A Real-Time Traffic Detection Method Based on Improved Kalman Filter
Xun Li, Nan Kaikai, Yao Liu, Zuo Tao · 2018
For the lack of current extraction methods of traffic basic data, a traffic information acquisition method based on improved Kalman filtering using traffic video was presented in this paper. The Gaussian mixture model was improved for multi-vehicle moving targets detection. In order to further improve the detection efficiency, a heuristic improvement method was proposed. For the matching problem of multiple targets in the continuous video frame, combining the vehicle movement characteristic, the Kalman filter was used to estimate the vehicle position optimally, a real-time traffic detection method of matching the target chain was proposed. Finally, the experiment was carried out with the actual transportation video, results show that the proposed method can effectively improve the noise interference and foreground blurring in Multi-target vehicle detection, and can extract the vehicle moving target information from different traffic environments with high accuracy, different models and vehicle color The lowest detection rate was 93.08%.