INTELLIGENT REAL-TIME VIDEO SURVEILLANCE SYSTEM USING 2-D SAD MOTION DETECTION AND MASCOT METHODOLOGY FOR MODULAR EMBEDDED DESIGN

Abstract

This paper presents an intelligent real-time video surveillance system using a 2-D Sum of Absolute Differences (SAD) motion detection algorithm, integrated with the MASCOT (Modular Approach to Software Construction Operation and Test) methodology to ensure scalability, modularity, and reliability. The system processes live video fed from multiple IP/Bluetooth cameras, detects motion using SAD values exceeding a defined threshold, and transmits and stores frames over a TCP/IP network. By applying MASCOT, we decompose the system into well-defined activity and communication modules, enabling formal verification, system-level testing, and easier maintainability. Simulation results on a DSP platform (TMS320C6416) demonstrated the feasibility of the system for real-time motion detection and multi-camera synchronization, with potential applications in security monitoring, smart environments, and automated surveillance.