Research On The Methods Of Road Extraction From High Resolution Remote Sensing Images

As an important land surface type,roads play an important role in people's life.Extracting road information from remote sensing images is of practical scientific value and practical significance.But with the development of remote sensing technology and the improvement of the resolution of remote sensing images,the information describing the objects becomes more and more abundant.The difficulty of identifying and extracting road information is also increasing.In this paper,a variety of remote sensing image road extraction methods have been deeply studied,and based on remote sensing image characteristics and road characteristics,we have made some beneficial improvements to the existing road extraction methods.First of all,for the road extraction method based on the edge line,the detection method of traditional optical image edge can not be well applied to the remote sensing image because of the different characteristics of remote sensing images.Therefore,on the idea of the canny edge detection algorithm,this article firstly adopts a smoothing and self-adapting Gaussian filter to reduce the noise of remote sensing image and less the noise interference as well as reserve the edge and details.Then,in the edge judgment of the dual threshold,select the high and low thresholds on the basis of local characteristics within the object scale of the pixel point and enhance the exact judgment performance of edge.The experiment results show that the new method can effectively improve the accuracy and positioning accuracy of the edge detection,obviously reduce the misjudgment of road edge extracted and remarkably increase integrity and consecutiveness,with high automation.In light of the difficulty in improvement of the extraction accuracy of road segmentation caused by the complex scene of high-resolution remote sensing images,a different object with the same spectra characteristics and the same object with the different spectra characteristics of surface features,a method of road extraction of mean shift algorithm based on the combination of multilayer features is proposed.First of all,according to the edge structural information which is representative and can distinguish from other surface features satisfactorily,and on the basis of edge statistics model,extraction of image edge statistical features is made.By means of the mean shift algorithm,the position feature,spectral characteristics and edge statistical feature of the images are combined to segment the remote sensing images.At last,there are still a few misjudgment and leak-judgment of road edge and region.According to the geometric features and morphological methods of the road,the fine processing is carried out.The experimental results show the effectiveness and superiority of the two methods of road extraction.