| Summary: | 碩士 === 國立成功大學 === 工程科學系碩博士班 === 100 === In the machine tool industry, high precision and automation are required for manufacturing and assembly processes. The accuracy of components fabricated by a machine tool is lower than that of the machine itself. Scraping technology plays an important role in eliminating the accumulated tolerance of products of a machine assembled from products of the first generation of the machine. Scraping work-pieces are hand-made, and the wear detection of the scraper tool depends on the human operator. Standard scraper tool wear detection is becoming increasingly important. In order to study the monitor of machine scraping tools condition in machining based on machine vision, the machine vision on the state of a scraper tool wear is therefore designed. The present study has developed a standard detection system for a scraper tool wear that uses image processing technology to detect the scraping pattern of surface textures from a work piece. The measurement results are analyzed by using wavelet technology and Otsu's method. The preprocessed images are decomposed under the use of wavelet technology, whereby the sub-image contain rich intermediate frequency information are obtained, and then the sub-image is used Otsu's method to automatically perform histogram shape-based image thresholding, the characteristics reflecting the scraping tool wear condition is extracted by calculating wavelet technological energy distribution of the images. The results has concluded that the method of judging the scraper tool wear condition through extracting the wavelet energy distribution of machine surface texture is simple, and the scraper tool wear condition can be determined in the way.
|
|---|
