IOT and ML Based Multitasking Drone System for Solar Panel Fault Detection

Abstract

As per the increasing demand for solar energy has highlighted the need for efficient and automated maintenance solutions for large-scale solar farms, where manual inspection is time-consuming, labor-intensive, and prone to errors. To overcome limitations of traditional methods using fixed sensors and thermal cameras, this research proposes an IoT and ML-based drone system equipped with RGB and thermal cameras to autonomously survey solar panels and capture image data. The images are uploaded to cloud storage in real time and combined with an existing dataset to train a CNN model for accurate detection and classification of faults such as dust accumulation, cracks, hotspots, shading, and diode failure. A cloud-based model comparison identifies faults and provides their GPS location along with severity levels via a user-friendly dashboard, enabling timely maintenance and improved efficiency. This system reduces operational costs, enhances fault detection accuracy, supports predictive maintenance, and significantly improves the lifespan and performance of solar installations, with potential scalability to other sectors such as wind turbine inspection, agriculture, and infrastructure monitoring.