New Developments in Automated Vision-Based Methods for Power Line Inspection

Inspecting electric utility assets is vital for ensuring the electricity supply's reliability, availability, and sustainability. Traditionally, these inspections have relied on labor-intensive methods such as foot patrols and helicopter-assisted surveys, which are slow and costly and pose significant safety risks. However, with advancements in technology, particularly in ar-tificial intelligence (AI) and machine learning (ML), there has been a significant shift towards automating these inspections. Integrating unmanned aerial vehicles (UAVs) equipped with advanced imaging sensors and deep learning algorithms has shown great potential for improving the efficiency and accuracy of power line inspections. Despite these advancements, significant challenges remain in achieving comprehensive automated power line inspections. Recent developments in generative and zero-shot learning AI models offer promising opportunities to address these challenges and expedite the development of automated inspection techniques across a wide range of power line inspection use cases. This paper reviews the power line inspection process and highlights the use of deep learning computer vision techniques and existing challenges. We explore some future opportunities to leverage the latest developments in multimodal large language models and zero-shot learning for defect detection, segmentation, and synthetic data generation.