ENVISAT ASAR for Land Cover Mappingand Change Detection

Abstract: The principal objective of this research is to investigate the capability of multi-temporal,multi-incidence angle, dual polarization ENVISAT ASAR imagery for extractinglanduse/land cover information in the rural-urban fringe of the Greater Toronto Area (GTA)using different image processing techniques and classification algorithms. An attempt todetermine the temporal change of landuse is also made.The multi-temporal ASAR imagery was first orthorectified using NTDB DEM and satelliteorbital models. Different image processing techniques, such as, Adaptive Speckle Filtering,Texture measures, Principal Component Analysis (PCA) were applied to the ASAR images.Backscatter profiles were generated for selected land cove classes. K Nearest neighbor (kNN)classifier was used to extract eleven land cover classes. Artificial Neural Network (ANN) wasalso tested with some selected combinations of ASAR imagery. The classification schemewas adopted from USGS alnuse/land cover classification scheme. Average accuracy, overallaccuracy and Kappa coefficients were calculated for all classifications.The raw ASAR images gave very poor results in identifying landuse/land cover classes due tothe presence of immense speckle. Enhanced Frost (EF) filtering significantly improved theclassification accuracies. For texture measures, eleven date Mean images produced the bestresult among all single set processed data. Combined Mean and Standard Deviation,combinations of different texture measures, further improved the results. Standard deviationprovided vital auxiliary boundary information to the classification resulting in theimprovement. The best kNN was achieved with combined Mean and Standard Deviation withmulti-incidence angle, dual polarization eleven date ASAR images. ANN further improvedthe classification results of the textured images. As for comparison of classifiers, It was foundthat, with complex combinations (dual polarization, multi-incidence angle), ANN performssignificantly better than kNN. The overall accuracy was 9.6% higher than that of kNN. Theresults were more or less similar in filtered images.Post classification change detection is largely dependent on classification accuracy ofindividual images. Even though, the classification results were somewhat satisfactory, theclassified ASAR image still had a significant amount or omission and commission errors withsome classes. The classification errors contributed a significant amount of noise in changedetection. The change detection procedure, however, was able to identify the areas ofsignificant change, for example, major new roads, new low and high built up areas and golfcourses.In brief, ENVISAT ASAR data was found to have vast potential in extracting land coverinformation. Especially with its all weather capability, ASAR can be used together with highresolutionoptical images for temporal studies of landuse/land cover change due to urbansprawl.