中厚钢板轮廓检测中线激光端点提取算法研究

Research on Medium Thickness Plate Laser Line Endpoint Extraction Algorithm of Contour Detection

提出了一种基于二次投影的边缘信息提取算法,并将这种算法应用于线激光端点的坐标提取.利用面阵CMOS相机获取入射在中厚板带钢表面的线激光图像,利用一次投影算法对预处理后图像进行处理,提取图像中激光线中心线的纵向位置坐标,并以它为基准,上下各取2行共5行作为提取图像中激光线端点位置坐标的目标行;然后,利用二次投影改进算法对这5行激光线分别进行处理,获取每行激光线端点的亚像素位置坐标,对该坐标值进行加权平均处理,所得值即为图像中激光线的端点位置坐标.所提出的二次投影改进算法不仅实现了目前中厚板形貌的非接触式测量,也提高了实际应用的检测精度.该算法光线端点检测精度可达0.003 5 pixel,在工程中具有较大的应用价值.

This paper presents an algorithm of edge information extraction based on secondary projection,and applies this algorithm to the coordinate extraction of line laser endpoints.First plane array CMOS camera is used to collect the incident in the medium plate steel strip line on the surface of the laser image,using the first projection algorithm to deal with image after preprocessing,to extract the image of the center line of the laser line longitudinal coordinates,based on the center line of the vertical coordinates 2 of up and down the line,a total of 5 as the laser line endpoint location to extract the image coordinates of the target row,recycle the second projection algorithm to deal with the five lines of laser line,respectively for each line on the subpixel location coordinates of laser line endpoint,the weighted average of the coordinates processing,as in the image and the average laser line endpoint location coordinates.In this paper,extraction of secondary projection in the improved algorithm not only has realized the current plate morphology of non-contact measurement,also improve the detection precision that up to 0.0035 pixel of the practical application,hasing great practical value in engineering.