交叉路口道路环境辨识之初步研究——建立智能型执法系统的首部曲

Road Environment Recognition System at Intersection

利用影像处理技术结合自行发展之道路环境辨识算法,建立智能型执法系统之平台,并选定"未依规定变换车道"违规行为建立违规行为判定系统,以期作为智能型执法系统平台建立之先驱研究。利用影像中位数法建立背景、结合递归式背景建构模式进行背景更新;透过"特征判断"与"样板比对"辨识路面标线;最后将车辆侦测、追踪等影像处理技术结合道路环境辨识结果进行违规行为的侦测与辨识。经过实际验证,于道路环境辨识以及违规行为辨识皆达良好的辨识水平。

In this study, the image processing techniques and the road environment recognition algorithm were used in the intelligent enforcement system platform. The ＂Change-Lane at Will＂ was selected for the traffic violation detection. First, the background was constructed using temporally median filter and combining it with recursive and non-recursive background updating algorithms to update our background image. Next, the extracted features and template matching algorithm were used to obtain the lane edge trace. Finally, the moving car detection and its tracking algorithms were also used to recognize the traffic violation. The simulation results show that the road environment recognition and the traffic violation detection both get well detection caliber.