Aircraft Surface Defect Detection Based on Improved YOLOv8

In aircraft maintenance, automated detection of aircraft surface defects is critical to ensure flight safety. In this paper, an aircraft surface defect detection method based on improved YOLOv8 is proposed. In order to enhance the feature extraction capability and global context awareness of the model, this paper introduces the CoTAttention module on the basis of YOLOv8, and adopts the SPD-Conv convolution operation for replacing the C 2 F convolutional layer. The experimental results show that the improved YOLOv8 outperforms the baseline model in terms of detection accuracy and recall for three types of defects, namely Crack, Dent and Rust. These improvements validate the effectiveness of the proposed method in complex defect detection tasks.