the basis of the selected features. Five other classification methodologies were conducted for comparison to verify the contribution of optical and PolSAR data integration and to test the superiority of the proposed object-oriented classification methodology. Comparison results show that HJ-1A data, RADARSAT-2 data, polarimetric decomposition, ReliefF–GA, and SVM have a significant contribution by improving land-cover classification accuracy.

In the northern region of India, hailstorms are common phenomena during winter. Four cases of hailstorms around Delhi (28° 58975′ N, 77° 22195′ E) region during winter have been analyzed in the present study. It is mainly based on the observations of polarimetric radar variables (ZH and ZDR), as observed by Delhi C-band Doppler weather radar and supplementary thermodynamic variables (CAPE, CIN, wind shear). The thermodynamic properties of the atmosphere during the hailstorm events have been studied using radiosonde observations. The hailstorms over the study region are classified into two types with and without large CAPE. Although ZH is higher for all events, the ZDR differs for the two categories. The events with small CAPE and strong shear produce storms with larger ZDR (rain mixed with small hail), while those with large CAPE and weak shear produce smaller ZDR (strong hail).

The performance of synthetic aperture radar (SAR) image interpretation directly depends on the image quality. However, conventional SAR image-quality indicators are measured to check whether the SAR system has maintained its performance specifications, not to assess how well the SAR image can serve image interpretation. An SAR image-quality assessment method based on the modulation transfer function (MTF) is proposed for image interpretation, which jointly reflects the resolution and the contrast of the SAR image. In addition, it describes the imaging performance of the SAR system at different spatial frequencies. Specifically, we propose a feasible MTF test field and realize it in Shanghai Jiao Tong University, Shanghai, China. Then, we give the MTF measurement procedure and utilize the MTF curve to evaluate SAR image quality. Simulation results demonstrate that the proposed MTF-based method is more accurate to assess the SAR image quality than the conventional methods. In addition, the spaceborne SAR experiments are carried out by the TerraSAR-X sensor and the experimental results are given to confirm the benefits of the proposed method.

The significance of laser return intensity has been widely verified in airborne light detection and ranging (LiDAR)-based forest canopy mapping, but this does not mean that all of its roles have been played. People still ask such questions as “Is it possible using this optical attribute of lasers to investigate individual tree-crown insides wherein laser intensity data are typically yielded in complicated echo-triggering modes?” To answer this question, this study examined the characteristics of the intensities of the laser points within 10 Quercus robur trees by fitting their peak amplitudes into default Gaussian distributions and then analyzing the resulting asymmetric tails. Exploratory data analyses showed that the laser points lying within the distribution tails can indicate primary tree branches in a sketchy way. This suggests that the question can be positively answered, and the traditional restriction of airborne LiDAR in canopy mapping at the crown level has been broken. Overall, this study found a unique way to detect primary tree branches in airborne LiDAR data and pointed out how to explore more ways this optical intensity attribute of airborne LiDAR data can measure tree organs at fine scales and further learn their properties.

Conventional methods for monitoring salt accumulation within irrigation schemes involve regular field visits to collect soil samples for laboratory analysis. Remote sensing has been proposed as a less time-consuming, more cost-effective alternative as it provides imagery covering large areas throughout the year. This study evaluated the efficacy of very high resolution (VHR) WorldView-2 imagery to map areas affected by salt accumulation. Classifications based on thresholds obtained from Jeffries–Matusita distance, regression modeling, classification and regression trees, as well as supervised classification approaches, were evaluated for discriminating between salt-affected and unaffected soils in Vaalharts, South Africa. The WorldView-2 bands were supplemented with salinity indices (SIs), principal components, and texture measures to increase the number of predictive variables. In situ soil samples were used for model development, classifier training, and accuracy assessment. The results showed that a simple threshold implemented on a normalized difference SI was the most successful in separating classes, with an overall accuracy of 80%. The findings suggest that VHR satellite imagery holds much potential for monitoring salt accumulation, but more research is needed to investigate why the classification results tend to overestimate salt-affected areas. More work is also needed to evaluate the transferability of the techniques to other irrigation schemes.

Various polarimetric features including scattering matrix, covariance matrix, polarimetric decomposition results, and textural or spatial information have already been used for polarimetric synthetic aperture radar (PolSAR) image classification. However, color features are rarely involved. We propose an improved superpixel-based PolSAR image classification integrating color features. First, we extract the color information using polarimetric decomposition. Second, by combining the color and spatial information of pixels, modified simple linear iterative clustering is used to generate small regions called superpixels. Then we apply Wishart distance to the superpixels to classify them into different classes. This method is demonstrated using the L-band Flevoland PolSAR data from AirSAR and Oberpfaffenhofen PolSAR data from ESAR. The results show that this method works well for areas with homogeneous terrains like farms in terms of both classification accuracy and computational efficiency. Furthermore, the success of the proposed method signifies that more color features can be discovered in the future research works.

We constructed a lightweight unmanned aerial vehicle (UAV) remote sensing system and determined the ideal method for equipment setup, image acquisition, and image processing. Fields of rice paddy (Oryza sativa cv. Unkwang) grown under three different nitrogen (N) treatments of 0, 50, or 115  kg/ha were monitored at Chonnam National University, Gwangju, Republic of Korea, in 2013. A multispectral camera was used to acquire UAV images from the study site. Atmospheric correction of these images was completed using the empirical line method, and three-point (black, gray, and white) calibration boards were used as pseudo references. Evaluation of our corrected UAV-based remote sensing data revealed that correction efficiency and root mean square errors ranged from 0.77 to 0.95 and 0.01 to 0.05, respectively. The time series maps of simulated normalized difference vegetation index (NDVI) produced using the UAV images reproduced field variations of NDVI reasonably well, both within and between the different N treatments. We concluded that the UAV-based remote sensing technology utilized in this study is potentially an easy and simple way to quantitatively obtain reliable two-dimensional remote sensing information on crop growth.

Radar performance has improved substantially from the first prototypes until current modern systems. Nowadays, technical advances run parallel to the development of new materials and shapes able to shield different targets or even make them invisible to the radar. This paper introduces the shielding phenomena from another point of view, dealing with the usage of forested radio propagation channels to degrade radar performance. Based on measurements carried out in five different forest and meadow scenarios, we study the probability of detection of a target located within such environments. Depending on the type of target, the frequency, and the vegetated surroundings, we demonstrate that a forest can provide radar invisibility even to large targets, reducing the probability of detection to values below 0.1.

Three-dimensional (3-D) geodata of mass wasting sites are important to model surfaces, volumes, and their changes over time. With a photogrammetric approach commonly known as structure from motion, 3-D point clouds can be derived from image collections in a straightforward way. The quality of point clouds covering a quarry dump derived from terrestrial and aerial imagery is compared and assessed. A comprehensive set of quality indicators is calculated and compared to surveyed reference data and to a terrestrial LiDAR point cloud. The examined indicators are completeness of coverage, point density, vertical accuracy, multiscale point cloud distance, scaling accuracy, and dump volume. It is found that the photogrammetric datasets generally represent the examined dump well with, for example, an area coverage of up to 90% and 100% in case of terrestrial and aerial imagery, respectively, a maximum scaling difference of 0.62%, and volume estimations reaching up to 100% of the LiDAR reference. Combining the advantages of 3-D geodata derived from terrestrial (high detail, accurate volume calculation even with a small number of input images) and aerial images (high coverage) can be a promising method to further improve the quality of 3-D geodata derived with low-cost approaches.

Small baseline subsets interferometric synthetic aperture radar technique is analyzed to detect and monitor the loess landslide in the southern bank of the Jinghe River, Shaanxi province, China. Aiming to achieve the accurate preslide time-series deformation results over small spatial scale and abrupt temporal deformation loess landslide, digital elevation model error, coherence threshold for phase unwrapping, and quality of unwrapping interferograms must be carefully checked in advance. In this experience, land subsidence accompanying a landslide with the distance <1  km is obtained, which gives a sound precursor for small-scale loess landslide detection. Moreover, the longer and continuous land subsidence has been monitored while deformation starting point for the landslide is successfully inverted, which is key to monitoring the similar loess landslide. In addition, the accelerated landslide deformation from one to two months before the landslide can provide a critical clue to early warning of this kind of landslide.

This study evaluates Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and Hyperion hyperspectral sensor datasets to detect advanced argillic minerals. The spectral signatures of some alteration clay minerals, such as dickite and alunite, have similar absorption features; thus separating them using multispectral satellite images is a complex challenge. However, Hyperion with its fine spectral bands has potential for good separability of features. The Spectral Angle Mapper algorithm was used in this study to map three advanced argillic alteration minerals (alunite, kaolinite, and dickite) in a known alteration zone in the Peruvian Andes. The results from ASTER and Hyperion were analyzed, compared, and validated using a Portable Infrared Mineral Analyzer field spectrometer. The alterations corresponding to kaolinite and alunite were detected with both ASTER and Hyperion (80% to 84% accuracy). However, the dickite mineral was identified only with Hyperion (82% accuracy).

The quantification of grassland grazing intensity (GI) and its detailed spatial distribution are important for grassland management and ecological protection. Remote sensing has great potential in these areas, but its use is still limited. This study analyzed the impacts of grazing on biophysical properties of vegetation and suggested using biomass to quantify GI because of its stability and interpretability. In comparison to a single spectral index, such as the red edge index (REI), combining REI and a cellulose absorption ratio index calculated from hyperspectral data performs better for biomass estimation. Further, an auxiliary spectral index, called the grazing monitoring index (GMI), was developed based on differences in spectral reflectance in the infrared range. Experiments in a grazing area of the Inner Mongolia grassland indicated that GMI can identify GI, with three range intervals (GMI <0, 0–1, and ≥1) used to describe the biomass distribution. The results showed that combining GMI and biomass was more successful than existing approaches for identifying the grassland variability resulting from the spatial heterogeneity of grazing behavior. The thresholds of biomass for four GI levels (ungrazed, lightly grazed, moderately grazed, and heavily grazed) could be determined by the intersections of biomass distributions. In addition, the approach developed at the on-ground canopy scale was extended to remotely sensed Hyperion data. The results showed that the approach could successfully identify the grazing treatments of blocks in the experimental grazing area. Overall, our study provides inspiration and ideas for using satellite remote sensing for evaluating plant production, standing biomass, and livestock impacts.

A modification of the iterative procedure of the surface energy balance was purposed to expedite the convergence of Monin–Obukhov stability correction utilized by the remote sensing based flux calculation. This was demonstrated using ground-based weather stations as well as the gridded weather data (North American Regional Reanalysis) and remote sensing based (Landsat 5, 7) images. The study was conducted for different land-use classes in southern Idaho and northern California for multiple satellite overpasses. The convergence behavior of a selected Landsat pixel as well as all of the Landsat pixels within the area of interest was analyzed. Modified version needed multiple times less iteration compared to the current iterative technique. At the time of low wind speed (∼1.3  m/s), the current iterative technique was not able to find a solution of surface energy balance for all of the Landsat pixels, while the modified version was able to achieve it in a few iterations. The study will facilitate many operational evapotranspiration models to avoid the nonconvergence in low wind speeds, which helps to increase the accuracy of flux calculations.

The urban thermal environment is an important element for the urban ecological environment and climate. As megacities are affected by severe thermal environment, this paper selected Landsat 8 to retrieve land surface temperature (LST) studying the thermal environment of five megacities in China including Beijing, Shanghai, Guangzhou, Tianjin, and Chengdu. Three methods have been applied, quantifying the surface urban heat island intensity, landscape pattern metrics, and spatial autocorrelation. Three main conclusions have been drawn as follows. First, high-LST area is located in the central urban area. Second, the medium-temperature region is the most prevalent. The class-based and the landscape-based metrics can detect the pattern of thermal landscape. The fragmentation is low both in low and high temperature level classes. Third, global Moran’s I suggests there is spatial clustering of thermal landscape. Local Moran’s I map was able to detect several high-high and low-low clusters, which are the main types of thermal landscape.

Precise estimation of carotenoids (Car) content in plants, from remotely sensed data, is challenging due to their small proportion in the overall total pigment content and to the overlapping of spectral absorption features with chlorophyll (Chl) in the blue region of the spectrum. The use of narrow band vegetation indices (VIs) obtained from hyperspectral data has been considered an effective way to estimate Car content. However, VIs have proved to lack sensitivity to low or high Car content in a number of studies. In this study, the carotenoid triangle ratio index (CTRI), derived from the existing modified triangular vegetation index and a single band reflectance at 531 nm, was proposed and employed to estimate Car canopy content. We tested the potential of three categories of hyperspectral indices earlier proposed for Car, Chl, Car/Chl ratio estimation, and the new CTRI index, for Car canopy content assessment in winter wheat and corn. Spectral reflectance representing plant canopies were simulated using the PROSPECT and SAIL radiative transfer model, with the aim of analyzing saturation effects of these indices, as well as Chl effects on the relationship between spectral indices and Car content. The result showed that the majority of the spectral indices tested, saturated with the increase of Car canopy content above 28 to 64  μg/cm². Conversely, the CTRI index was more robust and was linearly and highly sensitive to Car content in winter wheat and corn datasets, with coefficients of determination of 0.92 and 0.75, respectively. The corresponding root mean square error of prediction were 6.01 and 9.70  μg/cm², respectively. Furthermore, the CTRI index did not show a saturation effect and was not greatly influenced by changes of Chl values, outperforming all the other indices tested. Estimation of Car canopy content using the CTRI index provides an insight into diagnosing plant physiological status and environmental stress.

Kriging interpolation provides the best linear unbiased estimation for unobserved locations, but its heavy computation limits the manageable problem size in practice. To address this issue, an efficient interpolation procedure incorporating the fast Fourier transform (FFT) was developed. Extending this efficient approach, we propose an FFT-based parallel algorithm to accelerate regression Kriging interpolation on an NVIDIA^® compute unified device architecture (CUDA)-enabled graphic processing unit (GPU). A high-performance cuFFT library in the CUDA toolkit was introduced to execute computation-intensive FFTs on the GPU, and three time-consuming processes were redesigned as kernel functions and executed on the CUDA cores. A MODIS land surface temperature 8-day image tile at a resolution of 1 km was resampled to create experimental datasets at eight different output resolutions. These datasets were used as the interpolation grids with different sizes in a comparative experiment. Experimental results show that speedup of the FFT-based regression Kriging interpolation accelerated by GPU can exceed 1000 when processing datasets with large grid sizes, as compared to the traditional Kriging interpolation running on the CPU. These results demonstrate that the combination of FFT methods and GPU-based parallel computing techniques greatly improves the computational performance without loss of precision.

Nanjing, a typical megacity in eastern China, has undergone dramatic expansion during the past decade. The surface urban heat island (SUHI) effect is an important indicator of the environmental consequences of urbanization and has rapidly changed the dynamics of Nanjing. Accurate measurements of the effects and changes resulting from the SUHI effect may provide useful information for urban planning. Index, centroid transfer, and correlation analyses were conducted to measure the dynamics of the SUHI and elucidate the relationship between the SUHI and urban expansion in Nanjing over the past decade. Overall, the results indicated that (1) the region affected by the SUHI effect gradually expanded southward and eastward from 2000 to 2012; (2) the centroid of the SUHI moved gradually southeastward and then southward and southwestward, which is consistent with the movement of the urban centroid; (3) the trajectory of the level-3 SUHI centroid did not correspond with the urban mass or SUHI centroids during the study period and (4) the SUHI intensity and urban fractal characteristics were negatively correlated. In addition, we presented insights regarding the minimization of the SUHI effect in cities such as Nanjing, China.

Dynamic assessment of heavy metal contamination in crops is essential for food security and the farmland ecological environment. A new index for monitoring heavy metal stress based on the assimilation of synthetic aperture radar (SAR) data and the crop growth model is performed. The improved World Food Study (WOFOST) model was used in this study, which is embedded with two stress factors to improve the accuracy of assimilation. Biomass (BM) values retrieved by SAR data were assimilated into the improved WOFOST model to simulate dry weight of rice roots (WRT), and the root mass ratio (RMR, WRT/BM) was calculated as an index for monitoring heavy metal stress. SAR shows enormous potential for monitoring crop growth status in cloudy area. Compared with other physiological indices, RMR could weaken the weight change of rice caused by other background factors. In the temporal scale, RMR showed a faster significant decrease when the stress was greater. The spatial distribution of RMR and the stress factors exhibited good consistency. These results suggest that RMR derived from the assimilation method based on SAR data and the improved WOFOST model is effective for dynamically monitoring the rice growth status in cloudy regions under heavy metal stress.

The use of landscape metrics to characterize the morphological behavior of a landscape has been extensive in the last few years. It is recognized that a single metric is insufficient to characterize a landscape. Such metrics are used individually to derive the morphological aspect of a landscape. No joint use of various metrics has been reported. Therefore, we considered the joint use of landscape metrics in a multivariate classification. We derived the value of a number of landscape metrics of patches from several case studies. A multivariate classification was applied using a hierarchical clustering algorithm. The multivariate classification was carried out using the least correlated landscape metrics. To consider the multivariate classification, a normalization of metrics range was used. The results provided the morphological structure of patches grouped into four or five classes. The classes depicted a morphological structure of patches that ranged from simple to very complex. An index was proposed to quantify the morphological structure of a class-patch. Such an index was defined as the average of the landscape metrics for a class-patch. The distance among the class-patch was given by means of the Jeffries–Matusita distance.

Terrestrial laser scanning (TLS) has been used to extract accurate forest biophysical parameters for inventory purposes. The diameter at breast height (DBH) is a key parameter for individual trees because it has the potential for modeling the height, volume, biomass, and carbon sequestration potential of the tree based on empirical allometric scaling equations. In order to extract the DBH from the single-scan data of TLS automatically and accurately within a certain range, we proposed an adaptive circle-ellipse fitting method based on the point cloud transect. This proposed method can correct the error caused by the simple circle fitting method when a tree is slanted. A slanted tree was detected by the circle-ellipse fitting analysis, then the corresponding slant angle was found based on the ellipse fitting result. With this information, the DBH of the trees could be recalculated based on reslicing the point cloud data at breast height. Artificial stem data simulated by a cylindrical model of leaning trees and the scanning data acquired with the RIEGL VZ-400 were used to test the proposed adaptive fitting method. The results shown that the proposed method can detect the trees and accurately estimate the DBH for leaning trees.

This PDF file contains the errata for “JRS Vol. 10 Issue 02 Paper JARS-2016-0418-ERR” for JRS Vol. 10 Issue 02