Unmanned Aerial Vehicle (UAV) Detection in Video

Abstract

Unmanned Aerial Vehicle (UAV) or Drone is a type of unmanned aerial vehicle. They’re gaining in popularity, and they’re reaching the general population at a quicker rate than before. As a result, the likelihood of a drone being misused is increasing. To avoid unlawful and unwanted drone intrusions, automated drone detection is required. As a result, we’re working on a deep learning model that can recognize drones of various sorts and forms in order to fill the gap left by earlier deep learning models that can detect drones in this study. Current drone detection deep learning models, for example, are unable to recognize drones with a bird-like form, but our model can bridge this gap. We’re employing the YOLO object detection model with OpenCV for our automated UAV detection challenge. We are training the YOLO object identification model using our unique datatset, which contains images of various types of drones as well as drones that resemble birds in real time with the maximum level of accuracy, we have achieved 96.98%@0.5 mAP. We’re merging DeepSORT with YOLO to track detected items.