Assessment of the Manufacturing Quality of Metallic Surface Finishes in Artificial Intelligence

Abstract

This chapter describes a prototype of a system that automatically checks the surface of precision ground metallic surfaces. It specifically focuses on bearing rolls. The authors study and test how light bounces off the surface using optical experiments. The aim is to figure out how to make things look sharp and bright, so it's easy to see the contrast between surfaces that are damaged and those that aren't. A new method for choosing a value to separate different parts of an image is presented. In addition, the text explores various search methods that go through a list one by one and have been made available to the public. These algorithms are used to find the most suitable group of traits for organizing or categorizing something. The text also looks at how much computer power these algorithms need. Finally, the authors display the results of sorting 540 blemished pictures.