United States Department of Transportation
ROSA P serves as an archival repository of USDOT-published products including scientific findings, journal articles, guidelines, recommendations, or other information authored or co-authored by USDOT or funded partners.
As a repository, ROSA P retains documents in their original published format to ensure public access to scientific information.
i
Developing a Portable Railroad Crossing Monitoring System Based on Artificial Intelligence and Image Processing Technology
-
2024-08-01
-
[PDF-2.43 MB]
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Resource Type:
-
Geographical Coverage:
-
Edition:Final Report, September 2023-August 2024
-
Corporate Publisher:
-
Abstract:This project addresses the urgent need for improved railroad grade crossing safety, where traditional systems often fail to detect and manage unforeseen hazards such as vehicles or pedestrians obstructing tracks. This project develops a cost-effective, portable system for real-time identification, counting, and categorization of objects at railroad grade crossings, including vehicles and pedestrians. The system integrates a cost-effective camera and an edge-computing device, capable of performing object detection and classification in the railroad crossing area. A field-testable prototype has been assembled and proven effective in detecting track intrusion. Utilizing a specialized deep neural network and edge computing platform, the system could enhance collision warnings and inform traffic management and urban planning. Tested on the CDnet 2014 dataset, it achieves an F-measure of 87.67%, surpassing state-of-the-art models in foreground detection. Its deployment on a multi-core inference pipeline reduces latency from 120.82 ms to 49.63 ms and increases the frame rate from 8.28 to 20 FPS, ensuring real-time performance. Field testing validates its ability to identify track intrusions and enhance safety, representing a significant step toward proactive traffic management and reduced trespassing risks at railroad crossings. This approach could shift railroad crossing monitoring from passive to proactive traffic management, aiding urban development and reducing trespassing risks.
-
Format:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type:
