Image Recognition of Wind Turbines Blade Surface Defects Based on Mask-RCNN

Wind turbine blade failure will reduce power generation efficiency and increase operating costs. Serious faults will lead to production accidents. In this paper, a method of blade fault diagnosis based on deep learning algorithm is proposed. Mask-rcnn model is used to identify the defects in blade images, and the method is verified by the pictures of unmanned aerial vehicle (UAV) patrolling blades. Good recognition results are obtained. Most of the blade defects can be identified and the phenomenon of misidentification is very few. It can be used to upgrade wind farm power generation. Efficiency and daily operation and maintenance provide reliable information support.