The Research And Realization Of Image-based Traffic Jams State Recognition System

With the development of Intelligent Transportation Systems, China's rapid economic development and urbanization process, transportation has become heavy because that the increasing of private vehicles bring us a number of transport problems needed to be solved, though they have improved our standard of living. So, that makes us conduct more on studying of ITS. In the Intelligent Transportation (ITS), the timely access to the traffic flow technology is a smart transportation system based on using digital image-related technology, which has become the mainstream of the field of traffic control.In this paper, the traffic congestion of vehicles measured the composition of various parts of the traffic system. Innovation and significance of accurate estimates of the program lies not only on avoiding distortion of the data caused by the practical device that without different standards, but also further deepens the hot topic of research based on color image segmentation. It helps to reduce the errors caused by data distortion. Rigorous scientific research on this has been reflected across the border to be able to detect the exact number of vehicles.Observation points can be adjusted according to road conditions. All the observation points cover the entire city area of the traffic network. The system's image data derived from data obtained CDD camera images. It relies on "Eye of Heaven" system as a platform for the entire system. On this basis,discussion further will continue in this paper by identifying the vehicle positioning. Experimental results show the implementation of this system function properly.Graphic image based on the vehicle congestion identification is made at 80s of the last century, in which the measured thinking constantly improved and updated. Many of the data has showed a large number of research methods and the subject data, such as: background difference method, frame difference method, the road coil labeling method, optical flow method, and some other equipment, hardware-based implementations. Its main characteristics are complicated to operate, high cost, environmental adaptive capacity is low, the calculation volume, detection accuracy is low, and reliable ability to adapt is poor. This paper proposes method based on image processing that has made well-targeted research.This paper has done the main work as follows:1. This paper choose a method based on image after analyzing the various methods of traffic detection and traffic environment. In the process of realization, useful collection maps of traffic environment are cut purposefully. In addition, on the base of color-based image segmentation model, the image boundary extraction, edge patch and a series of traffic flow image processing operations implemented, this paper also carries out based on the weight of the traffic flow and comprehensive evaluation of traffic statistics.2. This article's research data is derived from the actual environment. Images collected in a real environment not only impacted by noise, but also by low degree of contamination due to light damage which reducing the images quality. This study and experiments uses many different algorithms to avoid the disturbing factors. In addition, some methods were combined and improved, and the results shows the improved methods have a good effect on avoiding the noise. It dose a very good preparatory work for following-up research.3. Using the recognition on the base of edge detection and a "handling - Patch - identify" strategy, you can quickly capture images under conditions of first-time data in a very short period of time. And in a very short time, the data processed will be feedback to the user.