Research On Road Segmentation And Extraction's Algorithms Of High-resolution Remote Sensing Images

The main purpose of this article is how to achieve the automatic recognition and extraction of road target from high-resolution remote sensing images. In the process of research, first uses all sorts of algorithms of remote sensing images segmentation to segment the road, then uses the relevant knowledge of mathematical morphology and pattern recognition to achieve automatic extraction of the road target.The main content of this article is:(1) First do pre-processing before the segment of images, pre-processing will have great influence on segmentation effect (particularly in image enhancement). This article summarizes some successful methods of the image enhancement, and for the image's enhancement processing by using remote sensing image histogram equalization and method of mathematical morphological.(2) This paper focuses on the Canny edge detection algorithm,threshold segmentation algorithm,K-means clustering segmentation algorithm etc, and improved them, so for meeting the conditions of automatic extraction of road features from high-resolution images.(3) Practice has proved, any kind of segmentation methods are difficult to get satisfactory segmentation results, however when some segmentation methods fusion applied can obtain good results. Therefore, this paper combined some segmentation methods to application, the research studies the method of combining edge detection segmentation algorithm and K-means clustering segmentation algorithm to extract the road information.(4) The good segmentation results have obtained, using the method of mathematical morphology image processing for further processing of the segmentation results, then using the knowledge of pattern recognition to recognize road target, getting a good road information.