摘要
针对车辆压线的违章行为,提出一种基于计算机视觉的车辆压线检测方法。首先利用Hough变换对交通视频图像中的黄线实现自动提取;然后用背景差法从图像中提取前景车辆,并用大津法将车辆分割出来,从而获得车辆位置;最后根据黄线区域和车辆区域是否有重叠来判断车辆是否压线。实验结果表明,该方法能满足准确高效的要求。
Aiming at detecting the violation behavior of vehicle's line pressing, this paper proposes a method of vehicle's pressing line detection based on computer vision. First, the Hough transform method is applied to extract the yellow lines from traffic video images automatically; second, the background difference method is used to extract vehicles from images, and the otsu method segments the vehicles out, thus getting vehicle's position; Finally, the overlap of the yellow line area and vehicle area can determine the vehicle's line pressing. Experimental results dem- onstrate that this method can meet the requirements of accuracy and efficiency.
出处
《电子科技》
2013年第2期71-73,共3页
Electronic Science and Technology
关键词
图像处理
直线提取
车辆检测
image processing
line extraction
vehicle detection