Computing Shortest Path for Transportation Logistics from High-Resolution Satellite Imagery

One of the most key facets of transportation logistics systems is the traffic management where city planning, road monitoring and speed of transportation play significant role. Regardless of the firm technologies available, road extraction from high-resolution satellite imagery has been an interesting research field focused in recent years. The study deals with extraction of topographical features like roads from high-resolution satellite imagery and computation of shortest path for transportation logistics. In this study, the topographical features and model from high-resolution satellite imagery were analysed. Then the comparison study of various road extraction algorithms was performed. After performing the preliminary processing like histogram visualization and grey-level thresholding, path opening and closing morphological filter was used. The response of the filter was then used to extract the curvilinear structure which represents the road. Post-processing morphological operation like thinning was also applied for removing distorted artefacts. Finally, shortest path between the source and the destination was approximately commuted using quasi-Euclidean distance method. The algorithm was extensively tested using several satellite imageries, and some of the selected results were presented in the paper. It is evident that one type of topographical features is not enough to obtain good results. The road extraction would be combined with other algorithms based on the requirement of applications to yield optimal solution for finding the shortest path. The proposed algorithms would also have the application of map generation for speedy transportation.