Vehicle Detection System Based On The Virtual Detector And Implementation

Intelligent Transportation Systems (ITS) is the future direction of development of transportation systems, it has provided the convenience for people because of its real-time, accurate, and efficient utilization of traffic information. Collecting real-time traffic information by video detection method has many unique advantages such as convenient and effective. It has gradually become the mainstream of the field of traffic information collection technologies.In this paper, Moving vehicles on the road is our study object in the video detection system. Difficult problems in video detection has been studied, such as background extraction and updating, the shadow suppression, vehicle tracking and information extraction. In the end, the paper introduces how to develop a prototype of video vehicle detection systems.In the aspect of background extraction, based on the comparison of existing theories and algorithms for system-specific objects (city expressways or high-speed Road), combined with the virtual loop set method, the paper proposed an extraction and updating algorithm based on the sub-blocks of invariant background, which greatly increased the efficiency of the background extraction time. This algorithm provides a strong guarantee for further follow-up.In terms of target detection, based on the virtual detector technology and the background extraction and updating, the author will detect whether there were any specific areas of the current existence of the vehicle using the image characteristics of low-order high-order statistics, and get better effect we eliminate shadows impact through the shadow suppression method.In the aspect of vehicle tracking, analyzing the classical target tracking method based on Kalman filter and through the establishment of the corresponding simulation model, we understand that this tracking model'process was clumsy, and each step impact track and test results, therefore, the paper proposed Kalman tracking based on virtual loop method through the analysis of actual scenes and improving classical tracking method. The tested result shows that this method fully meets the system requirements with a higher practicality.At last, the paper introduces how to develop a prototype of video vehicle detection systems by using OOP technology. The system achieve to record traffic parameters and query functions,and get more valuable traffic information by the further treatment on the traffic parameters.