Research on Bus Passenger Flow Statistics Based on Video Images

In order to meet the travel needs of inhabitant, buses as a travel mode of “low-carbon, environmental protection and green travel” have been widely used, so the statistics of public transport passenger flow has gradually become an important reference data source for rational traffic planning and optimization of public vehicle route scheduling. Aiming at the impact of incorrect counting caused by complex conditions such as movement, crowding and occlusion in the actual bus surveillance video scene, this paper adopts the passenger flow statistics system with the passenger's head area as the target detection. The algorithm proposed in this paper mainly uses the head area of the passengers getting on and off as the target feature to be extracted, and on this basis, the image processing and recognition operations are performed. Then input the above results into the target tracking algorithm to be used to estimate and preprocess it, and then use the data information obtained by the algorithm to achieve the association matching between multiple targets. Finally, the counting is completed by analyzing the target motion trajectory, so as to realize the statistics of boarding and alighting passenger flow. The experimental results show that the passenger flow statistics have strong robustness, track and count the target passengers accurately, and meet the needs of actual passenger flow statistics.