Dashcam-Enabled Deep Learning Applications for Airport Runway Pavement Distress Detection
-
2023-07-01
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Aug. 2022 to July 2023
-
Corporate Publisher:
-
Abstract:Pavement distress detection plays a vital role in ensuring the safety and longevity of runway infrastructure. This project presents a comprehensive approach to automate distress detection and geolocation on runway pavement using state-of-the-art deep learning techniques. A Faster R-CNN model is trained to accurately identify and classify various distress types, including longitudinal and transverse cracking, weathering, rutting, and depression. The developed model is deployed on a dataset of high-resolution dashcam images captured along the runway, allowing for real-time detection of distresses. Geolocation techniques are employed to accurately map the distresses onto the runway pavement in real-world coordinates. The system implementation and deployment are discussed, emphasizing the importance of a seamless integration into existing infrastructure. The developed distress detection system offers significant benefits to the Utah Department of Transportation (UDOT) by enabling proactive maintenance planning, optimizing resource allocation, and enhancing runway management capabilities. Future potential for advanced distress analysis, integration with other data sources, and continuous model improvement are also explored. The project showcases the potential of low-cost dashcam solutions combined with deep learning for efficient and cost-effective runway distress detection and management.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type: