Infrared Computer Vision for Utility-Scale Photovoltaic Array Inspection

TL;DR

This paper reviews infrared thermography and computer vision techniques for detecting faults in utility-scale solar panels, introducing a dataset and evaluating image processing methods for thermal anomaly detection.

Contribution

It presents a new infrared thermography dataset for PV arrays and evaluates computer vision algorithms for thermal fault detection in solar panels.

Findings

Infrared thermography effectively detects PV panel faults.

Hough Transform improves PV segmentation accuracy.

Benchmark dataset supports future research in thermal anomaly detection.

Abstract

Utility-scale solar arrays require specialized inspection methods for detecting faulty panels. Photovoltaic (PV) panel faults caused by weather, ground leakage, circuit issues, temperature, environment, age, and other damage can take many forms but often symptomatically exhibit temperature differences. Included is a mini survey to review these common faults and PV array fault detection approaches. Among these, infrared thermography cameras are a powerful tool for improving solar panel inspection in the field. These can be combined with other technologies, including image processing and machine learning. This position paper examines several computer vision algorithms that automate thermal anomaly detection in infrared imagery. We demonstrate our infrared thermography data collection approach, the PV thermal imagery benchmark dataset, and the measured performance of image processing transformations, including the Hough Transform for PV segmentation. The results of this implementation are presented with a discussion of future work.