Research On End Mill Wear Detection Based On Machine Vision

In the milling process,there is inevitably end mill wear.when the wear and tear is not timely feedback to the staff,it will be impossible to meet the accuracy requirements for a large number of parts,and may lead to a large number of parts being incomplete within the specified time,which make work efficiency reduce and the progress of work be affected.In this paper,the wear of the end milling cutter is detected by computer during the end milling cutter change.According to the test results,we can judge whether the replacement end milling cutter can continue to be used in time.Through the analysis of the various forms of wear and the various testing methods of the wear of the end mill,the testing method suitable for the processing conditions in this article is selected,and according to the blunt standard of the tool given in the international standard,milling the judge of the eligibility.In this paper,it is necessary to calibrate the camera because it is necessary to study the wear image of the milling cutter.In the process of calibration,the camera parameters calculated from multiple images are summarized and optimized by genetic algorithm.Finally,the optimal distortion parameters of the camera are obtained,and the image correction is realized.In the pre-processing of wear images,Considering vibration noise and other environment of the machine tool,it is necessary to preprocess the collected images to prevent the influence on the calculation of the later eigenvalues.The gray level and denoising process are selected.In the process of noise processing,the method of adaptive median filtering is selected,which can not only remove the noise well,but also protect the image information well.After the de-noising image is still some information is lost,and the latter by enhanced image processing,which makes the image denois and the original image information be well protected.When extracting the eigenvalues of each wear quantity of the image,it is necessary to detect the edge of the image after preprocessing,to compare the first and second order edge detection operators and to select the optimal Canny operator.Canny operator is used to extract the wear image of the end milling cutter,the edge of the end milling cutter without noise is obtained.However,this method is only suitable for edge detection of whole pixels,but not for subpixel edge detection with high accuracy.In this paper,the former operator of Zernike moment is improved,the template is amplified.And through the experimental verification of each operator,the rationality of the operator is fully verified.Since a single feature can not fully reflect the wear characteristics of the end milling cutter in the past tool detection,this paper puts forward the qualification judgment of multiple features.Finally,through a large number of experiments,the rationality of the multi-feature system is fully proved.