Research On Cutting Tool Wear Detection Based On Computer Vision

Cutting tool wear detection technology is very important in enhancing automatically manufacture efficiency, cutting tool utilization and manufacturing quality, and cost-saving. During various cutting tool wear detection methodologies, image processing technology based on computer vision has the obvious advantages on convenience, fast-process, non-contact and accuracy in detection. This thesis has deeply investigated the key technology of image processing for cutting tool wear detection, which has certain guidance in cutting tool wear detection.The thesis has analyzed the characteristics of cutting tool wear condition and standard of abrasion , focused on cutting tool flank wear image, carried out image preprocessing,edge detection,threshold segment and other techniques on the original wear image in line with the characteristics of wear image result from laboratory and relevant graph treatment techniques, and calculated the wear value of cutting tools.Through a series image preprocessing, such as image clipping, enhancing and median filtering, etc, Otsu threshold segment method based on GA to obtain the optimized Th has been adopted for the original images in the thesis, and finally come to a clear cutting tool wear edge with the edge detection of Th as the best value of canny-edge detection. Moreover, a more accurate wearing edge has been reached through sub-pixel edge detection method based on grey moment,space moment and Zernike moment. An optimized algorithm has been arrived by the comparison of the three calculation methods in the experiments. Finally, the chain code edge searching algorithm has been adopted to describe the wear area edges, wearing values of flank of cutting tool has been further calculated with different algorithms. With the comparison of the actual value and the calculated value, the result shows that, the wear value obtained by image processing is in line with the actual value. On the other hand, if the wear value is exceed the abrasion standard value, and then the cutting tool has to be changed, otherwise, keep it in use.In conclusion, the algorithms in the thesis are deeply studied and the system of cutting condition detection has been well developed. This system can realize some functions, such as image reading, preprocessing, edge detection, wear measurement. The thesis surly has certain guidance in cutting tool wear detection based on computer vision.