Automatic Corona Discharge Detection for Cable Safety Inspection

The smooth operation of today’s digitalized society is heavily reliant on the integrity of electrical infrastructure. Early detection of defects and accurate estimation of the remaining useful life (RUL) of electrical systems are therefore crucial. While various methods exist for diagnosing electrical infrastructure, such as infrared image inspection and partial discharge inspection, corona discharge diagnosis, is favored due to its non-contact and robust nature. In this study, we present a deep-learning based approach for diagnosing cable safety using corona discharge images captured by a UV camera. The images are classified based on their shape, and a deep-learning model is adapted to classify and localize these images. This approach has the potential to significantly improve the efficiency and accuracy of corona discharge diagnosis, serving as a reliable basis for inspector’s judgment and even potentially being implemented as a real-time inspection system.