滚珠螺母的机器视觉检测

Ball nut detection by machine vision

提出了基于机器视觉检测技术的滚珠丝杠副螺母的非接触测量方法。将用A102FCCD数字摄像头采集的图像经过IEEE1394数字接口卡传输到计算机。对原始数字图像实施预处理、二值化、轮廓提取及细化等处理后,将图像信息转变成宏观图形信息。在尺寸测量时,借助于图像处理中提取的边界,使用自编程序测量和计算螺母的螺距、珠心径、球面度误差、圆柱度误差及滚道锥度,其测量误差分别为0.001、0.023、0.010、0.016及0.006 mm。分析了误差产生的原因,指出光学误差是影响测量精度的主要因素。测量结果满足产品的技术要求。理论分析及实验结果表明了该方法正确可行。

On the basis of machine vision, a the non-contact detection method was put forward to measure the geometrical size and shape-position error of the ball nut. The image was collected by a digital camera A102FCCD, and was input into computer by a digital interface card IEEE1394. The original gray level image was changed into the macroscopical drawing information by pre-processing, threshold selection, image bivalency, edge detection and contour extraction and segmentation. The geometrical parameters were calculated and measured by the aid of the contour extracted from the image processing and a self-compiled measurement program. The ball nut thread pitch, radius of the ball center circle, the sphericity deviation, the cylindricity deviation, and ball-way taper were measured and their errors were 0. 001, 0. 023 , 0. 010, 0.016 and 0.006 mm respectively, The reason for error was analyzed, and concluded that the optic error is the main factor affecting the measurement precision. The results satisfy the technique requirement of the product. Theoretical analysis and experimental results show that this method is feasible and correct.