【作者】 佘永业； 李熙莹； 赵有婷；

【Author】 She Yongye,Li Xiying,Zhao Youting (Research Centre of Intelligent Transportation System, Sun Yat-sen University,Guangdong Provincial Key Laboratory of ITS,Guangzhou 510275)

【机构】 中山大学智能交通研究中心广东省智能交通系统重点实验室；

【摘要】 本文提出一种在整个图像中计算车辆宏观光流速度的算法。算法包括提取关键特征点和计算宏观光流速度两步。第一步提取关键特征点,必须满足强运动点和Harris角点两个特征。第二步首先计算光流主方向,去除不满足主方向的特征点;然后采用Lucas-Kanade金字塔算法计算筛选后的特征点的光流速度值,把其平均值作为车辆宏观光流速度值。实验结果表明,该算法能有效检测图像中的车辆特征点,并计算宏观光流速度。更多还原

【Abstract】 The paper presents a new algorithm of calculating the whole optical flow speed of vehicle in the whole image.The algorithm includes two steps.Feature extraction and whole optical flow speed calculation.In the first step,strong motion pixels selected by Horn-Schunck algorithm and corner pixel acquired by Harris algorithm are the two conditions of the feature extraction.In the second step,the optical flow main direction is calculated to exclude the features which not meet the main direction.Finally,we get the optical flow speed of each feature through Pyramidal Lucas-Kanade optical flow algorithm and put the mean value as the whole optical flow speed.The test shows the algorithm can detect the features in the whole image and calculate the whole optical flow speed effectively.更多还原