Swarm based urban road map updating using high resolution satellite imagery

Roads are of important elements of urban infrastructure which directly and crucially affect the life of people in cities. The criticality of the roads network in urban areas justifies the efforts made to use remotely sensed data especially High Resolution Satellite Imagery (HRSI) to generate fast and accurate knowledge about road networks and to update road maps. Besides, on the way towards inventing soft computation solutions for geospatial problems, Swarm intelligence approaches such as Ant Colony Optimization are known to be helpful as they hire the emergence from the collective behaviour of social animals to solve combinatorial optimization problems. In this paper ACO solutions are inspected to propose a hand for road map updating using HRSI. First, image and map are overlaid and road elements in the image are extracted. Then, an engine for evaluation of these roads is invented on the basis of supervised classification. To add newly constructed roads to the map, some random seeds in the area are selected and possible candid connections are generated. These candid new roads are heuristically inspected using ACO algorithm and the proposed engine. The results are generated using GeoEye pansharpen images of a sub urban area located in the city of Tehran. Visual investigations show satisfying results and it is presumed to present more accurate results, enhancing ACO optimization, evaluation and extraction procedures.