Key points are not available for this paper at this time.
This study addresses the challenges of accurately capturing and localizing surface damage on large wind turbine blades using UAV-based inspection systems. We propose an integrated framework consisting of:(1) An advanced image feature extraction module that combines U-Net architecture with the Sobel operator for precise edge detection;(2) A robust sequence image stitching algorithm based on equal-width feature matching criteria. The methodology integrates a damage localization system via a Cartesian coordinate framework, achieved by fusing a damage recognition model with geometric transformation equations. These equations establish quantitative links between physical measurements and pixel dimensions, enabling a coordinate transformation pipeline to map defect locations from image pixels to actual blade dimensions. Experimental validation shows the proposed algorithm achieves localization accuracy within a 5% error margin across diverse simulated damage scenarios. This performance enables reliable quantitative analysis for engineering applications, representing a significant advancement in wind energy infrastructure inspection technology.
Fan et al. (Tue,) studied this question.