By using visual anomaly detection, it is possible to train ML models with images from ideal or normal product only. If an anomaly is found, the location of the anomaly can also be determined, so that you can immediately know which part is damaged or missing. Speed, efficiency, accuracy and ease of deployment are why we use FOMO-AD in this project.
This project uses Edge Impulse's FOMO-AD (Faster Objects, More Objects - Anomaly Detection) learning block based on GMM (Gaussian Mixture Model) which clusters data points with similar characteristics, then will compare and provide an anomaly value for which a threshold can be determined. So that it can be found which parts are considered anomalies. In this ML model, the resulting anomaly value also contains the coordinates, the with a specific program we can change this to which part are broken or not complete. The setting is an electronic manufacture process in the conveyor belt with a sorting system or an LCD display for information on anomalies or missing parts. Hardware: Sony Spresense with camera