基于双目线结构光的三维测量系统研究

Research on 3D Measurement System Based on Binocular Line Structured Light

针对物体三维重建过程中,传统的双目立体视觉测量法和激光三角测量法存在的局限性,设计了一套双目线激光扫描式三维重建系统。首先,对双目线激光扫描测量系统的结构组成和数学模型进行分析,并推导出在双目对极约束下激光中心线三维点坐标的计算方法;针对传统算法在光条曲率变化较大的区域提取中心线效果不理想的问题,提出了一种基于改进灰度重心法的线结构光中心线提取算法;最后,进行中心线提取精度和效率对比实验,以及三维重建精度实验。实验结果表明,改进算法能够精确有效地提取光条曲率变化较大区域的中心线,平均提取精度达到1.462 pixels,平均提取速度为0.0734 s。同时,设计的结构可以准确地获取物体表面丰富的三维点云数据,测量误差在1 mm以内。

Aiming at the limitations of traditional binocular stereo vision measurement method and laser triangulation method in the process of three-dimensional reconstruction of objects,a set of binocular line laser scanning three-dimensional reconstruction system is designed.Firstly,the structural composition and mathematical model of the binocular line laser scanning measurement system are analyzed,and the calculation method of the three-dimensional point coordinates of the laser center line under the binocular epipolar constraint is derived.Aiming at the problem that the traditional algorithm is not ideal for extracting the center line in the area where the curvature of the light stripe changes greatly,a line structured light center line extraction algorithm based on the improved gray center of gravity method is proposed.Finally,the centerline extraction accuracy and efficiency comparison experiment,and the three-dimensional reconstruction accuracy experiment are carried out.The experimental results show that the improved algorithm can accurately and effectively extract the center line of the area where the curvature of the light stripe changes greatly.The average extraction accuracy reaches 1.462 pixels,and the average extraction speed is 0.0734 s.At the same time,the designed structure can accurately obtain the rich 3D point cloud data of the object surface,and the measurement error is within 1 mm.