占道车辆移动抓拍技术的研究.pdf

占道车辆移动抓拍技术研究 摘 要 快速公交系统是解决城市交通拥挤问题的有效手段，但是社会车辆侵占快 速公交专用车道的现象普遍存在，严重影响了快速公交的效率。本文采用车载 摄像机对前方车辆进行判别，自动抓拍违规占道车辆的图像，本文研究成果具 有重要的应用价值。主要工作如下： 1、车道检测与判别。传统 Hough 变换车道线检测方法的效果较好，但算 法较为复杂，实时性不能满足实际需要。为了提高其实时性，本文提出了一种 改进的Hough 变换车道线检测算法。通过限制检测区域，筛选边缘点，极大减 小了算法的计算量，实现对车道线的实时检测。另外通过 GPS 与车道线特征检 测相结合方式有效实现对车道线的判别。 2 、车辆检测与车型判别。本文提出车辆尾灯的检测与车牌定位算法相结合 的方式实现对车辆检测。相对于车辆检测最常用的车底阴影等检测方法，该算 法受环境影响更小。最后通过车牌颜色特征判别是否为公交车辆。 实验结果表明，算法具有较好的实时性和鲁棒性。 关键词：车道线检测 车辆检测 Hough变换 尾灯检测 车牌定位 Research on Mobile Capture Technology of Occupied Lane Vehicles ABSTRACT Bus Rapid Transit system is an effective means to solve the problem of urban traffic congestion . However, due to the widespread social vehicles occupation of bus rapid transit lanes, seriously affecting the efficiency of the bus rapid transit. In paper automatically by the on-board equipment discriminant moving vehicles ahead and capture images of cars occupy lanes is true. The results of this study has important application value. The main tasks are as follows: 1. lane detection and discrimination. The traditional Hough transform traffic lane detection method is better, but the algorithm is more complex, Real -time performance cant satisfy the actual needs. In order to improve the real -time, this paper proposes a about the lane line detection Hough transform algorithm is improved. Achieve the real -time detection of lane line by dramatically reducing the amount of calculation of the algorithm, due to the limited test area and screening of edge points . In addition to the lane line discriminant by combining GPS with lane line feature detection . 2. Vehicle detec