engleski

An Innovative Vision System for Automatic Tool Wear Classification

An innovative approach for automatic tool wear classification based on comparison between images of the tool in unworn and worn condition is presented. The algorithm is based on multiple image comparison methods such as chromatic comparison and statistical correlation. The algorithm parameters were optimized using genetic algorithms and training images. In order to validate the approach, the methods were tested on over 600 test images and its reliability and accuracy were detrmined. From the data analysis, it emerged that this approach is not suitable for reliable tool wear measuremnt, however it can be effectively applied for tool wear classification.

TCM; Vision system; Tool wear classification

nije evidentirano