Multi-temporal RADARSAT fine-beam SAR imagery for landuse and land-cover classification in the rural-urban fringe of the greater Toronto area.

摘要

This research investigates the capability of the multi-temporal RADARSAT C-HH Fine-Beam SAR imagery for extracting landuse/land-cover information in the rural-urban fringe of the Greater Toronto Area (GTA) using various image processing techniques and classification algorithms.; At first multi-temporal SAR imagery were orthorectifed using DEM and satellite orbital models. Then various image processing techniques, including Adaptive Speckle Filtering, Texture Measures, Principal Component Analysis (PCA) and Wavelet Decomposition were applied to the SAR images. Landuse/land-cover information was extracted from original and processed images using Maximum Likelihood Classifier (MLC), Artificial Neural Network (ANN) and Contextual classifier with a ten-class landuse/land-cover classification scheme adapted from USGS landuse/land-cover classification scheme. Signature Separability, Overall Accuracies and Kappa Coefficients were calculated and compared for all classifications. (Abstract shortened by UMI.)