行人违规的跟踪与行为预测

Pedestrian Violation Tracking And Behavior Prediction

研究行人违规的跟踪与行为预测问题,为降低行人违规率和伤亡人数,通过中值滤波法,抑制噪声的干扰,并通过背景差法,分离背景区域与运动区域,将运动的行人区域检测出来。运用基于颜色匹配和模板匹配的目标跟踪算法,对FP⁃GA开发板安装相关编程和相应环境搭建,把软件实现的检测算法加到开发板中,可有效达到行人违规的跟踪与行为预测的目的。

The tracking and behavior prediction of pedestrian violations were studied.To reduce the rate of pedestrian violation and the number of casualties,the median filtering method was used to suppress the interference of noise,and the background difference method was used to separate the background area from the moving area to detect the moving pedestrian area.By using the target tracking algorithm based on color matching and template matching,the FPGA development board is installed with relevant programming and the corresponding environment is built,and the detection algorithm realized by the software is added to the development board,which can effectively achieve the purpose of pedestrian violation tracking and behavior prediction.