基于机器视觉的刀具磨损特征值获取方法

Tool Wear Eigenvalue Acquisition Method Based on Machine Vision

为在生产加工过程中有效检测刀具磨损状态,提高加工效率,降低加工成本,提出一种基于机器视觉的刀具磨损特征值获取方法与检测装置。基于加工场合开发了一款具备回收与展开两种状态的在位刀具图像采集装置;运用灰度处理、高斯模糊滤波与二维中值滤波、自适应二值化、形态学等图像处理技术对刀具图像进行预处理,结合GrabCut算法分割图像背景,运用OTSU分割算法提取刀具轮廓,采用Canny算子边缘检测提取刀具磨损区域,结合刀具直径计算得出刀具磨损特征值;通过对6061铝合金航发叶片进行正交试验,获取刀具磨损特征测量值,与电子数码显微镜获得的实际值对比表明,测量值与实际值之间的误差(除少部分外)均保持在0.02mm内。

In order to effectively detect the tool wear status in the production process,improve the processing efficiency and reduce the processing cost,a machine vision based tool wear eigenvalue acquisition method and detection device are proposed.According to the requirements of the processing situation,a in place tool image acquisition device with two states of recycling and unfolding is developed.Image processing technologies such as gray-scale processing,Gaussian blur filtering,two-dimensional median filtering,adaptive binarization and morphology are used to preprocess the tool image,combine GrabCut algorithm to segment the image background,use OTSU segmentation algorithm to extract the tool contour,use Canny operator edge detection to extract the tool wear area,and then combine the tool diameter to calculate the tool wear characteristic value.Through the orthogonal test of 6061 aluminum alloy aero engine blade,the measured value of tool wear characteristics is obtained,and compared with the actual value obtained by the electronic digital microscope,the results show that the error between the measured value and the actual value is kept within 0.02mm except for a small part.