Design And Implement Of System For Road Recognition Based On Image Partition

Road recognition plays a very important role in many areas, the identification ofmilitary topography road in battlefield is beneficial to open up channels, the departmentof surveying and mapping can accomplish various landform track and detection by theway of road recognition, and so as to improve the efficiency of mapping, the departmentof environment also can compare to the protection of the natural environment make useof the road features, the center of national earthquake rescue can estimate the damagesituation of road with the road feature recognition. In addition, the road terrainrecognition in tourism information, geological exploration, environmental monitoringand protection of land resources and other aspects, has an urgent and realistic demandand application.The digital image processing technique applied in road recognition that is theinnovation of this paper, which has certain theoretical and practical value, and animportant effect on road recognition based on the image segmentation technology. Inthis paper, we first briefly reviewed the existing image segmentation methods, analyzedthe method of threshold segmentation, edge detection and region segmentation method.Then we combined with the characteristics of the road image, and adopt a gray leveltransformation, histogram equalization method to preprocess the image. After thecontrast expansion, we propose an improved median filtering algorithm on image filterprocessing in the image, which may lay the foundation of feature extraction andrecognition for the road. According to the characteristics of road and terrain, this paperdesigned a road recognition algorithm that used threshold segmentation, image skeletonto extract the features, and then analyzed the target of the extraction process in detail.Finally we used Visual C++to design and realize the road terrain recognition system.The main work of this paper includes the following aspects:(1) Analying the technology of image segmentation deeply using the comparativemethod; (2)Proposing an improved median filtering algorithm for image processing makeuse of adopting the methods of image algorithm and histogram equalization algorithmaccording to road and terrain characteristics of the image;(3)Using the image processing technology of median filter, threshold segmentationto extract the image feature of the image skeleton, and puts forward a set of relativelycomplete knowledge driven, top-down, target recognition algorithm and process forsuch a goal according to the road terrain features;(4)Realizing the road terrain target recognition, accomplishes the road terrainrecognition system design, and verifies the proposed algorithm correctness, reliabilityand feasibility by using the object-oriented design language Visual C++6.0forprogramming.