AI POWERED SMART SURVEILLANCE FOR CAMPUS SECURITY AND MONITORING

Abstract

As the need for improved campus safety grows, intelligent surveillance systems that can monitor and determine potential risks in real-time are being developed. This is particularly relevant for surveillance in large areas, as traditional methods are manual, which is bound to produce both inefficiency and oversight. This paper proposes an AI based intelligent surveillance system to help monitor and improve campus safety. The system being developed will utilize the You Only Look Once (YOLO) framework to provide real-time detection of people and vehicles. Detected people will undergo face detection using the Haar Cascade Classifier, and then the detection and/or recognition of the face will be determined using the Local Binary Pattern Histogram (LBPH). Detected cars will have their license plates determined from using OCR. Because of this, the system will have an interactive database that provides the logs of the in and out of people and cars. There will be real-time notifications regarding people of concern, and responding to the repeated suspicious behaviour is managed by the system. The results of the system are the reduction in the need for human monitoring and the increased campus safety, as the system will have the ability to complete the tasks of an individual. This system is extendable, and a smart campus system is being developed for the system to be implemented in an economically careful way.