硅棒特征点三维坐标视觉检测技术研究

Study on 3D coordinate detecting technology for characteristic pionts of silicon rod

针对人工接触式硅棒参数检测中存在的耗时、低效、误检等问题,提出一种采用四CCD机器视觉系统作为硅棒参数实时检测的非接触式测量方案。系统基于多目视觉检测原理,以四CCD多目视觉检系统作为机器视觉检测前端数据采集模块,拍摄待测硅棒表面轮廓图像信息。相邻两CCD构成正交双目交汇数据采集子模块,子模块基于正交双目交汇测量原理实时采集两路含有目标空间信息的数字图像数据并保存到上位机。上位机图像处理系统读取采集图像,完成多目相机标定,特征点提取、同名点匹配以及视差计算后提取出目标特征点三维坐标。该检测技术有效避免了人工主观因素造成的漏检、错检,满足实时在线检测要求,实现了硅棒特征点空间三维坐标高精度检测的目的。

As manual contacted detection of the parameters for silicon rod is time-consuming, low efficient and easy to cause detection errors, a non-contacted detection plan of machine vision system using 4 CCDs is proposed. As the front-end data acquisition module of the whole machine vision system, the 4 CCDs vision system is based on multi-camera vision measuring principle to take pictures of silicon rodrs surface profile. Any adjacent two CCDs vision systems connect and form an orthogonal binocular intersection sub-data acquisition module, which is based on orthogonal binocular intersection measuring principle to collect two-way real-time digital image data containing the target information and save them to the host computer. Then the image processing system in the host computer starts to read the collected images. After the multi-camera calibration, characteristic points extraction, the corresponding point matching and the parallax computation were finished, the three-dimensional coordinates of the target feature point can be extracted. This detection technology can effectively avoid detection missing, failing and errors caused by subjective factors in manual detection. It also can meet the requirements of real-time online testing, and achieve the high accuracy detection purpose ofthree-dimensional space coordinates for the characteristic points of the silicon rods.