Despite animal husbandry being one of the most important industries in our society, the technology used is relatively rudimentary. This lack of technological advancements, make it a difficult and arduous job for humans and limit the quality of life given to the animals. The goal of this project is to create an animal detection application that could serve as a baseline for future projects to assist livestock farmers with tasks such as monitoring, care taking, herding and protection of the animals. To do this, different computer vision techniques have been analyzed, finally choosing the deep learning algorithm YOLO to develop the animal detection application. A detector is trained with a dataset of cows and sheep images, which is later improved by applying data augmentation to the dataset. To further improve the predictions of the detector, a pre-processing and post-processing step is applied to the input image. The detector obtained after applying data augmentation to the training dataset offers fast predictions, making it viable for real-time applications, although obtaining worse results for far away views of the animals. The added processing steps, on the other hand, offer good prediction results for far away views of the animals but cannot be used for real-time applications. Depending on the requirements of each application, one or the other approach can be taken.