基于图像识别的信号灯路口辅助驾驶方法

发布时间：2018-03-10 04:02

本文选题：信号灯路口　切入点：信号灯倒计时识别　出处：《浙江大学学报(工学版)》2017年06期 　论文类型：期刊论文

【摘要】：为了提高有信号灯路口的通行效率,将通过路口的车辆行驶状态分为4种情况(匀速通过、加速通过、减速通过和停车等待),并针对不同情况对驾驶员提出相应的驾驶行为提示.提出基于单一摄像头的图像识别系统,该系统通过加权K近邻方法识别出信号灯的颜色和倒计时.根据信号灯识别结果,建立车载摄像头与信号灯之间距离的计算模型,得到车辆与信号灯之间的距离和车辆车速.根据信号灯状态以及距离和车速信息判断车辆所处的行驶状态,给出合理的车速控制建议.实验结果表明:信号灯识别准确率高于97%;速度计算的平均误差小于5%;仿真实验验证结果表明:所提辅助驾驶方法能够减少车辆在信号灯路口的停车等待时间,提高通行效率.
[Abstract]:In order to improve the efficiency of traffic at signalized junctions, the vehicle driving state of the crossing is divided into four situations (passing at a uniform speed, speeding through, Decelerating passing and stopping waiting, and according to different situations, the driver is given corresponding driving behavior hints. An image recognition system based on a single camera is proposed. The system recognizes the color and countdown of the signal lamp by weighted K nearest neighbor method. According to the result of the signal light recognition, the calculation model of the distance between the vehicle camera and the signal lamp is established. The distance between the vehicle and the signal light and the speed of the vehicle are obtained. The driving state of the vehicle is judged by the state of the signal light and the information of the distance and the speed of the vehicle. The experimental results show that the recognition accuracy of the signal lamp is higher than 97; the average error of speed calculation is less than 5; and the simulation results show that the proposed auxiliary driving method can reduce the vehicle signal. Waiting time at a light junction, Improve the efficiency of traffic.
【作者单位】：吉林大学交通学院;长春工程学院机电工程学院;
【基金】：吉林省科技厅重点科技攻关资助项目(20150204052GX)
【分类号】：TP391.41;U463.6

【相似文献】