Engraved Character Recognition of Train Wheelset Based on the Total Least Square Method

摘要

According to the characteristics of depth information of the engraved characters and the lack of color difference to the background, a line structured light sensor was applied to acquire train wheelset engraved character point cloud information. The total least square method was proposed for a 3D character point cloud processing method, in which the observation vector error and the coefficient matrix error are taken into account. According to the standard deviation of the point cloud data, abnormal points and engraved character points can be eliminated for plane fitting. A high-quality outline of the character information can be extracted from the character point cloud depth data.