边缘检测识别道路边界时消除标识线干扰的图像处理方法

Image Processing Method to Eliminate the Interference of Marking Lines When Identifying Road Boundaries by Edge Detection

针对通过对道路图像进行边缘检测来识别道路边界时道路中道路标识线的边缘特征对判断造成干扰的问题,提出了一种处理图像的方法.该方法基于边缘检测算法检测到的边缘对道路的灰度图像进行逐行处理,通过对该行中边缘通过的点附近的灰度值进行判断和修改来降低道路标识线的边缘特征,同时保留道路真实边界的边缘特征.在该方法基础上搭建了基于边缘检测的道路边界识别系统并对方法进行检验,使用130余张道路图片作为输入,验证了方法的有效性和实时性.实验结果表明,该方法能有效消除道路标识线对基于图像边缘检测的道路边界检测方法的干扰.

An image processing method of solving the problem of the interference caused by the edge features of road markings lines when recognizing road boundaries by edge detection was proposed.The results detected by the edge detection algorithm were used to process gray road image line by line.The gray values near the point where edges pass by the line were determined and modified to reduce the edge features of road marking lines while retaining the edge features of real boundaries of the road.Additionally,a road boundary recognition system based on edge detection was built to test the method.The validity and real-time performance of the proposed method were verified by using more than 130 road images as input.Experimental results show that the proposed method can effectively eliminate the interference of road marking lines on road boundary detection techniques based on edge detection.