In order to prevent damage to farms caused by wild animals, we propose a wildlife detection and tracking automation system using artificial intelligence technology to effectively manage wild animals. Artificial intelligence has found the best algorithm for detecting and tracking objects so that they can predict the paths of wild animals or steadily search for populations. A total of 11 models were Object Detected using various versions and backbones of the YOLO algorithm. Multiple object tracking was conducted through the Strong SORT algorithm based on the detection results by selecting the best algorithm. We find object detection algorithms and multi-object tracking algorithms that show the highest accuracy and efficiency. Through this, the effect of object detection results on object tracking was identified.
In addition, the relevance of object tracking results according to environmental factors such as length, complexity, density, daytime, and nighttime of CCTV images was found.