An Improved Canny Operator Of Image Segmentation For The Application In Vehicle Detection

With the development of computer technology and the rise of artificial intelligence,people start to study how to process image with computer and then the subject of computervision comes into being. The image processing is an important part in the newly appearingsubject. Besides, the technology of image segmentation is worthy of paying more attention to.Image segmentation is to obtain and perceive the location and the presence of the target object,and further to complete the goals of identification, tracking and other tasks. In recent years,Intelligent Transportation Systems have also made considerable progress, with thedevelopment of computer vision technology. The video-based vehicle detection, not onlyrelieves the pressure on the manual operation, but also plays an important role of guidanceand protection in traffic control, safe driving, vehicle statistics, traffic planning and designaspects. This vehicle detection system is automated and intelligent, and provides informationfor decision making in the increasingly complex and high-traffic environment. VehicleDetection in Intelligent Transportation System was crucial and played a leading role.Detecting a vehicle belongs to the problem of image segmentation, and the accuracy androbustness of image segmentation are related to the performance of the vehicle detectionsystem.Image Segmentation is the technology and process of segmenting interested subject ortarget from an image in need of different people. For example, in the process of vehicledetection, we extract the edge of the vehicle, and obtain the area of the vehicle, therebysegment the moving vehicles. This kind of segmentation aims to segment an image intodifferent regions without common characteristics according to the textural features of animage, the gray scale, the color and so on. Image segmentation is widely applied, however;there is not yet a technology so far that can meet the requirements of all applications.The purpose of this paper is to segment the vehicle images from a vehicle video, byemploying an improved Canny operators. We extract the vehicle video images, that is thetarget object of interest, then we finish the task of extracting the vehicles. Firstly, some basicknowledge of image segmentation and vehicle detection is introduced, including the meaningof studying image segmentation and current study situation at home and abroad. Secondly, thedefinition of the image segmentation and the classification of the segmentation algorithm are also stated in this paper, including several commonly used segmentation algorithms and theiradvantages and disadvantages. The next is the detailed presentations of the imagesegmentation based on improved Canny operator and its specific application in vehicledetection. The last experiment results show that the image segmentation based on improvedCanny operator suggested in this paper can be achieved and can get good effects though someshortcomings exist. Finally, a summary about this paper is made; the faultiness of thesegmentation algorithm is explained and the future research direction is proposed.