Development of an embedded system with machine vision for quality inspection and process automation

Abstract

Today manufacturing market is getting more competitive with cutting-edge technologies and efficient manufacturing processes. Process automation is one of the key areas where industries can give more emphasis to face this challenge. Embedded system with machine vision can play a significant role in quality assurance and process automation. Besides process automation, right line stop policy will enable the industries to reduce the costs related to defective items and initiate maintenance task for improving systems and processes. This work presents the design, development and performance evaluation of a prototype system with automated color error detection and line stop facility by using low-cost hardware available in the local market. An open sourced Linux operating system, interfaced with the camera module has been used in this work. The sorting process depends on capturing the image of the product and analyzing the image by comparing with the predefined color to discard the defective products. A servo motor driven actuator is used to discard the defective items from the conveyor line. A program based on a maximum number of allowable defective items on conveyor line is developed to stop the production automatically. The system continuously monitors the process to reduce undesirable production of defective items. The result from our test shows above 98% success rate of defective item detection and system stoppage because of continuous production of defective items. Due consideration and attention are given to drive color based quality inspection efficiently and keep the cost low while designing this embedded system with machine vision. The study will provide the guidance for adopting color error detection based system to sort defective items and implement line stop policy in many industrial fields for continuous evaluation of product quality.