基于多尺度特征融合的飞行器总装质量检测方法及系统
By improving the YOLOv7 network and multi-scale feature fusion, the shortcomings of digitalization and intelligence in aircraft final assembly quality inspection have been solved, achieving efficient and accurate inspection results and meeting the inspection needs under complex and small sample conditions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHANDONG UNIV
- Filing Date
- 2026-05-15
- Publication Date
- 2026-07-17
AI Technical Summary
Current aircraft assembly quality inspection relies on manual visual inspection, which lacks sufficient digitalization and intelligence, making it difficult to achieve high-precision inspection. Furthermore, the YOLO network suffers from missed detections and false positives in small target detection, and its model has poor generalization ability, failing to meet the inspection requirements under high complexity and small sample conditions.
The YOLOv7 network is improved by adopting a multi-scale feature fusion method. By segmenting high-resolution images, adding feature layers, introducing transfer learning and adaptive prior boxes, a multi-module detection system is constructed to achieve comprehensive, accurate, efficient and intelligent detection of the overall assembly quality of the aircraft.
It improves detection efficiency and accuracy, meets the needs of complex, high-precision and mass production, realizes more comprehensive assembly quality inspection, and reduces the rate of missed detection and false detection.
Smart Images
Figure CN122223016B_ABST