Moving Vehicle Detection And Tracking Based On Computer Vision

Video Monitor system is a typical application of computer vision field, as the continuous development of the technology,the transportation detection and track technology of intelligent video monitor system, which is a branch of computer vision, has become new and hot research field. On the basis of analyzing and summarize current kind of detection and track technology, the text mainly studies the detection and track technology of multi-targets in the fixed single camera shooting vision background on the road.The object is moving vehicles detection and track, and the emphasis is the detection of moving vehicles, with the focus is the detection of multi-target vehicle in the condition of fixed single camera. The moving target detection algorithm is mainly background subtraction and adjacent frame difference algorithm.Furthermore,as for the improved surendra algorithm,the experiment is done and it proposes improved algorithm based on the Gaussian model, and improved adjacent frame difference algorithm,analyzing and reducing the complexity traditional gray space shadow detection algorithm. Combining the robustness and real-time of the algorithm to choose the algorithm suitable for the system. With the comparison and experiment of variant algorithm, it derives improved algorithm and their advantages and disadvantages of the traditional algorithm. It can achieve the detection of more vehicles in the vision scope well;with high accuracy of track and the robustness has arrived a certain degree of demand and with good real-time control of track process.