Drone Based Outdoor Insulator Inspection

Abstract

Over the past few decades, interest in unmanned aerial vehicles (UAVs) and in particular quadcopters has increased due to the wide range of possible research applications that can benefit from the use of quadcopters. Insulator inspection on overhead power lines has traditionally relied heavily on visual inspection. The task is both cumbersome and relies on the experience of the inspector. It is also extremely dangerous as the inspector needs to work in close proximity with overhead power lines, and contact with these lines can lead to instant death. This thesis focuses on the development of a quadcopter based system that is able to inspect insulators on overhead power lines. The proposed system consists of a quadcopter that is able to inspect the health of insulators on an overhead power line. The quadcopter, via the help of its onboard cameras and Raspberry Pi based computer, is able to detect the health of an overhead power line insulator and simultaneously send images to the ground station where they are processed. The main contribution of this thesis is the development of a complete quadcopter based system for overhead power line insulator inspection. The offshore image processing algorithm presented in this thesis has a mean average precision of 0.66 and an average processing time of 0.55 seconds. The onboard image processing algorithm has a mean average precision of 0.26 and an average processing time of 1.28 seconds. These numbers show that quadcopter based insulator inspection can be carried out successfully using both offshore and onboard image processing techniques both in terms of precision and image processing time.