在机高精度刀具直线轮廓异物过滤算法研究

Research on Foreign Object Filtering Algorithm of High Precision Tool Linear Contour on Machine

为解决在机高精度刀具图片测量时刀具轮廓粘有少量异物的问题,本文提出了一种基于机器视觉的刀具直线轮廓异物过滤算法。以铣刀轮廓异物过滤为例,首先提取轮廓像素点,然后选取适当大小的ROI区域,滑动窗口对路径上的每个ROI区域进行处理,在每个ROI区域中,使用凸包检测算法对区域进行迭代检测以提高去除率。实验结果表明,对于(5~30)像素内的异物,用该方法进行轮廓异物过滤,去除率高达95%,使后续的刀具测量精度有明显提升。

In order to solve the problem of a small amount of foreign objects sticking to the tool contour when measuring the high-precision tool image on the machine,this paper proposes a machine vision-based tool linear contour foreign object filtering algorithm.Taking the milling cutter contour foreign object filtering as an example,the contour pixels are first extracted.Then,select the ROI area of appropriate size,and slide the window to process each ROI area on the path.In each ROI area,the convex hull detection algorithm is used to iteratively detect the area to improve the removal rate.The experimental results show that for foreign objects within(5~30)pixels,the removal rate of this contour foreign matter filtering method is as high as 95%,which significantly improves the accuracy of subsequent tool measurement.