Research Of Video Detection And Peccancy Analysis Based On ITS

Video detection and peccancy analysis based on ITS is a hotspot in IT and is a fairly active field, which combines advanced technologies in image processing, pattern recognition, automatic control, artificial intelligence, computer and other relative fields. Besides of this, it is also continuously enriched itself by adopting the new concepts and new methodologies from related fields and is of great importance in daily life and national defence, military affairs, so this subject research has important theoretical significance and practical value.Based on the fruits related to LPR, both domestic and abroad, this paper is aimed at the characteristic of transportation of China, giving a new idea of video detection combine peccancy analysis. Finally the video detection module and peccancy analysis recognition module were designed and realized. We can use video detection module to collect traffic data and use peccancy analysis recognition module to recognize and record the car witch is got out of line. In this paper, the traffic parameter measurement algorithm and plate recognition were both researched. At last, based on the experiment results, we give some available arithmetic and designed and realized the system based on virtual loop detection techniques. The major works of this paper are summarized as follows:This paper is aimed at traffic video sequences acquired through a static monocular camera in specific traffic scenes, use the method of moving target detection. In this paper, we analyse some arithmetics, including adjacent frame difference algorithm, background difference algorithm and model-based algorithm. Based on this analysis, an improved model based algorithm is proposed which effectively solve the problem of model update after changing of the background. The shadow model and main shadow detection algorithm is also introduced in this paper, and the shadow detection algorithm based on HSV spaces is discussed in detail, experiment is given to verify the validity of this algorithm.The most important technique in peccancy analysis recognition module is plate recognition technique. The plate recognition mainly contains three processes: license plate location,character syncopation,character recognition. In this paper, we will introduce the principles of realization and algorithm in detail. In the process of plate location, we use the arithmetic both based on transcendental knowledge and horizontal gradients of the car plate. In the part of eliminate car light, this paper gives a new arithmetic by which we count each maximal gradient in 4 rank matrix in the image block before we use Sobel arithmetic operators. By this we greatly improved the plate location rate. In the process of character syncopation, we use the arithmetic which is based on image projection and transcendental knowledge. And in the last process of character recognition, we projected a three level character classifier, they are coarse classifier based on fuzzy logic, pinpoint classifier based on the matrix element comparison idea we proposed and similar character classifier based on special point characteristic.In the dissertation, we also discussed image contrast enhancement, slant rectification and image binary, which are very common in LPR. In this paper, we analyse some common traffic data collection and plate recognition arithmetics, and modify some parts in necessary. And the software is designed and realized using model method in the VC++6.0 environment. At last the system is tested using the video data captured at ShunGeng Station on JingShi Road. The results of the test indicate that the system is with real time capability and achieves enough accurate results.