Urban impervious surface extraction method under SAR (Synthetic Aperture Radar) image space-time similarity analysis

The invention relates to an urban impervious surface extraction method under SAR (Synthetic Aperture Radar) image space-time similarity analysis. The method comprises the following steps: preprocessing an SAR image data set and constructing a pixel-level SAR image time sequence through high-accuracy matching; calculating DTW (Dynamic Time Warping) values of a sampling mixed pixel and a pure pixel as maximum threshold values for similarity extraction by taking the DTW as similarity measurement of the pixel-level SAR image time sequence; fully considering the better similarity of the same land feature pixel time sequence and the proximity feature of the space by using a time sequence similarity-based spatial context analysis method, calculating the DTW values of the corresponding time sequence curves in a template window and a slide widow respectively using a slide window technology by taking the time sequence curve of a typical land feature pure pixel as the template window, and further determining the land feature type of a central pixel by using the spatial proximity rule of the pixel. According to the method, the land feature extraction and breaking phenomena can be improved, and the information extraction accuracy of an urban impervious surface is improved.