Research Of Road Network Extraction From High-resolution Remote Sensing Image

With the rising of domestic high resolution remote sensing technology, the needof it in transportation industry drives its development tremendously in China. Trafficnetwork, as the geographic data source for multiple businesses, is playing a key rolein urban planning, vehicle navigation, disaster relief and military application.Meanwhile high resolution remote sensing technology provides a precise as w ell asprompt method for the collecting and updating of traffic network. In the past days,many automatic, semi-automatic road extraction algorithms have been produced withthe rapid growth of road extraction from remote sensing images. Nevertheless, as th ecomplexity of it, there is still a big distance between the application of thistechnology and the reality.This paper summarizes the characteristics and research methods with the supportof related projects in Ministry of Transport. Based on real practice, it proposes animproved road extraction method and a transport network automatic extractionmethod. After a series of Algorithm experiments and results tests, the proposalsbrought up in the thesis is proved to be highly effective and useful.At last, to achieve a broad application of the transport network automaticextraction technology proposed in this paper, a satellite monitoring analysis andremote sensing platform is designed on the basis of an optimized algorithm andtechnological process. Furthermore, the paper highlights a transport networkautomatic extraction plug-in system which can greatly reduce workload and therequirements for workers’ specialized knowledge. With an efficient time and costcontrol, it reveals a promising and broad application in engineering field.