Automated gear inspection using Image Processing and Machine Learning Techniques

Abstract

The creation of a robotic transmission inspection system that harnesses the power of image processing and artificial intelligence systems is presented in this report This system consists of components that have been developed to perform three tasks, which are the fault diagnosis, sorting, and wear evaluation of gear teeth. The automatic fault diagnosis module use image processing to find the problems like cracks, chips, and wear at the surface of gear teeth. The sorting module has an imaging system that analyzes the gear images and moves the gears further to their own subgroup with the use of a conveyor belt and a robotic arm. The inspection algorithm takes advantage of image processing and machine learning in the analytic tasks of assessing the surface texture and wear zones of gear teeth. The system consists of the modules without direct gear measurement, profile measurement as well as analysis of wear via the neural networks. Different types of gears are being tested using image processing and machine learning to assess their condition and establish the credibility of the system in detecting gear faults, sorting gears, and analyzing gear wear.