激光条纹亚像素中心精确提取方法  

An accurate extraction method for subpixel center of laser fringe

作　　者：张伟康 甄国涌 单彦虎 储成群 ZHANG Weikang;ZHEN Guoyong;SHAN Yanhu;CHU Chengqun(School of Instrument and Electronics,North University of China,Taiyuan 030051,China)

机构地区：[1]中北大学仪器与电子学院,太原030051

出　　处：《激光杂志》2023年第3期75-80,共6页Laser Journal

基　　金：国家自然科学基金面上项目(No:62075199)。

摘　　要：利用激光三角法对工件进行三维测量时,激光条纹中心位置的提取是实现高精度测量的重要因素之一。为了能够更好地取投射在工件形状突变处的激光条纹信息,通过改进最大类间方差法在激光条纹图像中选取多阈值以实现二值化分割。所提算法基于最小二乘法获取激光条纹骨架上的法向,避免了大量卷积计算,从而提高运行速度,最后在其法向上利用灰度重心法得到激光条纹中心的亚像素坐标点。实验结果表明:该算法在激光条纹中心提取过程中,相较于经典的Steger算法,在运行速度上提高近6倍,分辨率上提高近16%,在工业上适用于在线精确测量。When the laser triangulation method is used for three-dimensional measurement of workpiece,the extraction of the center position of laser stripe is one of the important factors to realize high-precision measurement.In order to better extract the laser stripe information projected at the sudden change of workpiece shape,multi threshold is selected in the laser stripe image by improving the maximum interclass variance method to realize binary segmentation.The proposed algorithm obtains the normal direction of the laser stripe skeleton based on the least square method,avoids a large number of convolution calculations,and improves the running speed.Finally,the sub-pixel coordinates of the center of the laser stripe are obtained by using the gray center of gravity method in its normal direction.The experimental results show that in the process of laser fringe center extraction,compared with the classical Steger algorithm,the running speed of the algorithm is improved by nearly 6 times and the resolution is improved by nearly 16%.It is suitable for on-line accurate measurement in industry.

关 键 词：三维扫描 中心提取 OTSU 骨架细化 灰度重心 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]