Advancing Wildlife Detection with YOLOv8 and YOLOv9: A Deep Learning Approach
Manav Majmudar, Harsh Shah, Dev Parmar +4 more
Wildlife monitoring plays a vital role in preventing animal-human conflicts and protecting biodiversity. With growing human encroachment into natural habitats, there’s an urgent need for intelligent systems capable of detecting animals in real-time. Traditional surveillance methods are often resource-heavy, inaccurate, or slow to respond. In this work, we explore the use of deep learning models, specifically YOLOv8 and YOLOv9, for wild animal detection from image and video inputs. Our custom dataset includes 80 classes of wild animals, each with 250–300 images, augmented to improve diversity and robustness under different lighting and environmental conditions. Both YOLOv8 and YOLOv9 were trained and evaluated on this dataset to compare performance. While YOLOv8 provided excellent speed and accuracy, YOLOv9 delivered further improvements in precision, especially for smaller or partially…