Typical behavior dataset of ewes

This dataset consists of images of typical behaviors of ewes from different periods. These images were collected from three fences at the Xin'ao Sheep Breeding Base in Lanzhou Bay Town, Manas County, Xinjiang Uygur Autonomous Region, China from October 1st to October 23rd. There are a total of 3923 images in the dataset, self-made by LabelImg open-source software, and its TXT format meets the common formatting requirements of YOLO. These images were enhanced using techniques such as mask addition, 30 degree and 60 degree rotation, and adaptive enhancement.
This dataset is jointly managed by researchers from the Northwest Agricultural Equipment Key Laboratory of the School of Mechanical and Electrical Engineering at Shihezi University in Xinjiang. The purpose of creating this dataset is to support the development of computer applications using machine learning and computer vision algorithms, while promoting the intelligence of animal husbandry. By studying the different behaviors of ewes, researchers can effectively monitor the activity status of each ewe and use multi-target tracking methods in subsequent video surveillance to achieve various task scenarios. This research method is beneficial for livestock enterprises to better understand the health status of ewes and help them minimize economic losses in the later stage.