• Version
• Download 31
• File Size 224.87 KB
• File Count 1
• Create Date 27 May 2025
• Last Updated 27 May 2025

Intelligent Video Surveillance using Deep Learning

Authors:

Abhishek Sunil Sarjine¹, Pramod R. Jadhao².

¹Master of Computer Application, Trinity Academy of Engineering, Pune.

²Assistant Professor (Dept of MCA), Trinity Academy of Engineering, Pune.

ABSTRACT: In recent years, the demand for automated and intelligent surveillance systems has increased significantly due to rising concerns over public safety and security. Traditional surveillance systems rely heavily on manual monitoring, which is time-consuming, error-prone, and lacks real-time responsiveness. This research presents the development of an intelligent video surveillance system using deep learning techniques, specifically employing the YOLOv5s object detection model for efficient and accurate real-time detection of objects such as people, vehicles, and animals. The system is implemented as a web-based application using Flask, with functionalities for both live surveillance through a webcam and video upload analysis. The YOLOv5s model ensures a balanced trade-off between speed and detection accuracy, enabling smooth operation even on limited hardware. The application also incorporates user authentication and object counting, offering a complete end-to-end solution. Experimental results demonstrate that the system performs well in diverse environments, making it suitable for use in schools, offices, and public places. This project highlights the effectiveness of deep learning in enhancing modern surveillance solutions.

Keywords:- Deep Learning, YOLO (You Only Look Once), Object Detection, Real-Time Tracking, Flask Web Application, Object Counting, Automated Surveillance System