Research On In-position Detection Technology Of Straight Blade Grinding Precision Based On Machine Vision

In the field of aerospace manufacturing,ultrasonic cutting technology for honeycomb materials has been increasingly used due to its green and environmental advantages.As a core component of this special machining technology,the dimensional and form parameters of the tool have an important impact on the machining accuracy,machining efficiency,cutting performance and service life.In-situ inspection of the machining accuracy of the tool can effectively prevent the problems of repeated positioning accuracy and secondary clamping in secondary machining caused by disassembly,and provide technical support for subsequent online error compensation.This paper takes ultrasonic straight edge tool as the research object,applies machine vision technology to tool machining accuracy detection,carries out the research of tool machining accuracy detection method,and finally realizes in-situ detection of tool in CNC machining.The main work content and results of the thesis are as follows.1、 Design of machine vision based tool machining in-situ detection system.Combine the requirements of system detection accuracy,working distance and field of view,select the camera,lens and light source,and design the lighting scheme for the shape characteristics of the tool.Study the calibration method of vision inspection system,and calibrate and verify the camera.2.Acquisition technology of tool geometry and geometric accuracy based on contour graphics.In order to ensure that the tool is in the best position when the camera is shooting and to avoid the measurement error caused by the wrong tool position,study the deflection of the tool position,divide the tool position into two kinds of C-axis position adjustment and A-axis position adjustment according to the relative position of the camera and the tool,and establish the corresponding mathematical model for the deflection conditions of these two positions,use the model to calculate the deflection angle of the tool C-axis and A-axis,and adjust the tool position through the CNC program The tool posture is adjusted by CNC program.Based on the contour shape of the tool,the calculation methods of tool grinding length,edge width,thickness and symmetry are proposed.3.Research on the tool profile generation technology based on image processing technology.The image is pre-processed by using grayscale transformation,filtering processing and binarization method to reduce the noise effect of the image and segment the tool in the image.The image is processed by binary morphological operations to improve the quality of the segmented binary map by filling holes and eliminating isolated pixel points.Study the detection effect and localization accuracy of pixel-level and sub-pixel-level edge detection algorithms,compare the detection effect of different pixel-level detection operators,use Canny as the pixel-level edge detection algorithm,and perform sub-pixel edge detection using Zernike moments for the edges after pixel-level detection to improve the localization accuracy of edges,and perform simulation tests by MATLAB to verify the accuracy of edge detection.Study the image stitching method and propose a curvature polar pointbased image stitching algorithm for forming a complete tool profile based on the characteristics of tool edge smoothing and grayscale distribution pattern.4.Develop image-based tool geometry and geometric accuracy inspection software.Divide the software into three parts: image processing module,image display module,and dimension calculation module,and develop the measurement software using App Designer of MATLAB.Experimentally compare the results of image processing calculations to verify the accuracy of the in-situ inspection system algorithm and the reliability of the software.