AI Image Inspection System - Mirror Surface Inspection Process
By implementing the AI image inspection system, we achieved a testing accuracy of 97%! After full-line implementation, the number of inspection workers was reduced by 70%.
The inspection system DEEP INSPECTION improves the speed and accuracy of the entire system and reduces the burden of visual inspections. Rist has developed a mirror surface inspection system using Deep Learning technology for the product inspection process of Murakami Kaimeido Co., Ltd., which boasts a high market share in rearview mirrors. This system has achieved an improvement in inspection accuracy (from 60% to 97%) and reduced the burden on inspectors who previously relied on visual checks. Rist's approach utilizes a convolutional network for multi-class classification. They developed a unique "Deep Inspection Confidence" metric, achieving a test result accuracy of 97%. 【Issues with Conventional Detection Systems】 - Difficulty in quantifying image color tones, etc. - Ambiguity in standards due to human determination of thresholds - Inspection accuracy around 60% - Final confirmation by inspectors through visual checks 【Features】 - After full-line implementation, the number of inspection workers is reduced by 70% - With an eye on overseas expansion, the data obtained from inspection automation aims to optimize upstream processes as well. *For more details, please refer to the PDF document or feel free to contact us.
- Company:Rist
- Price:Other