| With the development of science and technology, computer technology and the artificial intelligence technology, the artificial intelligence technology's application field becomes more and more wide. This causes some manual works which are very bored are replaced by the same functional equipments in this field. It not only raises the productivity, but also reduces the man-power labor intensity and mistake and misjudgment by human being greatly. Therefore, from the intelligent video supervisory system's birthday on, it is applied in the safety control, the fire prevention and other kind of management fields very quickly and widely. The intelligence supervisory system is mainly used to track and record the human being's movement and path. Furthermore, according to some kind of special need, the intelligence supervisory system can search the track record which the user is interested in through an intelligent way, in a short time.This paper analyzes and summarizes the current research status of pedestrian tracking at first. And then according to the project's physical demand, the paper designs a new pedestrian track and the performance evaluation system. The new system is based on the implementation of particle filter in pedestrian tracking, and uses the improved Particle Filter. It uses the Mean Shift algorithm which has the low computation complex rate to solve the problem that the movement model is difficult to establish by estimating the pedestrian's position. Descripting outward appearance accurately by multi-feature fusion, it can track steadily in a complex background and occlusion in the video. This system is implemented in IDE Visual C++, with OpenCV and MATLAB. Experiment results show that the system's tracking performance achieves the real-time level. It can capture the objective and track it in a real-time. Moreover, it can carry on feedback on the tracking result and analysis it and show the analyzed results. |
PDF Full Text Request