Research Of The Road Recognition Technology Based On Topographic Maps Of Road

Image segmentation technology has very important applications in many fields and has made some research achievements in recent years; especially it is relatively mature in the research on the following aspects:text, faces and license plates as well as the medical image segmentation algorithm. While topographic map of roads has its unique characteristics, such as path length is longer, width is almost the same, gray scale distribution is relatively uniform, road area and the background gray level difference is bigger, the road image impulse noise and gaussian noise exist at the same time, thus it is quite difficult to recognize the road based on these characteristics. The road image de-noising preprocessing effects, the accuracy and systematicness of road information and background separation are key factors that will influence road recognition in topographic map.This paper first compared and reviewed various existing image segmentation methods, then analyzed the main features of the method of threshold segmentation, edge detection and region segmentation. In combination with characteristics of landscape image of road, in addition to the method of image gray-scale algorithm, gray histogram equalization algorithm, on the basis of introduction of power plant parameters, an improved median filtering algorithm was proposed to have filtering de-noising processing after the extension of the image contrast. Experiments showed that we have achieved good results on the suppression of pulse noise and gaussian noise. Then in view of the road landscape features, this paper designed an image processing technology which includes the median filtering and the threshold segmentation, and adopted the image skeleton, to extract the features; and then based on the algorithm it proposed a set of relatively complete target recognition algorithm and process.Finally we designed and realized the road recognition system, which realized the position of road and terrain target recognition and thus verified the correctness, reliability and feasibility of the proposed algorithm. The running result of this system showed that this system had high efficiency on the recognition of the road and it might have a certain practicality.