基于车辆之间实际距离的快速交通状况检测算法

FAST TRAFFIC SITUATION DETECTION ALGORITHM BASED ON ACTUAL DISTANCE BETWEEN VEHICLES

交通状况检测与预警是交通信息检测系统中一个很重要的组成部分,它对城市交通管控和诱导有着重要的指导意义。基于视频图像的灰度和边缘特征,首先采用差分算法计算出各车道中轴线上像素点的灰度梯度,并对梯度函数结合"像素-距离"的映射关系,按距离聚类分析法完成其对应曲线上脉冲的合并与数值量化;然后根据量化结果,对各车道的有车区域和无车区域做出标识,并对有车区域的分布密度做出统计;最后,结合对车道中轴线上像素点的灰度帧差处理结果,对交通状况做出精确判断。经实验验证,该算法能对道路的交通状况做出准确、实时的判断,且算法简单稳定,具有很好的实用价值。

Traffic situation detection and early warning is an important part in traffic information detection system,it has significant guidance meaning to urban traffic control and inducement. Based on grayscale and edge features of video images,in the article we first use difference algorithm to calculate the grayscale gradient of pixel points in middle line of each lane,and then for gradient function,according to distance clustering analysis method we complete the aggregation of pulses on its corresponding curves and the quantification of number in combination with mapping relationship between pixel and distance; the next,according to the results of quantification we mark the regions with and without vehicles on each lane,and make the statistics on the distribution density of the regions with vehicles. Finally,combining the processing results of grayscale frame differences for pixel points on middle line of lane,we give precise judgement on the traffic situations. It is verified by experiment that the algorithm can make accurate and real-time judgment for traffic situations on road,in addition,the algorithm is simple and stable,and has good practical value.