AI-Based Cap Inspection with Instance Segmentation

This test video shows how we use artificial intelligence to inspect the caps of effervescent tablet bottles.

The technology behind it is instance segmentation: the model does not just classify the whole image, it finds and outlines each cap individually, so position, shape, and defects can be evaluated per piece even when several products are visible at once.

Fast-moving, rotationally symmetric products like caps are a classic inspection challenge – and a good demonstration of how modern segmentation models handle tasks where classic blob analysis becomes fragile. The same approach carries over to closures, seals, and completeness checks across pharma, food, and packaging lines.