Relative Fuzzy Connectedness In The Satellite Images Road Extraction

The technology of the road recognition is an active subject in the area of pattern recognition. In this paper, the new method of the road recognition, relative fuzzy connectedness algorithms , are probed. And the corresponding solutions are given.The work including:(1) Road extraction based on fuzzy connectedness theory, relative fuzzy connectedness theory in extraction of the multiple objects.Based on feasibility of fuzzy connectedness algorithms applied in image segmentation, this paper first introduced the fuzzy connectedness theory and algorithm Kv_XFOE, then the feasibility of application in road extraction was proved by the experiments. There were some scarcities in the results, So the new method based on the theory, relative fuzzy connectedness in extraction of the multiple objects, was probed to extract the road. In the end, the results of the experiments approved the validity of the new method.(2) Combined with the features of the road, fuzzy connectedness theory was partly applied in image to recognize the fuzzy segments of the road and finally perfected the information of the road.Considering the features of the road, fuzzy connectedness theory partly applied in the output images of the first experiments. Through the second extraction from the first outputs, the fractures of the road were eliminated. And on the basis of the existed results, the outline and the direction of the road was completed, the wrong information was also cleared up. In the end, the good output image offered good proof.