Research On The Algorithm Of City Road Extraction From High-resolution Images

City Road is an important part of the urban basic geographic information, it is very important for urban planning, vehicle navigation, geographic information database update, and cartography. How to extract accurately and efficiently the road information from remote sensing images become a research hotspot in recent years. Since the seventies of last century, experts and scholars from domestic and foreign began to research on the algorithm of road extraction, and made a lot of achievements. But most of the achievements are semi-automatic road extraction, there are many defects,such as large limitations, small scope of application and so on. With the continuous improvement of remote sensing image resolution, and further study of road extraction in high-resolution image is more theoretical and practical significance. In this article,the reserach is mainly about the urban road extraction from high-resolution remote sensing images.In the process of obtaining images, the sensor will be affected by narural, system and other factors, the image what we got must contains a lot of noise. In order to ensure road extraction accuracy, image preprocessing before road extraction is necessary. In this article, author study and compare the common algorithms of image preprocessing, provide basis for pre-processing algorithm selection.Hough transform is one of the commonly used algorithms for road extraction, the paper describes the principle and calculation of the hough transform algorithm, it tranforms the points of the edge of the image from image coordinates to parameter coordinates, so the lines detection from image become points detection in parameter coordinates. Then the author extract the road use bilateral filtering and hough transform.Introduces the basic theory of mathematical morphology of digital image processing, morphological operations include expansion, refinement, and open operation, closing operation and morphological thinning. Traditional hough transform has many defects,such as large memory usage, low computational efficiency and so on, It is not conducive to efficient extraction of complex road. The author introduce fuzzy set theory to calculate the cumulative matrix, introduce the fuzzy hough transform algorithm, and use the morphological thinning algorithm before fuzzy hough transform to further reduce the amount of computational.For the high-resolution images of urban roads, the paper analyzes the basic characteristics of the road in the image, and studys the road extraction algorithm, proposed some new ideas for road extraction, and in the extraction experiments we get some satisfactory results.