Downloads: 1 | Views: 97 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1
Research Paper | Computer Science and Information Technology | United States of America | Volume 13 Issue 3, March 2024 | Popularity: 4.7 / 10
AI-Powered Traffic Violation Detection Using CCTV Footage
Omkar Reddy Polu
Abstract: Urban congestion, accidents and fatalities are seriously impacted by traffic violations. Manual enforcement of traditional systems or the use of automated mechanisms to enforce such systems leave technology behind the violations these systems detect. On the basis of continuous video surveillance using CCTV, this paper proposes an AI powered Traffic Violation Detection System, which is able to automatically detect infractions like jump signal, over speed, lane violations, unauthorized movement of vehicle using deep learning based computer vision models. The system integrates YOLO based object detection for violations detection, transformer based action recognition to recognize different actions as well as spatiotemporal anomaly detection to prevent the violation tracking from noise. Furthermore, we propose an edge - AI approach that uses lightweight deep learning models on low power edge devices, thereby reducing dependence on centralized processing and enabling faster inference. Also included in the proposed architecture is the inclusion of blockchain powered storage for evidence against legal process to ensure that no one can ulterior tamper the record keeping. A federal learning framework provides model robustness between cities while maintaining privacy of sensitive vehicular information. At the end, it is tested against the real world urban datasets and is proven to have better accuracy and efficiency than existing techniques. The work in this thesis proposes a privacy preserving, legally compliant, and scalable framework for next - generation intelligent traffic monitoring systems that use AI for automation of modern urban mobility.
Keywords: AI - powered surveillance, traffic violation detection, deep learning, edge AI, blockchain security, YOLO object detection, federated learning, computer vision, urban mobility, smart city solutions
Edition: Volume 13 Issue 3, March 2024
Pages: 1956 - 1961
DOI: https://www.doi.org/10.21275/SR24038113847
Please Disable the Pop-Up Blocker of Web Browser
Verification Code will appear in 2 Seconds ... Wait