公路收费站横杆状态的视频检测方法及应用

以高速公路收费站监控视频为背景，提出一种可自动分析和检测横杆运行状态的方法，从而判断车辆通过与否并统计车辆数。该方法的实现，首先对横杆的颜色特征和轮廓几何特征进行学习；然后对监控视频图像进行颜色和几何特征匹配以滤除非横杆区域；再对得到的疑似横杆区域进行修复；最后计算横杆的重心坐标及其运动状态以判断开闭状态。样本测试显示，新方法能有效克服视频场景中各种干扰因素的影响，提高车辆计数的准确性。

Method and Applications of Video Detection for Toll-gate Cross-bar State

Based on the detection video of the highway toll-gate, anovel video-based-automatic cross-bar state detection method was proposed. The method can count the number of whether the vechile pass or does not. To achieve feature of the cross-bar should be learned the goal, secondly firstly, the color feature and geometric the suspect cross-bar region is estimated by color matching and geometric matching from of the cross-bar are estimated by calculating each frame; finally, the position and state the position of the cross-bar center and its movement. Experiments shown that the new method can decrease the false detections and increases the accuracy of the vehicle counting system effectively.