Dynamic Multilayer Virtual Lattice Layer (DMVL2) for Vehicle Detection in Diverse Surveillance Videos

View Sample PDF

Author(s): Manipriya Sankaranarayanan (National Institute of Technology, Tiruchirappalli, India)
Copyright: 2025
Pages: 28
Source title: Modern Advancements in Surveillance Systems and Technologies
Source Author(s)/Editor(s): Dina Darwish (Ahram Canadian University, Egypt)
DOI: 10.4018/979-8-3693-6996-8.ch007

Keywords: Information Science Reference / Security & Forensics / Surveillance & Monitoring / Surveillance Systems

Purchase

View Dynamic Multilayer Virtual Lattice Layer (DMVL2) for Vehicle Detection in Diverse Surveillance Videos on the publisher's website for pricing and purchasing information.

Abstract

Traffic surveillance videos play a critical role in various applications, from traffic flow analysis to incident detection. However, the variability in video quality and conditions poses significant challenges for developing robust algorithms to accurately enumerate traffic parameters. This work introduces the Dynamic Multilayer Virtual Lattice Layer (DMVL2), a novel framework designed to address these challenges. DMVL2 adapts to diverse video conditions by using multiple virtual lattice layers, ensuring accurate extraction of traffic parameters regardless of video variability. The framework also utilizes parallel processing to handle multiple videos efficiently and integrates adaptive enhancement techniques to adjust for varying illumination levels. Experimental results demonstrate that DMVL2 significantly improves detection accuracy, achieving 92.5% in urban traffic scenarios and reducing false positives to 4.2%. The framework outperforms traditional methods and deep learning approaches, proving its robustness and reliability in diverse traffic environments.

The IRMA Community

Research IRM

Dynamic Multilayer Virtual Lattice Layer (DMVL2) for Vehicle Detection in Diverse Surveillance Videos

Purchase

Abstract

Related Content

IRMA Sponsors