研究目的

To inspect the tool wear condition in the process of numerical control (NC) machining for difficult-to-cut material.

研究成果

An online tool wear inspection system was developed based on machine vision. The system was equipped with a CMOS camera, a bi-telecentric lens, and other devices. The self-matching algorithm was developed to inspect tool wear condition automatically, which was integrated into an independent software with a user-friendly GUI. The proposed method of image acquisition was efficient and easy to operate. All bottom edges were acquired at one time. Tool wear condition was measured by serving the worn tool itself as the reference object rather than by serving a new tool image as a template or by acquiring a different degree of worn images to have machine learning. The improved PCA rotation was compared with the improved Radon rotation. The former method was proved to be more effective and adaptable. The error of the maximum wear value reduced mostly by 0.042 mm using the former method. The absolute values of errors on the maximum wear width were less than 0.007 mm using the self-matching algorithm. In the case of severe wear or breakage, the absolute values of errors on the maximum wear width were less than 0.057 mm because of the uneven reflected light. A double octal ring light could improve the results.

研究不足

The severe worn area is rough and causes uneven reflected light. A double octal ring light could reduce the effect of uneven reflected light.