Aerial Object Detection and Tracking using YOLOv4 and DeepSORT

Aerial Surveillance can be used to monitor impor-tant governmental offices, restricted zones and border patrolling too. Drones can give exact location of object and hence can be used for surveillance. Earlier traditional computer vision techniques were used to do the feature extraction. Later the features were given to a machine learning classifier for detection and classification. These algorithms used were highly inaccurate and generated false detection and misclassifications. The Deep Learning algorithms are able to extract more information and provide better accuracy as compared to traditional algorithms. For object detection YOLOv4 is used which is one of the state-of-the-art algorithms. It uses Darknet 53 which is a type of CNN as a backbone for feature extraction. DeepSORT is used for real-time tracking of detected objects. Here, the YOLOv4 based proposed system detect and localize vehicles present in the restricted zone and then geotag and later DeepSORT is used to track them.