CAJ | 학술논문

交通信号灯的检测与识别系统是无人驾驶与辅助驾驶必不可少的一部分，同时还可以辅助色觉障碍人群与疲劳司机进行安全驾驶，因此本文提出一种能检测和识别交通信号灯状态的新方法。首先在HSV色彩空间对交通信号灯进行颜色分割，利用交通信号灯被黑色矩形框包围的特征进行形状分割，将颜色分割与形状分割相结合确认候选区域；使用颜色直方图统计候选区域的H分量，识别交通信号灯的颜色；使用模板匹配的方法，识别箭头型交通信号灯的方向。实验表明，该方法的检测与识别率达到97%以上。
교통신호등적검측여식별계통시무인가사여보조가사필불가소적일부분，동시환가이보조색각장애인군여피로사궤진행안전가사，인차본문제출일충능검측화식별교통신호등상태적신방법。수선재HSV색채공간대교통신호등진행안색분할，이용교통신호등피흑색구형광포위적특정진행형상분할，장안색분할여형상분할상결합학인후선구역；사용안색직방도통계후선구역적H분량，식별교통신호등적안색；사용모판필배적방법，식별전두형교통신호등적방향。실험표명，해방법적검측여식별솔체도97%이상。
Because of the trafifc signal detection and recognition system is an essential part of unmanned systems and auxiliary driving system,at the same time it can also auxiliary dys-chromatopsia crowd and fatigue driver safe driving,so this paper puts forward a way to identify the state of the trafifc lights.Segmenting the trafifc light in HSV color space.Using the typical characters of the trafifc light surrounded by a black rectangle to shape the trafifc light split.According to the shape of segmentation candidate areas to conifrm the position of color segmenta-tion.The number of H of labeled region images is counted by color histogram,as basis to judge the trafifc light color.Using template matching,conifrming the direction of the arrow trafifc lights.Experimental results indicate that the overal recognition rates of the proposed method are over 97%.