基于视觉检测的自动化精确寻边技术

Automatic Accurate Box Boundary Detection Technology Based on Visual Inspection

针对箱体蒙皮配制中的自动化和高精度需求,提出一种基于视觉检测的箱体自动化精确寻边方法。根据箱体边界特征设计基于Hough变换的方法,实现对所需边界点的提取。将图像处理得到的边界点恢复为三维点云,对不同位置拍照处理得到的边界点云进行合并、旋转投影得到二维平面内的边界点。通过曲线拟合最终得到完整有序的边界点云。利用箱体扫描平台和激光跟踪仪设计相应的试验,实验证明了该方法的可行性和准确性。

To meet the requirements of automation and high precision in the preparation of box skins,an automatic and precise boundary detection method based on visual inspection is proposed.According to the boundary feature of the box,the method based on Hough transform is designed to extract the required boundary points.The boundary points obtained by image processing are restored to three-dimensional point clouds,and the boundary point clouds attained by photographing at different positions are merged and rotated to gain the boundary points in the two-dimensional plane.A complete and orderly boundary point cloud is finally obtained through curve fitting.The corresponding test is designed by using the box scanning platform and laser tracker,and the feasibility and accuracy of the designed method are verified.