Jinghe Tian

A two-stage segmentation model based on improved Deeplabv3 + is proposed for segmenting Blade regions and defects in complex environments. Accurate segmentation is critical for timely maintenance and safe operation of wind turbines. The model consists of Blade-Deeplabv3 + and Defect-Deeplabv3+, collectively named BD-Deeplabv3+. In the first stage, Blade-Deeplabv3 + segments the Blade from the background using the Atrous Spatial Pyramid Pooling module to extract multi-scale features and suppress background interference. The resulting segmented Blade is then input to the second stage. In this stage, Defect-Deeplabv3 + extracts multi-scale features and refines boundaries of surface crack, hole, and spalling defects. DenseASPP replaces the original ASPP, employing densely connected dilated convolutions to enhance multi-scale feature fusion and improve semantic representation and boundary accuracy for minor defects. Experimental results show that the mean intersection over union for Blade segmentation reaches 98.97%, and for defect segmentation reaches 94.25%. Finally, Blade defect severity is quantified using the ratio of defect area to Blade area, enabling more reliable maintenance planning.

Timely and accurate detection of wind turbine blade surface defects is crucial for ensuring operational safety and improving maintenance efficiency with respect to large-scale wind farms. However, existing methods often suffer from poor generalization, background interference, and inadequate real-time performance. To overcome these limitations, we developed an end-to-end defect recognition framework, structured as a three-stage process: blade localization using YOLOv5, robust feature extraction via the large vision model DINOv2, and defect classification using a Stochastic Configuration Network (SCN). Unlike conventional CNN-based approaches, the use of DINOv2 significantly improves the capability for representation under complex textures. The experimental results reveal that the proposed method achieved a classification accuracy of 97.8% and an average inference time of 19.65 ms per image, satisfying real-time requirements. Compared to traditional methods, this framework provides a more scalable, accurate, and efficient solution for the intelligent inspection and maintenance of wind turbine blades.