Automated Recognition of Wild Animal Species in Camera Trap Images Using Deep Learning Models

Wild animal monitoring is one of the essential factors in balancing the ecosystem in protecting the animal habitat. Camera traps are used to capture the images, which play a vital role in the automatic monitoring of wildlife owing to their effectiveness in collecting wildlife data. The images captured from the camera traps are large volumes in nature, and they are inexpensive. However, the manual processing of such a large number of images is a time-consuming, expensive, and challenging task. Such a kind of process provides major hurdles to ecologists in monitoring wildlife. This paper evaluates deep learning models such as DeneseNet169, InceptionResNetV2, ResNetV2-50, and Xception for the automatic identification and recognition of individual animals in different patterns. The effectiveness of the models is measured using precision, recall, and F1 score indicators. Finally, the experimental findings show that the InceptionResNetV2 model outperforms the other models with an accuracy of 99%.