TUNNEL ACCIDENT DECTECTION USING YOLOv7
Year: 2025
Type: article
Abstract: This project Detecting accidents in tunnels is critical for ensuring the safety of commuters and maintaining theefficiency of transportation systems. This project proposes a deep learning-based approach utilizing YOLOv7(You Only Look Once version 7) to identify accidents within tunnel environments. The methodology beginswith importing essential libraries and acquiring a specialized dataset comprising tunnel accident images. Thedataset undergoes thorough preprocessing to ensure consistency and quality, followed by a structured split intotraining, validation, and testing sets. This division is crucial to train the model effectively, fine-tune itsparameters, and objectively evaluate its performance. The YOLOv7 model, known for its real time objectdetection capabilities, is selected and trained on the prepared dataset, using provided annotations or boundingboxes to accurately localize and classify accident scenarios within the images. Throughout the training process,progress is carefully monitored and logged to identify potential areas for improvement and ensure the model'soptimal performance. Once trained, the model is evaluated using key performance metrics such as accuracy,precision, and recall, providing a comprehensive understanding of its ability to detect tunnel accidents reliably.The results demonstrate the promising potential of deep learning, particularly YOLOv7, in enhancing accidentdetection systems within tunnel environments.
Authors: Dr. B. NANDITHA, Arupali Sri Vaishnavi, Gudupally Shiva Sai Reddy, Kethavath Vasantha, Kandibanda Gayathri
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W6893268117
FWCI:
Cites:
Cited by:
Related to: