基于圆结构光视觉三维点的孔洞缺陷识别及重构

Detection of pipe holes and their reconstruction based on circle-structured light

基于圆结构光视觉检测方法获取的管道内表面三维数据点以圆周形式存在的特点,本文提出了距离约束条件来检测孔洞的方法,即如果相邻两个圆周上对应的相邻两点之间的距离d均小于某一个阈值d',则该处视为非孔洞,否则视为孔洞,并提出了新的三角网格化方法对带孔洞点云数据进行三角剖分且保留出孔洞。最后,基于VC++软件平台利用OpenGL实现管道孔洞缺陷检测和三维曲面重构,实验结果表明该管道孔洞缺陷检测方法检测速度快、易于实现、具有较好的可视化效果。

Based on machine vision, three-dimensional data for the inner surface of a pipe were obtained by circlestructured light in the form of a circumference, and hence a distance constraint has been proposed in order to detect holes. That is, on two adjacent circumferences, if the distances between two corresponding adjacent points are all smaller than a certain threshold, then we view this place as a non-hole; otherwise, we view it as a hole. A new triangulation method was also proposed to triangulate the point cloud data with holes, and to retain the holes. Finally, OpenGL was used for surface reconstruction to achieve three-dimensional visualization. The experimental results show that this method is fast and easy to implement, and it has a good visual effect.