Real Time Mold Quality Inspection in Foundries using Image Processing Techniques

This paper inspects and analyzes the mold quality which rectifies the challenges faced by the metal casting foundries. Current quality control methods include, stopping the productivity to check the mold condition periodically leading to a production loss of about 5% a day. This requires manpower to check mold condition periodically. Using machine vision, we can capture the mold image in real time and analyze the quality of the mold and alert the operator on faulty mold and initiate cleaning procedures when necessary here by saving costs and improving productivity. The developed IoT inspection device consists of a camera which captures the mold images periodically whenever the mold is created and checks the mold condition and analyzes each mold’s quality by extracting features and comparing it to a reference image which is implemented by Raspberry Pi3.