【摘要】 通过现有车流量统计方法的分析,提出一种新颖的基于轨迹分析和车辆识别的车流量统计方法。该方法首先从视频中提取运动目标,并对运动目标的轨迹进行保存,然后利用支持向量机对车辆轨迹和目标进行学习,最后利用得到的分类器来统计真实轨迹和真实车辆的数量。实验结果表明,该方法在全天候环境下仍然能够保持较高的准确率和更好的鲁棒性。更多还原

【Abstract】 After analyzing the existed traffic flow statistic methods,a novel flow statistic method based on track analysis and vehicle recognition is proposed. In this proposed method,the moving objects are first extracted out from the video and a collection is used to save these tracks and moving objects,then the support vector machine is utilized to learn the real vehicle tracks and objects. Finally,the real number of tracks and vehicles can be got by means of the classifiers of support vector machine. The experiments show that the proposed method is able to keep higher accuracy and better robustness under different environments.更多还原