顾及道路约束条件的动态目标检测与跟踪方法

A Method of Dynamic Target Detection and Tracking Based on Road Constraint Condition

智能交通是智慧城市的重要组成部分,面对复杂多变的道路背景,如何能够快速检测、跟踪道路监控影像中的动态目标,是智能交通建设的关键技术难点。本文根据道路监控视频特点,提出了采用道路约束条件与颜色特征集相结合的动态目标跟踪方法,以道路约束条件确定运动目标搜索区域,利用HSV颜色特征集进行特征匹配,然后基于IIR滤波背景法对背景影像进行更新及动态目标的检测,并根据道路约束条件与颜色特征相结合跟踪方法实现对动态目标的跟踪及动作预测。试验结果表明该方法可准确对运动目标进行检测与跟踪,且对慢速运动目标也具有较好的响应能力,实现了对道路动态目标的实时检测与跟踪。

As an important part of intelligent city, intelligent transportation is an important guarantee for urban road safety. The difficulty is how to quickly detect and track dynamic targets in road monitoring images with the complex and changeable road background. According to the characteristics of road monitoring video, a dynamic target tracking method is proposed based on the combination of road constraints and color feature sets. Firstly, the moving target search region is determined by road constraints, and HSV color feature set is used for feature matching. Then the background image is updated and the dynamic target is detected based on HR filtering background method. Finally, tracking and motion forecasting of dynamic objects are realized based on the tracking method which combines the road constraints and color features. The experimental results show that the method can detect and track the moving target accurately. It also has good response ability to the small displacement target and realizes the real-time detection and tracking of the road dynamic target.