This is a list of reviewers who served the Journal of Applied Remote Sensing in 2020.

Vegetated roofs provide many ecosystem services and support urban biodiversity. While it would be interesting to study the contribution of vegetated roofs to ecological corridors, vegetated roofs are listed in no French databases. Because of their intrinsic nature as roofs, their small number, their small size, and the type of vegetation planted on them, vegetated roofs seem to be very difficult to identify. We propose a method to automatically identify vegetated roofs. Using infrared aerial photographs and building shape data, we were able to build a model detecting vegetated roofs using remote sensing and supervised classification techniques. The major difficulty lies in distinguishing between real vegetated roofs and roofs partially covered by tree foliage growing on the ground. In this operation, our classification model obtains an error rate of ∼18  %  . We improve the knowledge of vegetation detection in cities. Moreover, it opens interesting perspectives on the analysis of ecological networks in cities as a function of building height. In addition, it could be an interesting tool for municipalities to monitor urban vegetation development and to prioritize vegetated roofs planning.

In optical remote sensing images, the aircraft to be detected is very small; external environmental factors such as cloud occlusion, aircraft, and the site background are easily fused; and the interference of objects to aircraft has a great impact on the aircraft characteristics in remote sensing images. In response to the above problems, we designed a remote sensing aircraft detection method based on deep learning. First, to ensure the feature extraction capability and limit the number of calculations of the network, the LightNet v2 network unit is designed, and it constitutes an efficient backbone network. In addition, spatial pyramid pooling of residual ideas (Res-SPP) is performed on the output results of the backbone network. Res-SPP is used to separate more important contextual features while using almost no computing space. A multi-scale fusion prediction network (MFPN) is proposed to perform feature fusion from multiple angles to achieve a rich combination of gradients. The MFPN enhances the network’s ability to detect extremely small objects and can improve accuracy while ensuring that the method is lightweight. Finally, according to the judgment of the threshold, the dark channel defogging method, which enhances the ability to detect aircraft in cloudy and foggy scenes, is used for remote sensing images full of clouds and fog. The experimental results show that the proposed method can detect airplanes, especially very small airplanes, in various scenarios. The amount of calculation of the method is 23.56 BN, the model volume is 56 MB, the speed on a GTX 1080 platform reaches 157 frames per second (FPS), and the F1  %   on the remote sensing aircraft data set reaches 99.2. In particular, this method can be embedded in an ordinary field programmable gate array platform due to its lightweight characteristics, and the calculation speed can reach 32 FPS.

This study developed a quadratic discriminant analysis (QDA) model from the spectroradiometer reflections (400 to 1000 nm) and phosphorus (P) uptake in wheat under varying rates of P dosages (0, 25, and 50 ppm P) in the tillering (GS25) and heading (GS55) stages. Seventy-two experimental plants were grown under controlled greenhouse conditions. Stepwise multiple regression analysis was used to determine the wavelengths associated with different periods and P doses. Principal component analysis was employed to select the five wavelengths (418, 563, 639, 756, and 1000 nm) that best encompassed the total variance amongst the different reflection values. The QDA model assigned the training data to their real classes (0 ppm P: 79%, 25 ppm P: 50%, and 50 ppm: 83%) with 71% accuracy. For validation of the model, 36 randomly selected test data were used (0 ppm P: 75%, 25 ppm P: 42%, and 50 ppm P; 92%) and resulted in 69% accuracy. Results concluded that wheat P demand during different vegetation stages can be determined from the spectral wavelengths input into a QDA model; for future research, however, we suggest the nutrient dosage ranges are broad enough to provide sufficient variability. Nevertheless, discriminant modeling is a viable method of determining plant nutritional status by spectral data.

Crop type mapping provides essential information to control and make decisions related to agricultural practices and their regulations. To map crop types accurately, it is important to capture their phenological stages and fine spatial details, especially in a temporally and spatially heterogeneous landscape. The data availability of new generation multispectral sensors of Landsat-8 (L8) and Sentinel-2 (S2) satellites offers unprecedented options for such applications. Given this, our study aims to display how the synergistic use of these optical sensors can efficiently support crop type mapping research while integrating an object-based image analysis (OBIA). Through the applied methods, we used the enhanced spatial and temporal adaptive reflectance fusion model (ESTARFM) and the flexible spatiotemporal data-fusion model (FSDAF) to blend L8 and S2 data and obtain reliable normalized difference vegetation index (NDVI) datasets with fine spatial and temporal resolution. Then the crop phenological information was extracted using a Savitzky–Golay filter and fused NDVI time series. Finally, a model combining phenological metrics and fused reconstructed NDVI as classification features was developed using a random forest (RF) classifier/OBIA approach. The results show that the FSDAF method creates more accurate fused NDVI and keeps more spatial details than ESTARFM. The FSDAF model was then used to create fused, high-resolution time-series products that were able to extract crop phenology in single-crop fields while providing a very detailed pattern relative to that from individual sensor time-series data. Moreover, combined L8 and S2 data by FSDAF produced highly significant overall classification accuracies (90.03% for pixel-based RF to 93.12% OBIA RF), outperforming individual sensor use (82.57% for L8-only; 88.45% for S2-only). Our proposed workflow highlights the advantage of spatiotemporal fusing and OBIA environment in spatiotemporally heterogeneous areas and fragmented landscapes, which represents a promising step toward generating fast, accurate, and ready-to-use agricultural data products.

Airborne laser remote sensing is extensively used in geophysical explorations. Laser-remote-sensing studies are often undertaken to predict the availability of mineral resources in a target area. A major issue encountered in such studies concerns the acquisition of the ground-vibration information and analysis of the optoelectric characteristics of the corresponding target vibration. Our paper presents a scheme involving the use of a laser remote sensing system (LRSS) to realize the geophone-detection function and experimentally verify the optoelectric characteristics of the target ground vibration. As observed, the proposed LRSS demonstrates a 19- to 98-V s  /  cm sensitivity range for the low- to medium-frequency vibrations in the 0- to 600-Hz range. Overall, the LRSS could sense a wide range (0 to 4 kHz) of ground-vibration frequencies. Our research is expected to open avenues for electro-optical sensing technologies to be used in applications such as hidden-structure detection and non-line-of-sight imaging.

The use of the single machine learning classifier for high-resolution remote sensing (RS) image classification makes it difficult to improve the accuracy of classification results. To fully utilize the advantages of different classifiers for different types of ground objects, based on the Dempster–Shafer (DS) evidence theory, we propose a multi-classifier fusion method for classification of high-resolution RS images. Six machine learning classifiers: support vector machines, k-nearest neighbor, random forest, artificial neural network, classification and regression tree, and the C5.0 decision tree were selected for application in the fusion of multiple classifiers. We calculated a classifier difference index based on the accuracy and difference of the classification results of the base classifiers. Base classifiers with large differences were selected to perform integration based on the DS evidence theory. We also improved the classical DS evidence theory. First, based on the classification validity of the base classifier for different ground objects, the classification probability value of the base classifier for different samples was weight optimized. Then different fusion methods were selected according to the classification conflict coefficients between base classifiers. The results reveal that the overall accuracy and kappa coefficients of the fusion classifier are significantly better than those of the base classifier. The producer’s accuracy and user’s accuracy of the fusion results based on the improved DS evidence theory were higher than those of the fusion results based on the classical DS evidence theory.

Surface-water body maps are imperative for effective mosquito larvae control. This study aims to select a method for the automatic and regular mapping of surface-water bodies in rice fields and wetlands using Sentinel-1 synthetic aperture radar data. Four methods were adapted and developed for automated application: the Otsu valley-emphasis algorithm, a classification method based on the textural feature of entropy, a method using K-means unsupervised classification, and a method using the Haralick’s textural feature of dissimilarity and fuzzy-rules classification. The results were assessed using field data collected during the mosquito breeding periods of 2018 and 2019 in the region of Central Macedonia (Greece). The Otsu valley-emphasis technique provides the highest overall accuracy (0.835). The accuracy is higher at the beginning of the summer (0.948) than at the end of the rice-growing season due to higher density of vegetation. Results using this method were further assessed during the main larvicide application period. The presence of vegetation, built-up areas, floating algae in rice-paddies, salt-crust formations in wetlands, and water depth, were found to affect the performance of the algorithm. A WebGIS platform was designed for the visualization of the produced water maps along with other data related to mosquito-larvae presence.

Recently, there has been increasing attention to the digital elevation model (DEM) because of its ability to learn the Earth’s surface’s topography. Freely accessible DEMs such as the Cartosat-1, shuttle radar topography mission (SRTM), light detection and ranging, and advanced spaceborne thermal emission and reflection radiometer DEM, contain large vertical errors. The errors are aggravated over multifaceted geography and cannot rectify microgeographic deviations in the moderately flat landscape. As high-accuracy DEMs have limited availability, dated low-accuracy DEMs are still used in various models, specifically in the data-sparse areas. However, it is necessary to enhance the quality of these DEMs before their use in the geomorphometric analysis. We aim to investigate the effect of noise reduction filters on DEM’s accuracy and quality. The noise reduction filters such as weighted average filter, median filter, sharpen filter, Lee sigma filter, and local sigma filter are used for noise reduction in the stereo images. The quality and accuracy of the generated DEMs are further improved by selecting an optimum number of tie points in the image matching process. The effects are assessed by correlating the surface profiles for the final DEM obtained and SRTM DEM as a reference with a resolution of 1 arc sec.

The fusion of multispectral and synthetic aperture radar (SAR) images is of vital importance in many remote sensing applications. Spectral distortion and trade-off between the spatial and spectral quality of the fused image are significant issues in SAR-multispectral image fusion. Our study attempts to improve the performance of SAR-multispectral image fusion concerning these two issues. The primary objective of our study is to optimize the performance of hybrid fusion approach based on principal component analysis and discrete wavelet transform (PCA-DWT) using Taguchi orthogonal array. The fused data are evaluated using visual analysis and standard quality metrics. The results are compared with recent hybrid fusion approaches applied to the SAR-multispectral image fusion. The utility of the fused data is evaluated based on the remote sensing application, namely, land use land cover classification. The classification results are compared to a standard thematic map available on the Bhuvan geoportal to check classification accuracy. A comparative analysis with recent hybrid approaches conclusively demonstrates that the proposed optimization in the PCA-DWT based fusion is superior to conventional hybrid methods.

Deforestation and forest degradation are two important contributors to atmospheric emissions of carbon. To claim carbon credit for conservation plus enhancement of existing forest carbon stock under reducing emissions from deforestation and forest degradation (REDD+) program, a post-Kyoto climate change mitigation initiative, developing countries need to assess the level of respective forest biomass-carbon and corresponding emissions of carbon in a complete and verifiable manner. We aimed to develop a comprehensive approach following the recommended inter-governmental panel on climate change good practice guidelines to estimate emission factors and emissions of biomass-carbon due to deforestation and forest degradation at Raghunandan Hills Reserved Forest, Bangladesh, for REDD+ implications. The estimation was done using a combination of field and satellite data employing geospatial techniques. The critical assessment provided a comprehensive estimate of the emission factors and emissions of biomass-carbon in the form of maps and statistics with acceptable accuracy. The approaches and findings of this study may have important scientific and management implications including providing baseline information for biomass-carbon stock-related concerns such as REDD+.

The coffee crop spectral behavior identification throughout its cycle can contribute to its development monitoring under pest incidence. We aim to identify coffee development through time signatures of enhanced vegetation index (EVI), as well as to evaluate the use of seasonal autoregressive integrated moving average (SARIMA) models to identify coffee trees spectrum-time patterns under different irrigation management and design future scenarios. Three coffee fields were selected under different irrigation systems, whose EVI data of 8 years were obtained from the moderate resolution image spectroradiometer sensor. Each coffee crop model was subjected to residual autocorrelation test and classified according to information criteria, while its accuracy was assessed by means of prediction error measures and agreement index. The estimated and observed EVI values were similar, even for the predicted year. However, in agricultural years during which coffee diseases occurred, the crops showed vegetative vigor below the expected. We concluded that SARIMA models enabled the establishment of a reliable spectral signature expected for coffee crop, which could help with crop management defining, regardless of the irrigation system adopted. Based on the evaluation of divergence between expected and observed spectral signatures, early signs of coffee underdevelopment were detected, which could reduce economic loss risks on its commercial chain.

Road extraction as a significant role in traffic management, road monitoring, and autodriving cars have been important research topics in recent years. A deep learning method based on U-net and spatially squeeze and exciting channelwise (SE) is proposed for recognizing road. The SE attention block reweights feature maps from U-net layers and highlights only useful channels. We added this block to U-net and test out proposed method on two road datasets. Finally, we compare our method with other methods, and the results demonstrate that the proposed method outperforms other networks.

Mapping the refined land-cover classification mapping (RLCM) is a primary and essential strategy for evaluating the ecological change and understanding the ecosystem services. A common problem during the generation of RLCM is a scale mismatch between remote sensing (RS) data and field quadrat, which leads to inaccuracy of the classification result. A multi-scale transformation method was developed via integrating RS, unmanned aerial vehicle (UAV), and field surveys in Sanjiangyuan National Park (SNP). With the help of UAV, a large number of virtual biomass quadrats were resampled and interpolated, and the quantitative thresholds of different vegetation coverage in alpine meadow and steppe were determined to improve land-cover classification accuracy. Based on the spatial-temporal analysis of RLCM from 1990 to 2017, the whole ecological coverage was becoming better, and its driving factor was attributed to government policy and climate change. This study can provide a practical suggestion for the management and sustainable development in SNP.

Surface roughness is a crucial parameter for the estimation of soil moisture (SM). The present study attempted to optimize the surface roughness parameter (h) for the estimation of SM from Soil Moisture Active Passive (SMAP) using tau–omega (τ-ω) model and also downscaled the estimated SM product using a polynomial regression relation among Normalized Difference Vegetation Index (NDVI), land surface temperature (LST), and SM. The brightness temperature of SMAP available at two spatial resolutions (36 and 9 km) was used for two seasons intended for SM assessment. After assessment with in-situ SM data, 9-km SM data values were further used for spatial disaggregation to obtain the optimized downscaled soil moisture (ODSM) at 1 km. Results showed that the variation in the value of the roughness parameter strongly affects the performance of the τ-ω model and the downscaling performances. The investigation provided lowest values of root-mean-square error (RMSE) to be 0.0518 (at h  =  0.35) and 0.0480 (at h  =  0.25) for the SM estimation at 36 km for the different seasons used in this study while the lowest values of RMSE for ODSM were found to be 0.0365 (at h  =  0.4) and 0.0252 (at h  =  0.25, 0.3) for different seasons.

We propose the convolutional neural networks approach for detecting urban slums from very high resolution (VHR) satellite images. Slums are becoming an inevitable and growing phenomenon in cities of the global south, whose locations are inappropriate in official statistics and maps. Thus the automatic detection and identification of slums provides vibrant information to decision-makers for formulating pro-poor policies in urban planning. However, field surveys are used as the conventional methods for slum detection, which are expensive and inefficient. The challenge is to find an automatic approach for identifying slums from VHR imagery. Although numerous studies focused on detecting slums from satellite data, only limited number captured their differences. Due to the unique spectral signatures of urban slums, it could not be simply classified as other urban (buildings, vegetation, and roads) features. We explore the potential of the dilated kernel-based deep convolutional neural network (DK-DCNN) approach to the learning of discriminatory spatial features and the automatic detection of slums from other features. However, the result obtained by the proposed DK-DCNN achieves high accuracy. The open area of informal settlements and the roofing structures of formal settlements are misclassified as slums. Morphological spatial pattern analysis, based on mathematical morphology, is used as a postprocessing method to enhance the classification accuracy. The four distinct very high-resolution satellite images captured by WorldView-2 Sensor (1.84 m) of Madurai and Tiruppur city, South India, have shown the performance of the proposed method to distinguish urban slums from other features by producing higher accuracy than any other approach.

Manual evaluation of crop injury to herbicides is time-consuming. Unmanned aerial systems (UAS) and high-resolution multispectral sensors and machine learning classification techniques have the potential to save time and improve precision in the evaluation of herbicide injury in crops, including grain sorghum (Sorghum bicolor L. Moench). The objectives of our research are to (1) evaluate three supervised classification algorithms [support vector machine (SVM), maximum likelihood, and random forest] for categorizing high-resolution UAS imagery to aid in data extraction and (2) evaluate the use of vegetative indices (VIs) collected from UAS imagery as an alternative to traditional methods of visible herbicide injury assessment in mesotrione-tolerant grain sorghum breeding trials. An experiment was conducted in a randomized complete block design using a factorial treatment arrangement of three genotypes by four mesotrione doses. Herbicide injury was rated visually on a scale of 0 (no injury) to 100 (complete plant mortality). The UAS flights were flown at 9, 15, 21, 27, and 35 days after treatment. Results show the SVM algorithm to be the most consistently accurate, and high correlations (r  =    −  0.83 to −0.94; p  <  0.0001) were observed between the normalized difference VI and ground-measured herbicide injury. Therefore, we conclude that VIs collected with UAS coupled with machine learning image classification has the potential to be an effective method of evaluating mesotrione injury in grain sorghum.

The lack of data on water consumption by agriculture in the context of climate change led the authors to estimate the water balance of the Bekaa Valley. Plots were mapped according to the normalized difference vegetation index profiles of spring and summer crop with the Sentinels-2 images. Surface energy balance model python (PYSEBAL) estimated the seasonal evapotranspiration (ET_season) and net irrigation requirements (NIR_season). Spring and summer crop maps were accurate: 90% and 93%. The results of PYSEBAL were validated by the comparison between averaged daily ET of PYSEBAL and FAO-56 method (root mean square error from 0.38 to 1.93 and the mean average error from 0.33 to 1.64  mm  /  day). The water balance was negative: 0.16  km³ of groundwater was extracted against 0.10  km³ of rainfall available for aquifer recharge in 2017. This is explained by ET/ha/season (alfalfa: 1094  mm  /  ha  /  season, orchards 881, corn 719, tobacco 606, cucumber 558, late potatoes 487, summer vegetables 483, early potatoes 467, vineyard 390, spring vegetables 237, and wheat 313). ET_season varies according to temperature (strong correlation between daily ET and temperature, Pearson: 0.74), relative humidity (Pearson: −0.65), surface radiation (Pearson: 0.75), and percentage of sand in the soil and its depth (strong negative Pearson: −0.62 and −0.57).

Urban agglomerations (UA) are the fastest growing regional types during recent years, especially in developing countries. Monitoring and evaluating the eco-environment quality of UA is significant for sustainability. The previous remote sensing models of urban eco-environment are generally based on the vegetation-impervious surface-soil framework, which neglects the air quality in urban areas. We constructed a vegetation-impervious surface-soil-air (VISA) framework and derived a remote sensing model of UA eco-environment (RSUAE) based on the VISA. The RSUAE can integrate the greenness, dryness and imperviousness, moisture, heat, and air turbidity. The new model was evaluated by comparing with land cover types and the existing remote sensing-based ecological index and eco-environmental quality index models. Results demonstrated that RSUAE is valid to depict the difference of eco-environment quality among varied land cover types. The RSUAE model has general consistency with the existing models, while RSUAE takes into account well the air quality. The RSUAE was utilized to evaluate the eco-environment of UA along the Yellow River in Ningxia, China (NXUA) with Mann–Kendall trend test. Results identified 78.69% of the area has no significant change trend, 18.73% of the area has a significant increasing trend, and 2.58% of the area suffers significant decreasing trend in eco-environment quality from 2001 to 2019. The Yanchi county has the best improvement in eco-environment quality, whereas sensitivity analysis indicates that it is more vulnerable than other counties. The analysis tool and method presented in this study provide a reference for other UAs. The evaluation results in NXUA are noteworthy for local management.

The Orbita hyperspectral satellite (OHS) is the first commercial hyperspectral satellite in China that completed launching and networking. It can collect world-class hyperspectral data and obtain aerial hyperspectral imagery with 32 bands covering the spectrum range from 400 to 1000 nm at a 10-m resolution, which are of great significance for the quantitative analyses of remote sensing and refined recognitions of land covers. We explore the potentiality of the OHS image in land cover classification (LCC). Taking the Pearl River Delta region as the study area, we selected five feature indices from OHS data, i.e., original bands (OBs), vegetation indices (VIs), water indices (WIs), red edge indices (REIs), and gray-level co-occurrence matrix (GLCM) textures for the LCC. Then, data combination schemes were intended to analyze and compare the performance of different feature indices on the accuracies of the LCC. Last, feature optimization was performed on all input variables to determine the optimal variables combination to increase the accuracy and efficiency of the LCC. The random forest classifier was adopted in the above schemes, and the method of mean decrease in accuracy was used to determine the importance of the variables. The results show that (1) refined accuracy of LCC was obtained using only OBs; in addition, REIs can further improve the classification accuracy significantly. (2) The optimal variables combination achieves the highest classification accuracy (OA  =  93.21  %   and kappa coefficient  =  0.91), and the user’s and producer’s accuracies exceeded 90% for most land cover categories. (3) Variable importance analyses show that the importance of both red-edge bands and REIs were greater than those of near-infrared bands and VIs for the LCC. The importance ranking of different indices from high to low was REIs > OBs > GLCM > WIs > VIs. This research demonstrates the potentialities and values of the OHS image for the application of LCC.

Terra and Aqua MODIS have successfully operated for more than 20 and 18 years, respectively, and far exceeded their designed lifetimes of 6 years. MODIS has 36 spectral bands, among which 16 are the thermal emissive bands (TEB) covering a wavelength range from 3.75 to 14.24  μm. Observations from the MODIS TEB have been used to generate a number of data products, such as surface/cloud/atmospheric temperatures, cloud top altitude, and water vapor properties. To a large extent, the quality of the MODIS L1B radiance product has been maintained through the entirety of each mission via extensive on-orbit calibration and validation efforts. MODIS TEB calibration uses a quadratic algorithm and is referenced to an on-board blackbody with its temperature measurements traceable to the NIST standard. We provide an overview of the MODIS instrument operations, key TEB calibration activities, and algorithms used in the latest L1B data Collection 6.1 (C6.1) and describe the TEB on-orbit performance for both Terra and Aqua MODIS. The TEB gain trends have been stable to within 5% over the Terra mission after 2003 under the same electronic configuration, except for bands 27 to 30, and within 3% over the Aqua mission. The Terra MODIS optical crosstalk correction implemented right after launch has been very effective in removing the ghost images in its photoconductive detector bands 32 to 36. Starting from C6.1, an electronic crosstalk correction algorithm has also been applied to the Terra MODIS photovoltaic detector long-wave infrared bands 27 to 30. The noise characterization performance remains stable with most TEB detectors continuously meeting their design requirements. A gradual loss of the Aqua MODIS radiative cooler margin reached its maximum in 2013 and has been slowly recovering since then. On-orbit changes in TEB response versus scan angle are extremely small based on pitch maneuver observations and assessments using vicarious approaches. The remaining challenges in TEB calibration and proposed improvements for the upcoming data collection with L1B data production and reprocessing are discussed.

Detecting and monitoring surface water has received much attention in recent decades. Surface water is one of the most critical water resources for both human and ecological systems. Remote sensing technology has made it possible to have accurate and frequent updates of surface water. We propose a remote sensing multisensor fusion system using optical data including Landsat-8 and Sentinel-2 and RADAR data including Sentinel-1 for water body extraction. Using a data fusion approach, the spatial resolution of multispectral images increased from 30 to 10 m, and the spectral information of Landsat-8 and Sentinel-2 were preserved. Then, all features extracted from the high spatial resolution images, water index maps, and Sentinel-1 dataset were combined as a stacked feature space. The new data were subsequently classified by support vector machine, neural network, and random forest. Finally, all classifiers’ outputs were integrated using a weighted majority voting decision fusion strategy, which significantly improved surface water extraction. Our study illustrates the ability of remote sensing multisensory fusion, water indices, and decision fusion for water body extraction.

Windthrows induced by strong winds pose a major threat to both transport infrastructure and road users. Therefore, an exposure analysis of trees along the federal trunk road network was carried out and exemplarily applied for the federal state of North Rhine-Westphalia, Germany. The aim of this project was the development of a GIS-based method on the basis of freely accessible high-resolution airborne LiDAR data as well as RGBI orthoimage data that identifies and parameterizes single trees. For the determination of a suitable method, several models with different settings and parameters were calculated, validated, and iteratively adjusted to gain the optimal settings for tree identification. The identification of trees was finally realized by applying both a minimum-curvature technique for single trees as well as a local maximum approach for dense vegetation, especially for trees with small crown diameters. False classification results corresponding to nonvegetation areas have been corrected by the use of the normalized difference vegetation index. With our method, we accomplish detection accuracies from 65% to 75% in very heterogeneous environments and 75% to 100% in more specific settings. All tree candidates with potential to affect the road infrastructure and road users were retrieved and joined to the respective road sections to deliver a fast and effective method for analyzation and visualization of vulnerable parts of the trunk road network due to tumbling trees.

The global interaction of CO₂ flux was highly dynamic under seasonal and inter-annual variability. Thus, precise estimation of gross primary productivity (GPP) and net photosynthesis (PsnNet) under seasonal variations was important to understand the global carbon cycle and ecosystem response to climate variability. Considering this significance, in our paper, we have conducted a comparative investigation of correlation (i.e., Pearson product, Spearman rank, and Kendall rank) and supervised machine learning models (i.e., random forest, conditional inference forests, and quantile regression forests) to explore the agreement and prediction of GPP and PsnNet under seasonal variability from spectral indices, tasseled cap transformations, and reflectances over a mixed ecosystem assembling moderate resolution imaging spectroradiometer and Landsat data. Associated uncertainties and errors have been also compared. Our study was carried out in a tropical moist ecosystem under pre-monsoon (March) and post-monsoon (October) conditions. Overall outcome revealed that seasonal differentiation does not affect the simple correlations; however, it has a significant impact on the GPP/PsnNet prediction process. Near-infrared (NIR)-based spectral indices have been found with the best agreements and statistically significant (p  <  0.001) for GPP/PsnNet during March and October ranging coefficients from 0.85 to 0.44 in all three correlation models. However, Landsat 8 OLI band reflectances in different regions of the electromagnetic spectrum have gained importance as best predictors during March, whereas NIR-based indices have been found as best predictors for GPP/PsnNet during October. We have found that anomalies in environmental and meteorological conditions have a strong impact on plant functional activities that significantly differs prediction process rather than simple correlations.

Land cover mapping using remote sensing optical images with a very high spatial resolution (VHSR) plays an important role in observing the Earth’s surface. However, classification maps are usually affected by salt-and-pepper noise because VHSR optical images usually have a low resolution of ground targets in terms of spectral reflectance. An adaptive region-based post-classification framework (ARPF) is proposed for improving the initial classification map while using a VHSR optical image to further smooth the noise of initial classified maps. First, different from several traditional methods using a single classifier, our proposed ARPF needs more than four different initial classification maps acquired from different classifiers or image features. Second, an adaptive region around each pixel of the gray image is generated with two predefined parameters, and each adaptive region is applied to refine the corresponding pixel of each initial classified map. Finally, all the refined classified maps are merged to obtain the final classification map by coupling adaptive region and majority voting rules. In our experiments, three optical images with VHSR are used to evaluate the proposed ARPF. Compared with three typical relevant post-classification methods, the proposed ARPF can provide a classification map with less noise in visual performance and achieve higher quantitative accuracy while having an advantage in the constant detail of ground targets.

Sparse unmixing (SU) can represent an observed image using pure spectral signatures and corresponding fractional abundance from a large spectral library and is an important technique in hyperspectral unmixing. However, the existing SU algorithms mainly exploit spatial information from a fixed neighborhood system, which is not sufficient. To solve this problem, we propose a nonlocal weighted SU algorithm based on global search (G-NLWSU). By exploring the nonlocal similarity of the hyperspectral image, the weights of pixels are calculated to form a matrix to weight the abundance matrix. Specifically, G-NLWSU first searches for a similar group of each pixel in the global scope then uses singular value decomposition to denoise and finally obtains the weight matrix by considering correlations between similar pixels. To reduce the execution burden of G-NLWSU, we propose a parallel computing version of G-NLWSU, named PG-NLWSU, which integrates compute unified device architecture-based parallel computing into G-NLWSU to accelerate the search for groups of nonlocally similar pixels. Our proposed algorithms shed new light on SU by considering a new exploitation process of spatial information and parallel computing scenario. Experimental results conducted on simulated and real datasets show that PG-NLWSU is superior to comparison algorithms.

The imaging parameters of high-squint synthetic aperture radar (SAR) mounted on maneuvering platforms have obvious spatial variability, which cannot be effectively solved by traditional SAR imaging algorithms and limits the focus depth. To extend the focus depth of maneuvering SAR, an imaging method is proposed based on perturbation keystone transform (KT) and subregion phase filtering (SRPF). In the process of range cell migration correction, the azimuth-dependent range curvature is corrected by introducing the perturbation factor in two-dimensional (2-D) frequency domain, and then the linear range cell migration is corrected by conducting the azimuth KT. In the process of azimuth compression, after removing the negative influence generated by the 2-D frequency-domain perturbation, the precise azimuth time-domain phase is reconstructed. And then, the method of SRPF is proposed to extend the azimuth focus depth. This method divides the range cell into several subregions and carries out the phase filtering and deramp process in each subregion. Theoretical analysis and simulation results show that the proposed method has high efficiency and can effectively expand the focus depth of high-squint SAR mounted on maneuvering platforms.

We address the covariance matrix estimation problem for radar adaptive detection in a non-Gaussian clutter environment. We first propose an estimation method based on α log-determinant divergence, which estimates the true covariance accurately by solving the geometric mean of the sample covariance matrix (SCM). Since the estimation performance would be seriously degraded when the number of secondary data is insufficient, a knowledge-aided method is then proposed. Under the similarity constraint between the a priori covariance and the true one, a closed form expression is derived by minimizing the α log-determinant divergence between the real covariance and the SCM. Simulation results verify the accuracy of the proposed algorithms in covariance estimation and superiority in target adaptive detection.

Deep learning techniques have become increasingly popular for classifying large-scale image and video data. Remote sensing applications require robust search engines to retrieve similar information dependent on an example-based query instead of a tag-based query. Deep features can be extracted automatically by training raw data without having any domain-specific knowledge. However, the training time for a massive amount of multimedia datasets is high. Training complexity is reduced using pre-trained GoogleNet weights for initial feature extraction. To fine-tune the feature vector and reduce the dimensionality, a one dimension convolutional neural network (1D-CNN) is applied. There is a loss of information while resizing the input image to a pre-trained network with an acceptable input size. We proposed a new feature set by integrating handcrafted features at detailed scales and deep features to improve the system’s efficiency. The curvelet transform was used to decompose the image into coarse and detailed scales. Haralick texture features were extracted from the detail coefficients in four directions and fused with fine-tuned deep features. The proposed feature set was assessed using standard performance metrics from the literature. The proposed technique achieved improved performance with 89% accuracy for retrieval of the top 50 relevant results.

Marine floating raft aquaculture (FRA) monitoring is vital for environment protection and mariculture management. Synthetic aperture radar (SAR) could provide high-quality remote sensing images under all weather conditions compared with the existing optical remote-sensing-based methods. Traditional SAR monitoring methods extract the pixel feature of marine FRA in single patches, which commonly leads to poor generalization. We propose a self-attention semantic segmentation method based on modified U-Net++ (SA-U-Net++) for FRA segmentation, which could automatically extract semantic feature information, and provide superior performance under complicated scenes. The proposed self-attention backbone could help to extract more precise features and enhance the overall accuracy. Furthermore, we propose a FRA-ISAR data generation method based on inverse SAR (ISAR) imaging to alleviate the sample shortage problem. We introduce the semantic segmentation method into FRA-SAR segmentation for the first time. The experiments verify the effectiveness and superiority of SAR FRA segmentation based on the proposed SA-U-Net++ model compared with the existed semantic segmentation approaches.

The speed of the image clustering method based on the spectral clustering algorithm is greatly affected by the size of image resolution; in many cases, the processing of high-resolution images cannot be completed precisely on time. Thus, based on the connected area labeling and superpixel spectral clustering algorithm, this paper proposes a fast change detection method for remote sensing images based on the connected area labeling method and spectral clustering algorithm. In the initial stage, the method adopts the idea of non-local (NL)-means algorithm to generate NL difference image, then, through the OTSU method, several critical areas with large connected areas are discriminated according to a threshold, and the rectangular contour internal images of these areas are extracted and processed by the superpixel spectral clustering algorithm, to realize the change detection of the areas quickly. The results of the application experiments and the performance experiments proved that the CASC method could solve the feasibility of spectral clustering algorithm to deal with high-resolution remote sensing images to a certain extent, has good robustness, and the processing speed has been dramatically improved. The proposed method can quickly and efficiently extract sensitive sub-areas with significant changes in the study area and provide basis and decision support for geological monitoring and research in critical areas in the future.

Bistatic radar target tracking is challenging due to the fact that the measurements are nonlinear functions of the Cartesian state. The converted measurement Kalman filter (CMKF) converts the raw measurement into Cartesian coordinates prior to tracking and is superior to the extended Kalman filter for certain problems. The challenges of CMKF are debiasing the converted measurement and approximating the converted measurement error covariance. Due to no closed form of biases, we utilize the second-order Taylor series expansion of the conventional measurement conversion to find the conversion bias in bistatic radar and propose the unbiased converted measurement (UCM). In order to decorrelate the converted measurement error covariance from the measurement noise, we evaluate the covariance using the prediction in Bayesian recursive filtering, designated as the decorrelated unbiased converted measurement (DUCM). Monte Carlo simulations show that the DUCM is unbiased and consistent, and the DUCM filter exhibits an improved performance compared with the conventional CMKF and the UCM filter in bistatic radar tracking.

The accurate segmentation of remotely sensed hyperspectral images has widespread attention in the Earth observation and remote sensing communities. In the past decade, most of the efforts focus on the development of different supervised methods for hyperspectral image classification. Recently, the computer vision community is developing unsupervised methods that can adapt to new conditions without leveraging expensive supervision. In general, among unsupervised classification methods, sparse subspace clustering (SSC) is a popular tool that achieves good clustering results on experiments with real data. However, for the specific case of hyperspectral clustering, the SSC model does not take into account the spatial information of such images, which limits its discrimination capability and hampering the spatial homogeneity of the clustering results. As a solution, we propose to incorporate a regularization term to the SSC model, which takes into account the neighboring spatial information of spectral pixels in the scene. Specifically, the proposed method uses a three-dimensionall (3D) Gaussian filter to perform a 3D convolution on the sparse coefficients, obtaining a piecewise-smooth representation matrix that enforces an averaging constraint in the SSC optimization program. Extensive simulations demonstrate the effectiveness of the proposed method, achieving an overall accuracy of up to 99% in the selected hyperspectral remote sensing datasets.

Airborne LiDAR has become a kind of indispensable measurement device in the current field of remote sensing. However, target extraction using traditional airborne LiDAR based on single-point scanning requires filtering and point cloud segmentation operations, which are complicated and time consuming. Although some researchers have studied streak tube imaging LiDAR (STIL) before, there are few reports in which it is used as an airborne LiDAR for ground measurement in large-scale field. We propose a method of ground target extraction using STIL. Taking advantage of the structural properties of the STIL, complex filtering and point cloud segmentation algorithms are avoided in the target extraction method. The purpose of this article is to verify the feasibility of airborne STIL in ground target extraction. We analyzed the raw streak signal image collected by field experiment and used morphology and intensity information to extract features. After that, we employed the decision tree classifier to classify the four kinds of targets and evaluated the extraction results. The results show that the target extraction achieved satisfactory consequences under an acceptable level. That demonstrates that ground target extraction using STIL is feasible in the field of large-scale remote sensing.

Hyperspectral images (HSIs) are the spectral images that provide spatial and spectral information. Unlike multispectral images, these images consist of 100 to 200 bands, which provide a large amount of data to identify minute details of the scene with the help of the spectral signatures. This information is valuable and should be secured while transmitting the HSI over the network. Visual cryptography is a well-known cryptographic method for securing images. It helps in securely transmitting images among n users. It converts visual data into unreadable shares that are transmitted and, stacking these shares together will reveal the image. Many visual secret sharing (VSS) schemes have been proposed in the past decade, which made it easier to hide the visual information from unauthorized users. We have proposed a verifiable XOR-based VSS method for the secure transmission of HSIs. We have introduced a preprocessing step for the image with a band selection technique, which reduces the size of the image and eliminates redundancy. It also includes the detection of the tampered shares. Embedding a verifiable bit in each pixel is performed to test the integrity of the shares. We have assessed the visual quality of the recovered image using quantitative measuring parameters. We also compared them with existing VSS methods. The proposed method recovers the HSI with a self-similarity index of 95% to 99%. The proposed method’s experimental results show that the HSI is restored with better visual quality.

Recently, with the number of large-scale remote sensing (RS) images increasing, the demand for large-scale RS image object classification is growing, and many researchers are interested. Hashing, as a result of its low memory requirements and high time efficiency, has widely solved the problem of large-scale RS images. Supervised hashing methods mainly leverage RS image label information to learn hashing function; however, the similarity of the original feature space cannot be well preserved, which cannot meet the accurate requirements of RS images object classification. To address the aforementioned analysis, we propose a method named optimized projection supervised discrete hashing (OPSDH), which jointly learns optimized projection constraint and discrete binary codes generation model. It uses an effective optimized projection method to further constraint on the supervised hashing learn, and generated hash codes preserve the similarity based on the data label while retaining the original feature space’s similarity. The experimental results show that OPSDH reaches improved performance compared with existing hashing methods and demonstrates that the proposed OPSDH is more efficient for operational applications.

Hyperspectral image (HSI) classification is a challenging problem due to the high dimensional features, high intra-class variance, and limited prior information, and the classification is the basis for HSI applications. Active learning (AL) and semisupervised learning (SSL) are two promising approaches in the HSI classification. In AL, the traditional entropy query-by-bagging (EQB) algorithm only pays attention on uncertainty and ignore the diversity among the samples. Therefore, we propose averaged normalized entropy query-by-bagging (anEQB) algorithm. Meanwhile, the collaborative active learning and semisupervised learning framework (CASSL) may invoke many wrong pseudolabels and deteriorate the classification performance. To make up for the deficiency of CASSL, we complement different AL algorithms to constitute a multiple filtering mode semisupervised learning framework (MFMSLF). To further study, we introduce syncretic secondary filtering mode into multiple verification semisupervised framework and thus constitute a multiple secondary filtering mode semisupervised verification framework (MSFMSVF). We evaluate the performance of anEQB, MFMSLF, and MSFMSVF on different hyperspectral data sets and compare them with other state-of-the-art HSI classification methods. Numerical experimental results reveal the superior classification performance of anEQB, MFMSLF, and MSFMSVF, respectively. Experimental results also demonstrate that exploring the information and diversity of the samples from different criterion can improve the classification performance of the collaborative framework.

The existing methods of remote sensing image super-resolution reconstruction based on deep learning have some problems, such as insufficient feature extraction abilities, blurred image edges, and difficult model training. To solve these problems, a super-resolution reconstruction method combining residual channel attention (CA) is proposed. Based on the framework of generative adversarial networks, the residual structure is designed to enhance the ability of deep feature extraction ability for remote sensing images. The CA module is added to extract the deep feature information of remote sensing images, and the shallow features and deep features are fused using the skip connection. The perceptual loss function is combined with the loss function represented by the Wasserstein distance to improve the stability of model training. The experimental results show that this method is superior to the comparison algorithms in the objective evaluation criteria of the peak-signal-to-noise ratio and structural similarity of the reconstructed remote sensing images. After optimizing the model training process, the reconstructed remote sensing images are visually clearer and have sharper edges.

For an airborne passive radar with impure reference signals, the clutter caused by multipath (MP) signals involved in the reference channel (MP clutter) corrupts the space–time adaptive processing performances. To eliminate the influence of the MP clutter, two clutter suppression methods based on reduced-dimension (RD) transformation are proposed herein. RD transformation is exploited to reduce the size of the sparse recovery dictionary. Subsequently, the sparse recovery problem is revised, and the MP clutter is suppressed using the least mean square (LMS) algorithm and the exponentially forgetting window LMS algorithm. Compared with the existing L₁-based recursive least square algorithm, the proposed algorithms significantly reduce computational complexity without degrading the MP clutter suppression performance. In addition, the proposed algorithms provide more robust characteristics to the errors in prior knowledge than the modified blind equalization method. A range of simulations is conducted to test the proposed algorithms.

The classification of high-resolution (HR) synthetic aperture radar (SAR) image is of great significance to SAR scene interpretation and understanding. However, the HR SAR image contains rich ground information and complex spatial structural features. The extraction of effective distinguishing features under the influence of coherent speckle is still a challenging task. A multi-scale anisotropic convolution sparse coding (MACSC) algorithm is proposed for HR SAR image classification. First, the low-order statistical features are extracted to improve the richness of SAR image. Then an unsupervised MACSC model is utilized to learn a set of sparse feature maps and convolution filters for the above statistical features by solving the designed objective function. In MACSC, the multi-scale anisotropic Gaussian kernels are utilized to initialize the dictionary of MACSC. Compared with the isotropic Gaussian kernel of the original CSC, these kernels can more accurately describe the details of different directions and scales in the SAR scene. Meanwhile, the adaptive sparse control factor is introduced into the MACSC model, which can make the learned sparse feature maps suppress the speckle and capture the abundant edge and texture information of the SAR image. After that, the aggregation operation is conducted on multi-scale and multi-direction sparse feature maps to integrate the neighbor pixels and reduce the computing burden. Finally, the obtained feature vector is input into the support vector machine classifier to realize classification. Experiments on three challenging SAR data sets demonstrate that MACSC achieves better classification performance than related sparse representation methods.

Optical remote sensing images are frequently affected by clouds, haze, and mist in the atmosphere. We introduce an iterative minimization light-cloud removal method designed for the specific quality improvement needs of military reconnaissance panchromatic remote sensing images. The proposed method is required to fulfill the military reconnaissance demands for improvements in the quality of panchromatic high-resolution images while guaranteeing high fidelity between the restored and observed images. A heuristic approach based on contrast enhancement is proposed to solve the thin-cloud removal problem. We design the target function of a minimization algorithm that contains a fidelity term, a contrast penalty term, and an information loss penalty term. By minimizing the target function with the iterative steepest descent method, a high-quality image can be restored from the observed satellite cloudy image, and the details are preserved by the penalty terms. The application of our iterative method to Gaofen-1 (GF-1) and Ziyuan-3 (ZY-3) satellite data shows that the iterative method was applicable to GF-1 and ZY-3 satellite and the data showed that for panchromatic remote sensing images, the proposed method could reduce satellite image degradation caused by haze and thin clouds while preserving the details in the observed images.

Recent automated crop mapping via supervised learning-based methods have demonstrated unprecedented improvement over classical techniques. Classification accuracies of these methods degrade considerably in cross-year mapping. Cross-year crop mapping is more useful as it allows the prediction of the following years’ crop maps using previously labeled data. We propose vector dynamic time warping (VDTW), an innovative multi-year classification approach based on warping of angular distances between phenological vectors. The results prove that the proposed VDTW method is robust to temporal and spectral variations compensating for different farming practices, climate and atmospheric effects, and measurement errors between years. We also describe a method for determining the most discriminative time window that allows high classification accuracies with limited data. We carried out tests with Landsat 8 time-series imagery from years 2013 to 2015 for the classification of corn and cotton in the Harran Plain of southeastern Turkey. In addition, we tested VDTW corn and soybean in Kansas, the US for 2017 and 2018 with the Harmonized Landsat Sentinel data. The VDTW method improved cross-year overall accuracies by 3% with fewer training samples compared to other state-of-the-art approaches.

Bark beetle outbreaks are a significant cause of loss of vegetation cover, for which accurate monitoring of forest areas is required to detect and control bark beetle outbreaks as early as possible. A tool that processes aerial imagery from an unmanned aerial vehicle to automatically detect levels of damage caused by bark beetle outbreaks is proposed and evaluated. The true-color RGB flight imagery is combined into orthomosaics, enhanced, and then analyzed in reference to manually annotated samples to identify thresholding rules for training classifiers based on a cellular automaton that assigns to each nonbackground pixel in the image a class label corresponding to the estimated stage of infestation at the location—healthy (green), early-stage (yellow), late-stage (red), and dead (leafless). Also samples corresponding to the ground (nontrees) are annotated and processed. The resulting classifications are on average over 89% accurate over five flights and often near flawless; the view from above does not fully substitute a ground-based assessment for intermediate stages of the infestation.

Stellar map denoising and centroid positioning, which directly determine the postpositioning accuracy of star trackers, are key technologies in stellar map processing. Due to the influence of a complex starry sky background, there is often a large amount of noise in stellar maps, which makes it difficult to accurately locate the stellar centroid. A stellar map processing method based on dark channel denoising and continuous multiframe stellar map centroid positioning combined with centroid trajectory constraints is proposed. First, a dark channel noise template is used for denoising, and the single-point and multipoint noises in the denoising result are filtered. Second, in the process of stellar map positioning, if the maximum gray value of the stellar is not unique, it is constrained by the previous stellar positioning result, an adaptive window is established, and the gray-scale centroid weighting method is used to locate the centroid. Then, the star angular distance is used to analyze the precision of the centroid positioning. Finally, the jitter frequency of the satellite platform based on the continuous multiframe centroid positioning result is used to detect satellite attitude. The experimental results show that the performance of dark channel denoising, which can solve many strip noise and background noise problems in stellar maps, is better than that of the existing stellar map denoising method. The centroid positioning results improves the star angular distance by 18.85 arc sec compared with the Gaussian filter and by 8.03 arc sec compared with the global threshold segmentation method, significantly improving the accuracy of stellar map centroid positioning and laying a foundation for improving the star tracker positioning accuracy. A jitter frequency of 0.67 Hz was detected on the ZiYuan3 satellite platform based on the centroid positioning results. It enriches the jitter detection methods and provides a theoretical and technical basis for the design of future high-resolution ground observation remote sensing satellite platforms and geometric accuracy compensation.

Semantic segmentation is an important and foundational task in the application of high-resolution remote sensing images (HRRSIs). However, HRRSIs feature large differences within categories and minor variances across categories, posing a significant challenge to the high-accuracy semantic segmentation of HRRSIs. To address this issue and obtain powerful feature expressiveness, a deep conditional generative adversarial network (DCGAN), integrating fully convolutional DenseNet (FC-DenseNet) and Pix2pix, is proposed. The DCGAN is composed of a generator–discriminator pair, which is built on a modified downsampling unit of FC-DenseNet. The proposed method possesses strong feature expression ability because of its skip connections, the very deep network structure and multiscale supervision introduced by FC-DenseNet, and the supervision from the discriminator. Experiments on a Deep Globe Land Cover dataset demonstrate the feasibility and effectiveness of this approach for the semantic segmentation of HRRSIs. The results also reveal that our method can mitigate the influence of class imbalance. Our approach for precise semantic segmentation can effectively facilitate the application of HRRSIs.

Aiming at optimizing the allocation problem of limited resources in a radar network, a resource scheduling algorithm combining pulse interleaving with preallocation is proposed for multitarget inverse synthetic aperture radar imaging. The imaging method adopts compressed sensing, which only needs to emit a small number of pulses so that we can set the algorithms to schedule the allocation of the pulses over a period of time. The authors point out the problem of pulse conflict, which is ignored in the process of the scheduling algorithm and proposes a preallocation method to avoid the occurrence of the conflict. Meanwhile, pulse interleaving is added to increase the positions of the dispatchable pulses. Moreover, the combined algorithm can perform adaptive scheduling on the radar time resource according to the feature parameters after target feature cognitive. Finally, the feasibility of the combined algorithm is verified by the simulation, and two performance indicators, the hit value rate and the pulse utilization rate, are improved by the proposed algorithm.

Space-based light detection and ranging (LiDAR) sensors have provided valuable insight into the global, vertical distribution of aerosol and cloud layers in Earth’s atmosphere, and, more recently, of the distribution of phytoplankton in the ocean. However, the photodetectors in these sensors lack the performance necessary to capture the vertical structure of cloud tops and ocean phytoplankton to a fidelity sufficient for advancing our understanding of the global water cycle and ocean carbon cycle, respectively. Recent advancements in high-performance single photon avalanche diode (SPAD) arrays promise to enable these measurements, while also offering a sensitivity that will allow significant reductions in laser power and telescope size, with associated sensor-level size, weight, and power (SWaP) savings. To harness the unique benefits of SPADs for these measurements, we propose to develop a large-format array of photon counting SPADs with <10 ns dead time, along with readout integrated circuitry that sums and bins (histograms) photon counts in real time to the desired temporal resolution for the target application. The feasibility of this approach has been investigated with a small-scale 8  ×  8 SPAD array proof of concept hardware demonstration developed at Politecnico di Milano, with promising initial results. Progress is reported on designs that will allow scaling the array and readout integrated circuit electronics to the requisite of 128  ×  128 size in a chip-scale, low power, photodetector ideal for LiDAR remote sensing of the atmosphere and ocean from SWaP-constrained platforms.

A mobile system has been developed to be used to detect buried objects (and/or explosives) containing ferromagnetic material by magnetic anomaly method. Detection, in this system, can be performed without broadcasting any signal to the external environment, and a visual and audible alert can be sent to the user. A sensor network containing nine fluxgate sensors has been created; one of them was used as a reference sensor and the others as the measurement sensors. Data obtained due to the voltage variations at the output of fluxgate sensors are processed with an Atmega 2560 microcontroller. Images of buried objects are created using this data and displayed on a human–computer interface panel. Experimental studies have been carried out for eight test samples buried underground in different soil types, and the average classification success has been obtained as 91.75%.

The Advanced Baseline Imager (ABI) aboard Geostationary Operational Environmental Satellite (GOES)-16 and -17 satellites represent the next-generation geostationary multispectral imaging instrument. Since GOES-16 ABI imagery data became available, stray light was observed in ABI visible, near-infrared (VNIR), and 3.9  μm, i.e., CH07, channels. A stray-light characterization scheme was developed to quantitatively monitor stray-light variation in ABI imagery. The stray-light analysis is focused on ABI CH07, whose nighttime radiometric performance being impacted by stray light is of main concern. It is found that the stray light in the ABI imagery occurs over ∼3 months around spring and fall equinox each year. The maximum stray light of GOES-16 ABI CH07 is ∼0.65  K at 300 K scene in zone of normal performance (ZONP), i.e., region with relative solar angle >7.5  deg, which is within the radiometric requirement of 1 K. The analysis of Himawari-8 Advanced Himawari Imager (AHI) CH07 data indicates that its maximum stray light is ∼3.35  K at 300 K scene in ZONP, much higher than that of GOES-16. This confirms the effectiveness of reducing the major stray-light leaking path in ABI as a result of lessons learned from Himawari-8 AHI. The magnitude of GOES-17 ABI CH07 stray light is shown to be ∼0.45  K in ZONP, slightly lower than GOES-16. The analysis of AHI 3.9 and 6.2  μm channel stray-light radiance ratio is shown to match the ratio of solar irradiance spectrum, which suggests that the stray-light issue is due to direct leakage of solar radiation through the instrument. Further characterization and monitoring of GOES-16 ABI VNIR channel stray light also help understand the solar origin of ABI stray light. This paper also investigated cases of strong atmospheric refraction-induced stray-light contamination onto ABI detectors during solar eclipse seasons and presents scheme to reduce such potentially harmful contamination.

Control of orbital angular momentum (OAM) offers the potential for increases in control and sensitivity for high-performance microwave systems. EM waves with properties dependent on spatial distribution are said to be “structured.” Control of OAM in microwave systems is an example of a wave structure that exploits EM degrees of freedom, which most conventional systems do not use. OAM is characterized by an integer OAM mode in which zero represents the case of a plane wave and nonzero OAM modes propagate with a helical wavefront. A uniform circular phased array approach is utilized to produce helix-shaped OAM wavefronts. This method offers a critical advantage over common fixed-frequency dielectric lenses; the phases of all antenna elements are programmable across a wide frequency range, which is necessary for ultrawideband (UWB) radar imaging. Simulations and laboratory experiments are performed to determine the requirements and capabilities of an UWB OAM radar that uses the circular phased array approach. Key results include OAM phase front characterization, detection of specified OAM modes, and configuration of a network analyzer UWB radar with synthetic OAM mode-control via signal post-processing.

Hedgerows are one of the few remaining natural landscape features within European agricultural areas. To facilitate hedgerow monitoring, cost-effective and accurate mapping of hedgerows across large spatial scales is required. Current methods used for automatic hedgerow detection are overly complicated and generalize poorly to larger areas. We examine the application of transfer learning using two neural networks (Mask R-CNN and DeepLab v3+) for hedgerow mapping in south-eastern Germany using IKONOS imagery. We demonstrate the potential of such networks for hedgerow monitoring by investigating performances across varying input image bands, seasonal imagery, and image augmentation strategies. Both networks successfully detected hedgerows across a large spatial scale (562  km²), with DeepLab v3+ (75% F1-score) outperforming Mask R-CNN. Differences between band combinations were minimal, implying hedgerow detection could be achieved using RGB sensors. Results suggested that using all available training images across seasons is preferred and should have the same model generalizing effects as data augmentation. Experiments with varying data augmentations found augmentations effecting object geometries to greatly increase performance for both networks while results using augmentations modifying pixel spectral values showed concerning effects. Overall, our study finds that transfer learning in neural networks offers a simplified approach that outperforms previously established methods.

Aircraft detection in remote-sensing images is a fundamental task in civil and military applications. Deep learning techniques to achieve end-to-end object detection have attracted the attention of the Earth observation community. One of the primary factors behind the success of deep learning techniques is the utilized data. Several previous studies focused on designing the network infrastructure. Instead, this study pays more attention to the data. With the increasing number of available public datasets, whether directly employing a large number of instances with great variation will lead to a good performance has become a research topic. The ways in which these object instances are collected differ greatly. For example, the image sizes, object sizes in the training images, and geospatial resolution are varied. Therefore, herein, the factors influencing the detection performance, such as the object size, ground sampling distance, and Google zoom view, are investigated. A you-only-look-once-v3-based detection process is proposed for automatic aircraft detection. A nonmaximum suppression algorithm strategy is applied to filter unreliable and redundant bounding boxes detected in the overlapping image blocks. The model generalization ability under different training data combinations is evaluated in several challenging cases. The results prove that more variety of training instances from a greater variety of zoom levels will result in more false alarms. Instead, more variety in the object sizes under a constant zoom level is welcome. A large range of aircraft sizes (i.e., 7 to 77 m in length in this study) can be detected, with a promising F1 score of 0.98.

Estimating volumetric soil moisture (M_v) and surface roughness (S) are the key parameters for numerous agricultural and hydrological applications. Although these two parameters can be effectively retrieved from synthetic aperture radar (SAR) data, the presence of vegetation can negatively affect the results. A method was proposed to accurately estimate M_v and S over vegetated agricultural areas. The method was based on applying the machine learning inversion approach along with SAR data to invert a combination of the parameterized water cloud model (PWCM) and the calibrated integral equation model (CIEM). The soil backscattered component in water cloud model (WCM) was generated by CIEM to be applied to the WCM parameterization and dataset simulation. Three machine learning algorithms, including the support vector regression (SVR), multi-output SVR (MSVR), and artificial neural network (ANN), were employed to model the relationship between the simulated dataset variables. The genetic algorithm was also applied to optimize the models’ parameters. The inversion technique results demonstrated that the MSVR and ANN had the highest accuracy in estimating M_v and S due to their better structures. The SMAPVEX-16 in situ dataset, along with three Sentinel-1 images, was applied to evaluate the accuracy of the WCM parameterization and the proposed method for M_v and S estimation. The accuracies of the PWCM in the VV and VH polarizations of Sentinel-1 C-band data were reasonable for VWC  <  2.5  kg  /  m² [root-mean-square error (RMSE) = 1.44 and 1.77 dB, respectively]. Additionally, it was observed that the trained SVR, MSVR, and ANN had similar results for different VWC values. In summary, the proposed method had high potential in vegetated agricultural areas with VWC  <  2.5  kg  /  m², for which the RMSEs were 4 to 7 vol. % and 0.35 to 0.46 cm depending on the VWC values in retrieving M_v and S, respectively.

Efficient and accurate extraction of water areas from remote sensing images is a popular research topic. Currently, researchers have attempted to use neural networks to extract water from remote sensing images. However, most of these studies used computer vision techniques to improve the model results without considering the multi-band information unique to remote sensing images. Thus our study proposes an improved DeepLabv3+ network to increase the water body extraction accuracy in urban remote sensing images. The DeepLabv3+ network has the characteristics of extracting image features at multiple scales. We improved the network structure to incorporate multi-band features. By comparing several multi-band input methods, the feature map calculated by the normalized difference water index (NDWI) was converted into an input suitable for the neural network by comparing several multi-band input methods. Simultaneously, we developed a parallel convolution structure to combine the NDWI feature map with a standard false color remote sensing image during feature extraction. This allows the network to focus more on image areas that may be water bodies. We used atmospherically corrected Sentinel-2A L2A-level data, divided the training set at multiple scales, and conducted several experiments. The results show that the proposed network can improve water extraction accuracy when training subregions are unified from different sizes to 512  ×  512. Finally, we used the model to extract water bodies from remote sensing images from different regions. We combined the images with visual interpretation to verify the reliability of the model results. Moreover, the model scores of four types of multi-scale neural networks in the two categories are compared, which proves the effectiveness of the method.