 12 impervious surface extraction methods were compared for four spatial resolution images, Landsat-8, Sentinel-2A, Gaffin-2, in three testing areas to determine the optimal classification method for each spatial resolution. The support vector machine SVME method was found to be the best for medium spatial resolutions of 30 and 20 m. While the object-based image analysis, OBIA, method obtained the highest accuracy for high spatial resolutions of 10 and 4 m. The perpendicular impervious surface index, PZ, outperformed other indices in obtaining the impervious surface distribution with the highest accuracy for all for spatial resolution images. This article was authored by Shen Shen Feng and Feng Li Fan.