A Review Paper on Real-Time Video Analysis in Dense Environment for Surveillance System

Abstract

Dense environmental conditions such as snow, fog, lightning, heavy rain, and darkness drastically lower the quality of outdoor surveillance videos. The primary functions of video surveillance systems in crowded environments have received significant attention, particularly in detection, categorization, and event or object recognition. The methods and algorithms for real-time video analysis in various weather conditions have also significantly advanced with the advancement of technology. Examples include background extraction, the see-through algorithm, deep learning models, CNN for nighttime intrusions, the System for high-quality underwater Monitoring using optical-wireless video surveillance, the low-visibility enhancement network (LVENet), edge computing, and many others. Using various elements of these methodologies, the current research increased monitoring performance and avoided potential human failures. In-depth information about these video surveillance methods, systems, and supporting details is provided in this study. An overview of employed construction and architectural styles is given, and the critical assessments of these systems are then covered. Current surveillance systems and various methods for achieving accuracy in real-time video analysis in adverse weather circumstances are contrasted in terms of their features, benefits, and challenges, which are discussed in this paper, to provide a complete image and a broad view of the System. Future trends are also highlighted, pointing to the study that will be conducted in the future.