Anomaly detection using edge computing in video surveillance system: review.

Area of Science:

• Computer Science
• Artificial Intelligence
• Urban Planning

Background:

• Smart cities aim for modern, secure, and sustainable infrastructure to improve residents' quality of life.
• Video surveillance is crucial for enhancing public safety and well-being in urban environments.
• Detecting abnormal events in surveillance video remains a significant challenge, demanding extensive human analysis.

Purpose of the Study:

• To provide a comprehensive overview of anomaly detection methodologies in intelligent video surveillance.
• To systematically categorize existing approaches and identify key objects-of-interest and datasets.
• To explore the integration of edge computing for real-time anomaly detection in surveillance systems.

Main Methods:

• Literature review and survey of anomaly detection techniques over the past decade.
• Systematic categorization of anomaly detection methodologies based on context and application.
• Exploration of edge device capabilities and specific algorithms for real-time surveillance.

Main Results:

• A structured categorization of anomaly detection methods is presented.
• Key objects-of-interest and relevant public datasets for anomaly detection are identified.
• The role and challenges of edge computing in real-time anomaly detection are discussed.

Conclusions:

• Anomaly detection in intelligent video surveillance is evolving, with edge computing offering promising solutions for real-time applications.
• A systematic understanding of methodologies and datasets is crucial for advancing the field.
• Further research is needed to address the challenges and opportunities of edge-based anomaly detection in smart cities.