In view of the difficulties in urban road zebra crossing measurement and the drawbacks of existing point cloud data identification and extraction methods,the study optimizes the zebra crossing corner extraction algorithm based on the vehicle-mounted laser scanning point cloud data,and uses the vehicle-mounted scanning system point cloud scanning line to separate the zebra crossing point cloud with OTSU algorithm.The frequency distribution histogram of"number of scanning lines-horizontal distance"and moving discrimination window are respectively used to extract the short and short edges of zebra crossings combined with RANSAC algorithm.Finally,the corner coordinates of zebra crossings are precisely extracted based on the geometric relationship between the short and long edges of zebra crossings,and the accuracy of algorithm extraction is analyzed and verified by actual engineering projects.By comparing and analyzing the results of zebra crossing corner extraction with the measured data,the reliability of the zebra crossing corner extraction algorithm is comprehensively verified through the mean error and distance root mean square error,which provides a new technical method for the rapid acquisition and updating of road feature lines,such as zebra crossing and marking lines on urban operating roads.
关键词
点云数据/斑马线角点/直线拟合/扫描线
Key words
Point cloud data/Zebra crossing corners/Linear fitting/Scanning line
维普
万方数据