Stamped appearance parts dread a single scratch or shallow dent—small and low-contrast, exactly what manual eyes miss most. A body-shop plant used DaoAI AI-AOI to push the escape rate down to two per thousand.
Stamped sheet-metal parts are mostly visible outer panels, where any scratch, dent or impression directly affects vehicle appearance quality. At this plant such defects are often only a few tenths of a millimeter, with extremely low contrast on reflective metal surfaces. Manual inspection tires easily and misses defects, with a measured escape rate as high as 2.8%. Once defective parts flowed downstream or to the customer, rework and claim costs were high and the reputation of appearance parts took a direct hit.
Seeing a 0.3mm Scratch Clearly
DaoAI AI-AOI surface inspection optimizes imaging and lighting for reflective metal surfaces, pairing them with a deep-learning model that recognizes scratches, dents, impressions and pinholes, reliably catching fine defects at the 0.3mm class. Iterated continuously on on-site samples, the model holds a high detection rate while effectively suppressing false calls, reaching 97% overall accuracy and enabling full, part-by-part judgment of appearance parts.
- 0.3mm-class detection: fine scratches and shallow dents reliably caught
- 97% accuracy: high detection with low false calls, reducing needless rework
- Escape rate 2.8% → 0.2%: customer-side outflow risk sharply reduced
- Metal-surface optimized: imaging and lighting tailored to reflective outer panels
Cutting the escape rate from 2.8% to 0.2% puts the reputation of appearance parts back in your own hands.
After go-live, the plant achieved full surface inspection of stamped appearance parts, the customer-side escape rate fell from 2.8% to 0.2%, appearance complaints and claims dropped markedly, and both quality stability and delivery confidence rose in step.