In order to reduce errors and omissions in manual inspections of civil aviation aircraft before takeoff

enhance inspection quality and efficiency while reducing labor intensity

this paper proposes a deep learning-based augmented reality intelligent inspection method for civil aviation aircraft. Firstly

a data augmentation method based on pre augmentation evaluation was designed

which achieved large-scale automatic augmentation of a small number of civil aviation aircraft damage defect image sample datasets. Subsequently

focusing on the visual characteristics of damage defects

and improved YOLOv8 network is proposed to train the augmented dataset for damage defect detection

forming a damage and defect detection model. Finally

this method is integrated into the augmented reality recognition and display process

utilizing augmented reality glasses to achieve intelligent identification of aircraft damage and defects and augmented reality display and maintenance guidance for the identification results. The proposed method is validated on real-world scenarios

showing effective identification of common defects with a detection rate increased from 89.1% to 95.7%

and a maximum reduction in inspection time of 27.0%

thereby effectively assisting inspection personal in achieving intelligent inspection of civil aviation aircraft.