Medical Balloon Inspection
Multiple images of the medical balloon are captured with a 2D camera. Specialized optics, filters and lighting components are selected to generate repeatable inspection images. The model locates, measures and classifies defects found in the balloon. Results are communicated to the automation system and images are stored for audit.
The solution has improved repeatability, reliability and efficiency when compared to a human operator. It also executes faster than a typical visual inspector, leading to shorter cycle time and lower operating costs.
AI Pipeline:
Image ↦ Background separation↦ Scaling↦ Object detection↦ Position, Size and Classification