基于视频解析的高速公路车辆测速研究  

Research on the Measurement of Vehicle Speed on Highways Based on Video Analysis

作　　者：强添纲[1] 林雨[1] 吴占红 邵平[1] 欧阳冬冬 

机构地区：[1]东北林业大学交通学院,哈尔滨150040 [2]富裕县友谊民族学校,黑龙江富裕161200

出　　处：《森林工程》2016年第1期68-71,共4页Forest Engineering

基　　金：黑龙江省博士后特别经费项目(LBH-TZ0501);国家自然科学基金项目(51405075)

摘　　要：为了精确地测量高速公路车辆的速度,降低高速公路车辆测速成本,文章提出了运动目标检测算法,并自主验证并开发出了一种基于视频解析的高速公路多车辆的测速软件。在该方法中,首先利用改进的背景差分法有效地提取前景目标;然后,利用"伪目标"消除方法,保留有效的运动目标;再次,通过卡尔曼跟踪算法获取有效目标的运动轨迹;最后,通过摄像机标定获取运动目标单位时间内行驶的实际距离。经过验证算法有效的提高了车辆检测的精度,有着较好的参考价值。软件实现了高速公路车辆超速预警及事故责任认定的智能化,使事故认定更加科学准确。软件为高速公路工作人员处理交通事故提供了辅助决策软件,提高了硬件设备的利用率并减少了人力,具有良好的经济效益和社会效益。In order to measure the speed of vehicles in expressways accurately and reduce highway speed vehicle costs, a testing algorithm for moving objects was put forward and the corresponding software for measuring the vehicle speed in highways was designed based on video analysis. Firstly, the detection method of background subtraction was used to effectively extract the prospect goal; then, the method of ＂phony＂ elimination was adopted to keep effective kinetic target. Next, the target trajectory was acquired effec- tively through Kalman tracking algorithm; finally, the camera was used to calibrate the actual distance per unit time for the moving tar- get. After verification, the accuracy of vehicle detection by the algorithm was effectively improved and the algorithm has a good refer- ence value. The software can realize the intelligence of highway speeding warning and accident responsibility confirmation and the i- dentification of accidents will be more scientific and accurate. For staff, the software has provided decision support to deal with traffic accidents and improve the utilization of hardware and decrease the use of manpower, which has good economic and social benefits.

关 键 词：智能交通 视频解析 测速 高速公路 

分 类 号：S714.8[农业科学—林学]