Study Of Intelligent Transportation Surveillance And Control System Based On Images Recognition Technology

With the high-speed development of our national economy, there are more and more construct in the domestic expressway, urban road, and parking area, etc. The requisition on the traffic control, sagety management improves day by day. The intelligent traffic system (ITS) has already become the main direction of present traffic administration development, and image recogniting which is a key technology to ITS, plays a very important role. It wide application will contribute to the process of the traffic mechanography of our country.The dissertation mainly deadls with the illegibility of license plates in ITS that arises from bad weather, insufficient light and awful road condition, and tries to make the algorithms more adaptable to environment. The license plate recognition mainly includes 4 parts: image pre-processing, license plate location, characters segmentation and characters recognition. Some improved algorithms are presented in these approaches and the license plates are recognized correctly at last. In the image pre-processing, original images are enhanced, and then a denoising method based on wavelet analysis and genetic algorithms is proposed to denoising and take good effect to get ready for further work. In license plate location, a location method is presented. First of all, it gains the edge information which is obtained by decomposition with the parts of wavelet high-frequency. Secondly, the license plate is obtained according to the projection of characters. In segmentation of license plate characters, a segmentation method based on the improved Snake model is proposed for solving the segmentation of merged characters. In recognition of license plate characters, for higher recognition rate, the system of characters recognition is divied into four subsystems according to the distribution of characters. A method of number-plate recognition using AdaBoost algorithm, based on templates matching is used. It inherits merits of trondional AdaBoost algorithm, such as the high rate, speed of recognition, and etc. In order to solve the problem of overfit, which the classical AdaBoost algorithm brings, a weight-adjusting method based on the distribution of samples is brought forward.At the same time, for intelligentizing, miniaturizing and improving performance-to-price of the system, the intelligent transportation surveillance and control system based on DSP was presented, which use high-powered DSP to replace PC as central control unit. We describe the circuits of hardware detailedly. The hardware system is used to detect the proposed algorithms (from image preprocessing to segmentation of characters). The experiment results show that the performance of the system is satisfied.