Effective Detection of Weapons in Video Surveillance

Abstract

Surveillance cameras also known as Closed Circuit Tele-Vision (CCTV) play a major part in the enforcement of the law and the administration of justice. The present level of industrial production technology cannot differentiate between a good and a bad, or even a bit worse, situation. As a consequence, additional crime scene investigation or law order maintenance work is needed, which takes a lengthy time. The suggested work is being utilized for a number of purposes, including surveillance, weapon monitoring and classification, live tracking, and more. Video input is permitted as a type of raw input in this project for monitoring and detecting abnormal events utilizing realtime detection techniques such as You Look Only Once Version 3 (YOLO V3). The proposed project's operations make use of a processing module for object recognition using convolutional neural networks such as YOLO V3, which predicts classes and bounding boxes for the entire image in a single run of the algorithm. The circular area will be watched by CCTV, which will automatically execute all operations and be controlled. Before implementing in such a setting and delivering optimal results, numerous samples and datasets will be examined to find accuracy in detection and classification. The planned effort aims to significantly reduce crime rates while simultaneously providing improved protection in specific areas and decreasing the time it takes to capture a criminal.