结合双向光流约束的特征点匹配车辆跟踪方法  

Vehicle Tracking Method Using Feature Point Matching Combined With Bidirectional Optical Flow

作　　者：卢胜男[1] 李小和[1] 

机构地区：[1]西安石油大学计算机学院,西安710065

出　　处：《交通运输系统工程与信息》2017年第4期76-82,共7页Journal of Transportation Systems Engineering and Information Technology

基　　金：国家自然科学基金(61572083)~~

摘　　要：针对复杂交通场景中动态光照变化、目标尺度变化和部分遮挡等因素带来的影响,提出了一种基于特征点的稳定可靠的车辆跟踪方法.针对运动车辆高速行驶时具有较大帧间运动的特点,构造KLT算法的金字塔模型,根据前向和后向跟踪偏移量,对稳定性较差的特征点进行剔除.同时,采用SURF特征匹配算法对目标特征点集进行更新和校正.最后,利用特征点之间的位置信息,确定目标的尺度和旋转变化因子,从而实现当前帧中目标区域的定位.实验结果表明,提出的车辆跟踪方法可以有效地解决复杂场景中目标形变和部分遮挡等问题,对尺度和旋转变化也具有较强的鲁棒性.To handle the dynamic illumination changes, scale variation and partial occlusion in complex traffic scene, a stable and reliable vehicle tracking method based on feature point is proposed. With the characteristics of large interframe motion of moving vehicle, the pyramid model of KLT algorithm is constructed and the feature points with poor stability will be deleted according to forward and backward tracking offset. Meanwhile, SURF algorithm is used as a compensation mechanism for updating and adjusting the feature point sets. Then, according to the relative location and relative angle of feature points in the first frame, the scale and rotation variation of objects in the current frame is determined. Finally, the object region is determined in the current frame. The experimental results show that the proposed method effectively solves the problems of object deformation and partial occlusion, and it is robust to scale and rotation variation.

关 键 词：智能交通 车辆跟踪 特征点提取 光流法 SURF特征 KTL算法 

分 类 号：U495[交通运输工程—交通运输规划与管理]