Local urban sprawl accuracy from image segmentation uncertainties simulation

Urban sprawl can be monitored using remote sensing. One method is to segment and classify the images to obtain polygons of impervious areas and then derive urban patches. We evaluated the geometric accuracy of the impervious polygons using the boundary distance distribution signature (BDDS) and investigated their propagation to spatial indicators of urban sprawl indicators at municipality scale, south of France, from impervious polygons uncertainties simulation.