Composites are the material of the future – lightweight, strong and capable of performance that metals can only dream of. Yet ironically, when it comes to inspecting these cutting-edge materials, many manufacturers are still relying on one of the oldest tools in the factory: human eyesight. Walk through a composites production line and you will still find skilled inspectors running their hands along surfaces, peering at fibres under fluorescent lights and making judgment calls on whether a part passes or fails. It is careful work. It is intuitive. It is also, increasingly, a bottleneck.
The complexity that keeps automation out
Let’s be fair – composites are not easy to inspect. A defect in a metal part might show up as a dent or a scratch. In a composite, it might hide in the weave, whisper through a wrinkle or sit just beneath the surface. And then there is the geometry. Curves, corners, layers – each part can feel more like a sculpture than a standardised product.
This is exactly why the industry has held on to manual inspection for so long. Automation has tried to join the party, but conventional vision systems tend to get confused by texture, glare or natural variation in fibre placement. In other words, they do not know the difference between character and flaw. But that is starting to change.
AI that does not just see – it understands
The next wave of inspection tech is not about rules and templates – it is about understanding. Today’s best systems do not just detect anomalies; they learn what a “normal” part should look like, even if that part is never exactly the same twice.
This is the space where Artificial Intelligence (AI), computer vision and composites finally meet on equal footing. At the forefront is a platform called Spectron™, developed by Zetamotion. Instead of drowning users in data labelling or demanding massive defect libraries, Spectron creates its own curated synthetic training data from a single scan.
It is not magic, it is semantic learning. The system is trained to interpret the semantics, i.e. the deeper principles and meaning behind a part’s nature and the defects it may exhibit. That means it can distinguish between acceptable variation and a genuine defect, even when surfaces are non-uniform or slightly imperfect by nature. So yes, the material might be complex. But now, the inspection system can be intelligent enough to handle that complexity.
photo: Factory 2035: a self-training, semantic-learning scanner casts a live holographic defect map over the composite arc, while an engineer in AR glasses tunes the algorithm – symbolising the leap from hand-feel inspections to real-time, AI-orchestrated quality control
Lu public network security: 37140202000173
