Benefits
For the first time in the production sector, an artificial intelligence algorithm capable of detecting even unique errors that could occur for the first time was developed by Yeditek.
The system developed by Yeditek achieved a 100% accuracy rate in quality control within seconds, significantly increasing efficiency in the quality control process.
Quality control processes were carried out in 2.5 seconds per part, without exceeding the 4-second cycle time of the production line.
No defective samples were created during the commissioning and artificial intelligence training process. The artificial intelligence was trained using only flawless parts.
Flawless parts were automatically sorted into the boxing system, and defective parts were separated into the quarantine box with the integrated cobot.
The system replaced human visual quality controls on production lines, thus preventing defective products from reaching the customer.
Problem
YPS Automotive, one of Turkey's leading automotive parts manufacturers, faced challenges in ensuring the quality control of parts produced at the end of press lines through visual inspection, leading to missed defects. The production's 4-second cycle time was insufficient for conducting a detailed inspection.
Solution
For the first time in Turkey, an artificial intelligence-supported anomaly detection application was successfully implemented in press production processes. Without creating any defective samples during the commissioning and AI training phase, the system developed by Yeditek, based on artificial intelligence and image processing, used only flawless parts. It quickly and accurately inspected all parts in production, raising quality standards. The system managed to perform quality control without exceeding the production's 4-second cycle time, successfully enhancing the efficiency of quality control.