Texture analysis and classification of ERS SAR images for map updating of urban areas in The Netherlands

In single-band and single-polarized synthetic aperture radar (SAR) image classification, texture holds useful information. In a study to assess the map-updating capabilities of such sensors in urban areas, some modern texture measures were investigated. Among them were histogram measures, wavelet energy, fractal dimension, lacunarity, and semivariograms. The latter were chosen as an alternative for the well-known gray-level cooccurrence family of features. The area that was studied using a European Remote Sensing Satellite 1 (ERS-1) SAR image was the conurbation around Rotterdam and The Hague in The Netherlands. The area can be characterized as a well-planned dispersed urban area with residential areas, industry, greenhouses, pasture, arable land, and some forest. The digital map to be updated was a 1:250000 Vector Map (VMap1). The study was done on the basis of nonparametric separability measures and classification techniques because most texture distributions were not normal. The conclusion is that texture improves the classification accuracy. The measures that performed best were mean intensity (actually no texture), variance, weighted-rank fill ratio, and semivariogram, but the accuracies vary for different classes. Despite the improvement, the overall classification accuracy indicates that the land-cover information content of ERS-1 leaves something to be desired.