Skip to main content

High-Speed Unsupervised Anomaly Detection Demo

I built a high-speed industrial anomaly detection demo to show Emergent Vision’s software capabilities at the 2025 inVISION Days digital conference.

In this fun example, a high-speed EROS 10 GigE camera captures images of PCBs moving rapidly on a rotary stage. Each image is processed by an unsupervised anomaly detection neural network to determine whether it is “normal” or “anomalous”, i.e. has a defect, such as a scratch or a foreign object. If a PCB is identified as “anomalous”, an LED immediately illuminates it with a red X. All this happens in a fraction of a second!

Demo

You can watch the webinar, presented by Emergent Vision’s president John Ilett, here or here.