摘要
无人驾驶是当前智能汽车领域的研究热点,交通信号灯的有效识别对无人驾驶的实现起着决定性的作用。针对交通灯图像的特征,首先采用Sobel边缘检测算法提取信号灯图像边缘;然后对边缘图像采用数学形态学处理,获得连通区域,实现交通信号灯的定位;最后将定位后灯框区域转换成HSV颜色空间并只保留红色或者绿色的通道,再判断信号灯的颜色状态,从而识别出信号灯的颜色。用MATLAB进行仿真实验,结果表明,该算法的识别效果好,识别率高。
Driverless is a research hotspot in the field of intelligent vehicles.The effective recognition of traffic lights plays a decisive role in the realization of driverless.Firstly,Sobel edge detection algorithm is used to extract the edge of traffic light image;then,the edge image is processed by mathematical morphology to obtain the connected area and realize the location of traffic lights;finally,after positioning,the lamp frame area is converted into HSV color space,and only the red or green channel is reserved,and then the color state of the signal lamp is judged so as to identify the color of the signal lamp.The simulation experiment with MATLAB shows that the algorithm has good recognition effect and high recognition rate.
作者
陈蓉
罗山
Chen Rong;Luo Shan(School of Intelligent Manufacturing, Panzhihua University, Panzhihua Sichuan 617000, China)
出处
《山西电子技术》
2021年第6期11-13,共3页
Shanxi Electronic Technology
基金
攀枝花学院大学生创新创业训练计划项目(2020cxcy078)。
关键词
图像处理
交通灯识别
数学形态学
边缘检测
image processing
traffic light recognition
mathematical morphology
edge detection
