Real-time weapon detection using Drone

Abstract

Drones or Unmanned aerial vehicles (UAV's) are used widely because of their advantages like small size, low cost and portability. From past few years, drones have become important part in private as well as government sectors to reduce work load as well as to do work in less time. In agriculture field also drones are used to monitor crops, to supply water and to send live video of crops over network. Real-time object detection can play important role in security and surveillance. Drones can fly autonomously and it can capture live video. This video can be used for object detection. Proposed system captures live video frame using Raspberry pi camera module with help of drone. This video data is sent to real-time weapon detection system. If weapon is detected in video, then system give alert/notification on it, so operator can take appropriate action. Object detection module is created at computer side for fast object detection. OpenCV and YOLOv3 is used for weapon detection. Raspberry pi camera module work as a camera for computer. System has shown very good performance for detecting guns and rifle in stored as well as live streaming video.