A runway defect detection method of an airport FOD detection system
By deploying fixed-point cameras on airport runways and using an improved multi-scale grouped dilated convolution module, combined with quantized convolution technology, the problems of insufficient multi-scale feature extraction and missed detection in low-contrast environments in airport runway inspection have been solved, enabling accurate detection and real-time monitoring of minute defects.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING QIXING ZHILIAN TECHNOLOGY CO LTD
- Filing Date
- 2026-04-08
- Publication Date
- 2026-07-17
AI Technical Summary
Existing airport runway inspection technologies suffer from insufficient multi-scale feature extraction capabilities, easy omissions in low-contrast environments, and high computational demands on high-precision models, making real-time deployment difficult. Consequently, they struggle to achieve accurate detection and real-time monitoring of minute defects.
A road surface defect detection method based on computer vision is adopted. Image information is acquired by deploying fixed-point cameras, and preprocessing is performed by combining AOD-Net image dehazing algorithm and semantic segmentation technology. An improved multi-scale grouped dilated convolution module and quantized convolution technology are used to construct a road surface defect recognition network model to achieve efficient feature extraction and real-time analysis.
It achieves precise capture of minute defects, improves detection robustness in low-contrast environments, reduces computational latency, supports real-time processing of 4K high-definition video streams, and meets the 24-hour uninterrupted monitoring requirements of airport runways.
Smart Images
Figure CN122415464A_ABST