Paper Title
Advancements in Automatic Wind Turbine Blade Inspection Using Drones

Abstract
As pressure mounts on the energy sector to achieve renewable production, wind turbine shave proven to be a cost-effective solution. Governmental funding and research has contributed to the rapid growth of installations to meet energy production needs. However these large structures are subject to harsh environments and operating stressors which, overtime, cause faults, malfunctions, and even total failure. Our focus here is on the blades of these structures. To combat this, routine inspection is of the utmost importance for preventative maintenance. This study provides a comparative analysis of two detection architectures in You Only Look Once (YOLO) version 9 and Mask Regional-based Convolutional Neural Network (R- CNN) for the localization of wind turbine blade faults and their efficacy in automating this procedure. Utilizing a small-scale wind turbine for dataset creation, the results show a promising trend with YOLOv9 C achieving the highest mAP50 of 0.849 followed by Mask R-CNN achieving the mAP50 of 0.8372. Keywords: Autonomous Inspection, Wind Turbine Blades, Fault analysis, Deep Learning