Use of Multispectral and Hyperspectral Satellite Imagery for Monitoring Waterbodies and Wetlands

Abstract

Timely and accurate change detection (CD) of Earth’s surface features is important for understanding interactions between human and natural phenomena. Remote sensing (RS) as the most important information resource plays a role key in monitoring and assessment of the environment. One of most applications of hyperspectral imagery is CD. The hyperspectral imagery provides more details from CD compared to multispectral images. Wetlands are one of the most influential ecosystems in the natural environment for which it is very difficult to find an alternative. The monitoring wetland and waterbody areas based on RS imagery need special techniques due to some limitations (existence noise and condition atmospheric, need to high training data and threshold selection, and complexity of water body areas). Based on these problems it is necessary to CD methods to minimize problems so, this research proposed a framework for hyperspectral CD methods on wetland and water body areas. The proposed method is based on incorporating chronochrome, Z-score analysis, Otsu algorithm, SImplex via Split Augmented Lagrangian (SISAL), Harsanyi–Farrand–Chang (HFC), Pearson correlation coefficient (PCC), and support vector machine (SVM) to detect changes using hyperspectral imagery. The proposed method is applied in four main steps: (1) produce a training data for tuning SVM and kernel parameters, (2) predicted change areas based on a chronochrome algorithm and binary change map obtained using SVM classifier, (3) the amplitude of changes is created by Z-Score analysis and binary change mask, and (4) the multiple change map is produced based on the estimation of number and extraction of endmembers and similarity measure. To evaluate the performance of the proposed method, multi-temporal hyperspectral Hyperion images for Shadegan Wetland were used. The results show high accuracy and low false alarms rate of proposed method methods with an overall accuracy of more than 96%, kappa coefficient of more than 0.82. Besides, the proposed method can provide ‘multiple changes’ as well as the magnitude of the extracted changes.