基于动态阈值的高分辨率遥感影像道路提取

In this paper, an approach to extract road in the high resolution remote sensing image based on dynamic threshold is presented, and it can effectively resolve problems inherent in conventional methods with a single threshold value that cannot work well for extracting all roads from a whole image. Firstly the road information in the image is enhanced and the threshold is obtained according to the method of Otsu operation, then it is used for the image segmentation, the road information is included in the binary image. Secondly, a part of road with mathematic morphology erosion is obtained, stored in the queue. Then, the extracted road is taken as background. And then another road is extracted based on new threshold. Until the maximize information of road has been extracted. Finally, the centerline of the road is determined by morphological thinning and cropping. Experimental results show that the proposed method based on dynamic threshold values performs effectively on road extraction from high resolution remote sensing imagery.