Drone-based Elephant Detection and Counting With YOLOv5

Nowadays, object detection is popularly used in most vision-based AI and it is sole to the various computer vision tasks like face recognition, Animal detection. This study focuses on the identification and tally of elephants using drone photos using YOLOv5. Drones allow you to span a vast area and facilitate prompt and effective animal monitoring. YOLOv5 is the first YOLO model to be written in PyTorch framework which is fast and accurate. As every species plays a crucial role in the ecosystem, extinction of one specie may cause damage to entire ecosystem. So, it is important to detect the endangered species and ensure their protection. In this project YOLO algorithm is used to detect and count the endangered species from drone images. In this project elephant is taken as a reference as the population of elephant is declining over several decades due to poaching and loss of habitat. Existing model only has detection of animal so counting of animal’s feature is added to it to increase its efficiency. The model developed has mAP 89.4%, Precision 89.1% and Recall 83.2%.