Autonomous 3D Visual-Thermal Inspection for Utility-Scale Solar Farms

This paper presents Tricorder, an automated 3D visual-thermal inspection system for under-panel PV asset inspection. It combines LiDAR-inert, SLAM with geometrically calibrated RGB and thermal imagery, along with machine-learning-based object detection and segmentation. The system enables spatially fused and localized visual and thermal observations, demonstrating MC4 connector detection, thermal anomaly identification, and structural component detection in utility-scale PV data.