基于双目视觉的工件三维重建

Three-dimensional reconstruction technology of workpiece based on binocular vision

为了提高样品识别的智能化水平,基于双目视觉技术,配合工业机器人,对具有特征轮廓属性的工件进行三维重建.三维重建算法采用轮廓边缘线拟合算法以实现工件边缘轮廓提取;针对多条轮廓线相交情况,采用点圆拟合理论确定特征点;为了实现左右对应特征点的准确匹配,采用异面直线公垂线最小距离算法,通过对比左右公垂线距离大小判断两条直线对应特征点是否为匹配点;当单次拍摄无法满足完整尺寸测量时,引入了具有重叠特征点的二次拍摄重建算法.对5组不同尺寸、不同形貌的工件开展三维重建实验,整体还原出工件的三维尺寸形貌,各工件的主要轮廓尺寸的误差标准差基本在±0.5 mm以内.

To realize intelligent sample recognition,based on binocular vision technology and industrial robots,we carry out three-dimensional reconstruction of workpiece with characteristic contour attributes.In the three-dimensional reconstruction algorithm,the contour edge line fitting algorithm to extract the contour of workpiece edge is adopted.In view of the intersection of multiple contour lines,the point circle fitting theory is used to determine the feature points.To achieve the accurate matching of left and right corresponding feature points,we use the minimum distance algorithm of common vertical line of different plane line to determine whether these two lines correspond to the feature points,by comparing the distance between the left and right common vertical lines.When a single shot cannot meet the requirement of full size measurement,the research of secondary shot reconstruction algorithm with overlapping feature points is introduced.Five groups of workpieces with different sizes and shapes are used to carry out three-dimensional reconstruction experiments,and the three-dimensional shape of the workpiece can be restored as a whole.The variance of the main contour dimensions of each workpiece lies within±0.5 mm.