Method Reashch Of Damaged Road Information Extraction Based On SAR Images

The technology of Automatic extraction road damage from the synthetic apertureradar (SAR) images has become an important research topic of disaster assessment. Themain contents in this paper are road extraction, change detection and damageassessment.Three different features of road and road extraction algorithm based on thesefeatures are analyzes in this paper. First, the author discourse the influence of differentresolution of SAR images on road extraction algorithm, and a method for road detectionbased on Duda operator is proposed according the resolution of project SAR images. Inorder to extract the complete road information and remove non-road information, thistext combines a series of image processing algorithms, such as morphologic thinning,the intersection removing, the short lines removing. Finally, the author used computerprogram based on SAR images to verify this algorithm and achieved satisfactory results.In this paper, change detection algorithms and theirs advantages and disadvantagesare introduced, and eventually the text chose the change detection algorithm whichcombine Edgeworth approximation and cross entropy algorithm after algorithmscertificated on SAR image data. The algorithm extracted the difference information ofSAR images in complex scenes by weakening the speckle noise influence through thecomputer program. Compared with the optical remote sensing images, the algorithm isable to accurately extract the change differences from SAR images of post-disaster andafter reconstruction in Yinxiu.Finally, the limitations of road damage assessment model based on optical remoteare analyzed. Then, the author has chosen road damage evolution factor for SAR imagesand the article introduce an idea which combine road extraction and change detectionalgorithm to extract damage information from SAR images.