Real-Time Surface Defect Detection on Automated High-Volume Smart Device Component Manufacturing Lines



High-volume production lines for mobile device LED covers, process hundreds of parts per minute and quality inspection must occur in parallel. To enable cost-effective inspection of LED covers at high scale, this mobile device OEM turned to DarwinAI.

The Problem

The existing process suffered from rising costs and limited scale, due to escape rates, lead time, and a labor shortage.

DarwinAI developed and deployed a vision-based AI on ~100 edge inspection devices to detect and diagnose defects in real time.

The Results

annual savings per machine, from consolidating personnel
annual savings per machine, due to 50% lower escape and overkill rates
payback period
1 month

Difficulties with the existing inspection process

  • High Escape Rates
  • High Overkill Rates
  • Shortage of Inspectors

Alternatives couldn’t fulfill the requirements

Achieving high detection accuracy at high velocity was essential (stoppage is estimated to cost $10k per minute).

The manufacturer looked at a number of automation systems on the market, but found them to be imprecise and slow.

They needed a solution that would provide inspectors with root cause analysis—rather than simply flagging if a component had a defect—which was beyond the capability of available alternatives.

DarwinAI’s Solution

DarwinAI developed and deployed an edge-based quality control system that visually examines LED covers to detect and diagnose defects including cracks, scratches, and debris. 

As an additional long-term benefit, the inspection data collected from the edge can be analyzed to identify failure points earlier—allowing the manufacturer to optimize the quality control process.

By delivering the accuracy and scale needed, with fewer human operators, DarwinAI’s solution generated enormous savings—and delivered complete project payback in its first month of operation.

Contact us to learn more about this case study.