Accurate and spatially detailed mapping of complex urban environments is essential for land managers. Classifying high spectral and spatial resolution hyperspectral images is a challenging task because of its data abundance and computational complexity. Approaches with a combination of spectral and spatial information in a single classification framework have attracted special attention because of their potential to improve the classification accuracy. We extracted multiple features from spectral and spatial domains of hyperspectral images and evaluated them with two supervised classification algorithms; support vector machines (SVM) and an artificial neural network. The spatial features considered are produced by a gray level co-occurrence matrix and extended multiattribute profiles. All of these features were stacked, and the most informative features were selected using a genetic algorithm-based SVM. After selecting the most informative features, the classification model was integrated with a segmentation map derived using a hidden Markov random field. We tested the proposed method on a real application of a hyperspectral image acquired from AisaFENIX and on widely used hyperspectral images. From the results, it can be concluded that the proposed framework significantly improves the results with different spectral and spatial resolutions over different instrumentation.
Airborne and satellite hyperspectral remote sensing is a key technology to observe finite change in ecosystems and environments. The role of such sensors will improve our ability to monitor and mitigate natural and agricultural environments on a much larger spatial scale than can be achieved using field measurements such as soil coring or proximal sensors to estimate the chemistry of vegetation. Hyperspectral sensors for commentarial and scientific activities are increasingly available and cost effective, providing a great opportunity to measure and detect changes in the environment and ecosystem. This can be used to extract critical information to develop more advanced management practices.
In this research, we provide an overview of the data acquisition, processing and analysis of airborne, full-spectrum hyperspectral imagery from a small-scale aerial mapping project in hill-country farms in New Zealand, using an AISA Fenix sensor (Specim, Finland). The imagery has been radiometrically and atmospherically corrected, georectified and mosaicked. The hyperspectral data cube was then spectrally and spatially smoothed using Savitzky-Golay and median filter, respectively. The mosaicked imagery used to calculate bio-chemical properties of surface vegetation, such as pasture. Ground samples (n = 200) were collected a few days after the over-flight are used to develop a calibration model using partial least squares regression method. In-leaf nitrogen, potassium and phosphorous concentration were calculated using the reflectance values from the airborne hyperspectral imagery. In total, three surveys of an example property have been acquired that show changes in the pattern of availability of a major element in vegetation canopy, in this case nitrogen.
Pasture quality is a critical determinant which influences animal performance (live weight gain, milk and meat production) and animal health. Assessment of pasture quality is therefore required to assist farmers with grazing planning and management, benchmarking between seasons and years. Traditionally, pasture quality is determined by field sampling which is laborious, expensive and time consuming, and the information is not available in real-time. Hyperspectral remote sensing has potential to accurately quantify biochemical composition of pasture over wide areas in great spatial detail. In this study an airborne imaging spectrometer (AisaFENIX, Specim) was used with a spectral range of 380-2500 nm with 448 spectral bands. A case study of a 600 ha hill country farm in New Zealand is used to illustrate the use of the system. Radiometric and atmospheric corrections, along with automatized georectification of the imagery using Digital Elevation Model (DEM), were applied to the raw images to convert into geocoded reflectance images. Then a multivariate statistical method, partial least squares (PLS), was applied to estimate pasture quality such as crude protein (CP) and metabolisable energy (ME) from canopy reflectance. The results from this study revealed that estimates of CP and ME had a R2 of 0.77 and 0.79, and RMSECV of 2.97 and 0.81 respectively. By utilizing these regression models, spatial maps were created over the imaged area. These pasture quality maps can be used for adopting precision agriculture practices which improves farm profitability and environmental sustainability.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.