Deep transfer learning method for aero-engine blade encoded character detection and recognition based on YOLO-GE

Key Points

• The aim is to enhance character detection and recognition for aero-engine blade labeling using a deep learning approach.
• Utilized a deep transfer learning framework based on YOLO-GE.
• Focused on character detection and recognition specific to aero-engine blades.
• Implemented machine learning techniques to optimize accuracy.
• Achieved significant improvement in detection accuracy compared to previous methods.
• Demonstrated robustness in handling various blade characters.
• Reduced processing time for character recognition tasks.