Multi-lane traffic flow monitoring and detection system based on video detection
Xue Liu, Xiaowei Wang, Zhaosheng Yang · Journal of Intelligent & Fuzzy Systems · 2019
The paper adopts the improved ViBe algorithm, and initializes the background sample according to the original algorithm. Using the idea of image binarization, the current frame image is subjected to the foreground detection of the traditional ViBe, and then the second judgment is made according to the number of times the target is judged, and the original is changed. The fixed background factor in the algorithm background update strategy increases the background update probability during shadow elimination. Experiments show that the method can eliminate shadows in time and improve the accuracy of motion vehicle detection in different scenarios. At the same time, the paper designs a multi-lane traffic flow monitoring system based on improved ViBe algorithm combined with IoT ZigBee technology. The flow monitoring accuracy of the algorithm is verified.