Urban road extraction based-on morphological operations and radon transform on DSM Data

The main problem in extracting a road in an urban area from an aerial photo is the shadow cast by buildings. Because of this, we use Digital Surface Model (DSM) data, which are based-on the elevation of land surfaces. The problems associated with DSM data is non-road area with similar elevations like parking places, parks and so on. In this work, we propose a method to perform road filtering using Radon transform and morphological operations. The initial road is the result of intensity enhancement and instantaneous, simple thresholding. The road filter performs an opening operation on the initial road with a line shape structuring element, where the line orientation is obtained from the Radon Transform. Finally, the road is constructed based on B-Spline from the extracted road skeleton. The experimental result shows that the proposed method improved the quality and the accuracy average within an acceptable time.