Image Processing Technique Applied to Electrical Substations Based on Drones With Thermal Vision for Predictive Maintenance

Damage detection in electrical substations (ES) is vital to prevent breakdowns of devices. One of the main mechanisms of detecting failures and overloads in ES is the heat sources, which are concentrated in small areas with higher temperatures due to malfunctioning components or operating over from its current or voltage limits. Using drones equipped with a thermal camera is an exciting option to conduct technical inspections of ES to capture the zones with heat sources faster, safely, and efficiently. This paper proposes a practical and optimized procedure using thermal images based on open-source software and programming languages to detect and prevent malfunction or overload devices and then suggests a predictive maintenance system. A real ES is used as an example for testing and applying the proposal, and the results obtained from applying the digital image processing algorithms are provided to validate the proposed thermal image-detecting system.