Design and Fabrication of Smart Inspection System for Industrial Products Using ML Algorithm

Industrial 4.0 is revolutionizing the world including robotics, nanotechnologies etc., Still detecting defects on parts has a significant challenge to solve. The manufacturing factories and distribution centres have complex technologies for inspection of the products and packaging process. Especially the food and clothing industries have to deeply investigate the defects. Conventional inspection systems implement physical inspection techniques to find defects in the product which may have a chance to affect the final product. In some test methods only relatively non-porous surfaces can be inspected and in some cases, this can be affected by variations in magnetic permeability. A well-achieved visual inspection provides a reduction in the amount of repairs and overcomes the above-mentioned problems in the inspection system. By simply presenting the inspection system with several model images, deep learning enables it to learn to detect surface anomalies. The device which works on the concept of Deep Learning based image analysis for factory automation is our proposed theme. To train the machine, sample image sets are given. This can be made customized as data sheets are given to the software. Hence, the proposed system can be customizable according to the product dimensions.