Research On Road Extraction From High Resolution Images

In the remote sensing technology,intelligent computing and cloud computing is highly developed today,although the results of high resolution remote sensing image processing technology of high resolution image but the road emerge in an endless stream,automatic extraction is the research hotspot.In the field of navigation applications such as maps,land resources monitoring,urban and rural planning and other fields,automatic extraction of high-resolution image roads has become an indispensable and urgent need to improve the key technologies.How to get the road information accurately in the complex texture and geometry information of high resolution image is still the key problem of high resolution image target extraction.The traditional image road extraction has not taken into account the rich texture and geometric features of high resolution images.To this end,this paper puts forward a method of comprehensive analysis of various types of road based on high resolution images,high resolution image road extraction,extraction by eliminating positive and reverse thinking positioning,using the minimum bounding rectangle and minimum circumscribed circle,in the aspects of road extraction conditions,solves the problem of high resolution image road extraction.In this paper,we mainly study the extraction methods of various features of high resolution aerial images,get the texture features and geometric features of the road.To identify road use texture and Gabor extraction methods,using the minimum bounding rectangle and minimum circumscribed circle on the road outside the ground search and removed from the two aspects outside the road itself and the road surface,with the application of image segmentation method,using multi feature analysis,minimum bounding rectangle and circle extraction and the mathematical morphology for accurate road extraction.Comprehensive analysis of multiple features extracted in this paper process is as follows:(1)feature extraction of texture image Road(2)road image region segmentation(3)image road classification(4)feature analysis of mathematical morphology(5)fine processing.At last,we design and develop a high resolution image road extraction experimental platform using C++,and select the urban road image and mountain road image as the experimental image.The experimental results show that the proposed method can extract the road structure in high resolution remote sensing images,and can be used to set the threshold value for image processing.Compared with the previous extraction algorithm,the extraction effect is greatly improved.In the course of the experiment,it is found that there is a certain threshold setting rule for urban road extraction and mountain road extraction.Although the road extraction process and extraction and analysis method presented in this paper can,but in terms of adaptive threshold setting.It is necessary to make further research,at the same time for the extraction of shadows of serious road also need further study.