Aerial monitoring is critical for safeguarding governmental facilities, restricted areas, and borders, with drones playing a pivotal role due to their precise object location capabilities. However, traditional computer vision methods often lack accuracy, leading to detection errors. The developed system introduces a novel approach that integrates YOLOv4, an advanced object detection algorithm, with DeepSORT for real-time tracking, specifically tailored for drone surveillance. Developed alongside drone technology, the system effectively identifies and locates vehicles and humans within restricted zones. Utilizing YOLOv4 and its Darknet 53 backbone, it ensures precise object detection in drone-captured videos. DeepSORT further enhances geotagging and continuous tracking of identified objects. The system receives live video feeds from drones, processes data at a ground station, and activates surveil-lance protocols based on detected objects. This integration enhances surveillance capabilities, facilitating efficient monitoring and security enforcement in sensitive areas through drone technology.
Advanced Aerial Monitoring and Tracking System: YOLOv4 and DeepSORT Integration for Drone-Based Surveillance
26.04.2024
990371 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch