A Vision-Based Inspection Strategy for Large-Scale Photovoltaic Farms Using an Autonomous UAV

Abstract

The complex operational environment of photovoltaic farms has brought direct challenges to the conventional manual-style inspection. Small-scale quad-rotor unmanned aerial vehicles (UAVs) have been widely used in many civil applications in virtue of their obvious advantages. A typical UAV based inspection system aims to obtain infrared and visible images of polar panels to detect the failures; usually, the process of acquiring the images relies on manual work or autonomous flight control according to previously obtained waypoints, and the disadvantages of the above methods are inefficiency and inaccuracy respectively. In this paper, a vision-based image acquisition strategy is presented without predefined task points. Firstly, single PV string's edges are acquired by processing video frames in real time; then the flight direction and offset for the UAV can be calculated; through the velocity control algorithm, automatic tracking for PV arrays can be achieved. The proposed solution is evaluated through experiments and the numerical results demonstrated the effectiveness of the suggested approach.