Abstract
Insulators are critical components in power transmission systems, ensuring electrical stability and preventing current leakage to supporting structures. Faults in these components can cause power outages, equipment failures, and safety hazards. Traditional fault detection methods, such as manual visual inspection, are time-consuming, error-prone, and inefficient for large-scale grid monitoring. This systematic review explores recent advancements in power insulator fault detection using unmanned aerial vehicles (UAVs) integrated with advanced imaging systems and machine learning (ML) techniques. UAVs equipped with binocular vision and high-resolution cameras enable multi-angle, high-fidelity imaging of insulators in remote and hazardous environments. The review examines various ML and deep learning approaches, particularly convolutional neural networks (CNNs), for detecting cracks, contamination, and surface anomalies in aerial imagery. It also addresses key limitations, such as the lack of annotated datasets, weak model generalization under variable conditions, and challenges in real-time deployment due to computational constraints. A comparative analysis of existing techniques is presented, highlighting accuracy, scalability, and application readiness. Finally, the study identifies future research opportunities, including lightweight model design, multi-sensor data fusion, and explainable AI integration. The goal is to enhance fault detection reliability, reduce operational costs, and promote the intelligent maintenance of power transmission infrastructure.