基于夜间车辆检测的智能矩阵车灯控制方法

Intelligent matrix lighting control method based on night vehicle detection

文章提出一种智能矩阵车灯控制方法。其系统使用自身车灯的明暗截止线进行图像分区,采用掩模阈值分割操作,同时提取汽车尾灯和汽车前灯后进行判断识别,该方法能够有效识别同向行驶车辆和相向行驶车辆,从而自动控制矩阵车灯。实验结果表明,该方法计算量小,适合多种行车场景,在“树莓派3 B+单片机”下平均每帧耗时89 m s,综合环境下场景测试识别率为88%,车灯控制准确率为96%。

The article proposes an intelligent matrix vehicle lamp control method.The system uses the cut-off line of its own light to partition the image,uses the mask threshold segmentation operation to extract the tail light and the front light at the same time,and then judges and recognizes them.This method can effectively identify the vehicles in the same direction and the vehicles in the opposite direction,so as to automatically control the matrix light.Experimental results show that this method has a small amount of computation and is suitable for various driving scenes.It takes an average of 89 ms per frame under the“Raspberry PI 3B+microcomputer”,and the scene test and recognition rate under the comprehensive environment is 88%,and the vehicle lamp control accuracy is 96%.