Built-Up Area Mapping for the Greater Bay Area in China from Spaceborne SAR Data Based on the PSDNet and Spatial Statistical Features

Built-up areas (BAs) information acquisition is essential to urban planning and sustainable development in the Greater Bay Area China. In this paper, a pseudo-Siamese dense convolutional network, namely PSDNet, proposed automatically extract BAs from spaceborne synthetic aperture radar (SAR) data Area, which considers spatial statistical features speckle SAR images. The local indicators of association, including Moran’s, Geary’s, Getis’ together with divergence feature, are calculated for data, can indicate potential BAs. amplitude images corresponding then regarded as inputs PSDNet. framework, network independently learn discrimination ability original image features. DenseNet adopted backbone each channel, improve efficiency while extracting deep Moreover, it also has multi-scale sizes by using decoder. Sentinel-1 (S1) China used experimental validation. Our method BA extraction achieve above 90% accuracy, similar current product, demonstrating that our mapping data.