Flycatcher

Safe, reliable, cost-effective AI-driven inspections

Problem

Traditional inspection and monitoring approaches for infrastructures, industrial plants, fleets, and sites rely on large datasets to define all possible issues and malfunctions and then monitor them. These methods are limited, inefficient, costly, and unable to detect new, unforeseen problems or anomalies in dynamic and evolving systems.

Solution

Flycatcher disrupts inspections by utilising AI-driven anomaly detection models, focusing solely on normal operational behaviour. The solution requires minimal data, integrates with existing low-cost sensors, and highlights deviations in real-time, improving efficiency without extensive historical data.

Core technology

Flycatcher technology is built upon Unsupervised Anomaly Detection, applicable to any sensory data. It requires small sets of normal (day-to-day) data and is optimised to run on various platforms, from Cloud to hardware like NVIDIA Jetson, enabling efficient and real-time detection of both known and unexpected anomalies in structured and unstructured environments.