论文题目：ASD-YOLO: An Aircraft Surface Defects Detection Method Using Deformable Convolution and Attention Mechanism

论文摘要:

Aircraft surface defect (ASD) detection is crucial to ensure flight safety. Aiming at the problems of defects detection with large scale variations, irregular shapes, and sample imbalance, this paper proposes a network ASD-YOLO, which is based on the YOLOv5 and proposes multiple improvements to enhance the recognition capability. First, a new deformable convolutional feature extraction module DCNC3 is designed to better learn defects of different shapes, which is combined with a global attention mechanism GAM to pay more attention to defects region information. Secondly, the feature representation of defects for small targets is enhanced by the contextual enhancement module CEM. Finally, the exponential sliding average weight function EMA-Slide is introduced to solve the sample imbalance problem. The experimental results on two datasets show that the mAP is improved by 5.7% and 3.4% , which is better than the mainstream algorithms, and provides a novel solution for the ASD detection task.