An Improved Moving Target Detection Method Based on Vibe Algorithm
Xiaoqiang Shao, Xi Chen, Kangle Li, Zhichao Lv, Hua Zhu · 2018
The visual background extractor (ViBe) algorithm has a significant disadvantage in vehicle detection, that is when there is a moving vehicle to be detected in the first frame of the video, in the vehicle detection process of the subsequent frame, corresponding to the first Ghosting occurs at the position of the vehicle in the frame and the ghost will continue to disappear for a while, thereby interfering with the detection effect of subsequent frames, and the real-time performance of the background update is insufficient due to the fixed threshold. In order to overcome the shortcomings of the algorithm and improve the accuracy, realtime and detection rate of moving target detection, this paper proposes an improved vibe modeling moving target detection method, which uses multi-frame continuous image instead of single-frame image to initialize and reduce ghosting. For the influence of the model, the adaptive threshold is used instead of the global fixed threshold, which improves the robustness to the illumination mutation. Experiments show that the proposed algorithm can effectively improve the efficiency and accuracy of moving target detection.