How accurately do inspection robots monitor industrial facilities?
High-end sensors enable inspection robots to perceive beyond a human’s natural perception, which allows the robots to collect very accurate, high-quality data. They cover hundreds of inspection points per mission and can consistently conduct multiple, precise tasks at any time of the day. The data quality depends on the robot’s data capturing, contextualization, processing, and quality optimization capabilities.
Capturing data: The robot’s ability to consistently locate and access all predefined and ad hoc inspection points and reliably capture high-quality raw sensor data from the same viewpoint.
Data contextualization and fusion: The robot’s ability to associate data with time, location, viewpoint and create additional insights through the fusion of extrinsically calibrated sensors (e.g., visual, thermal, and acoustic measurements, gas leak detection, and reality capture).
Inspection intelligence: The robot’s ability to firstly, verify the state of instrumentation and equipment (e.g., manometers, gauges, counters, LCDs, levers, and valves) and, secondly, through AI-based inspection algorithms, detect anomalies and turn information into insights.
Quality optimization: The robot’s ability to assess data through edge computing and to make adjustments for poor quality. This happens through object recognition, correctly framing equipment (even after movement), unwarping perspectives, removing background objects, and reacting to visibility issues such as occlusion or glare.