An Undergraduate Internship on InfiltraWatch

Abstract

This project develops a web application that makes use of modern computer vision and existing CCTV infrastructure to solve the growing security issues in industrial buildings. The main goal is to effectively and instantly hinder potential attackers and prevent unlawful entry. The solution makes use of a state-of-the-art AI model to intelligently evaluate the CCTV cameras’ video stream to recognize and categorize any individuals in the picture. To enable thorough surveillance of the premises, the model is particularly trained to recognize human figures, count the number of people, and mark off forbidden zones. When it detects any human presence in the designated areas after hours. The MySQL database is populated with the combined date, time, human count, and camera position by a Python script. A simultaneous alarm is sent off inside the designated region with the intention of frightening and discouraging the intruder. It is anticipated that the implementation of this web application would dramatically improve security controls, decrease potential weaknesses, and significantly lower the danger of property damage and loss in industrial settings. The project not only seeks to secure the factory property but also contributes to the investigation of the use of AI in surveillance for more extensive applications