PigDetect 和 PigTrack
收藏资源简介:
PigDetect 和 PigTrack 是两个用于猪只检测和跟踪的数据集,由德国哥廷根大学经济学院统计与计量经济学系、基尔大学动物育种与繁殖研究所、德国科学数据处理公司共同创建。数据集包含来自真实猪舍环境的多样化图像和视频材料,包括遮挡、光照不良等挑战性场景。PigDetect 数据集包含 2931 张图像,其中包含 568 张挑战性图像,用于对象检测;PigTrack 数据集用于多目标跟踪。这些数据集旨在提高猪只监测的自动化水平,确保动物福利和有效管理。数据集和训练代码已公开发布,以便于可重复性、再利用和进一步发展。
PigDetect and PigTrack are two datasets for pig detection and tracking, co-developed by the Department of Statistics and Econometrics at the School of Economics, University of Göttingen, Germany; the Institute of Animal Breeding and Genetics, Kiel University; and a German scientific data processing company. The datasets include diverse image and video materials collected from real pig housing environments, covering challenging scenarios such as occlusions and poor lighting conditions. The PigDetect dataset contains 2,931 images, among which 568 are challenging ones, and is used for object detection; the PigTrack dataset is designed for multi-object tracking. These datasets aim to elevate the automation level of pig monitoring, ensure animal welfare and enable effective management. The datasets and training code have been publicly released to facilitate reproducibility, reuse and further development of relevant research.

- 1Benchmarking pig detection and tracking under diverse and challenging conditions德国哥廷根大学经济学院统计与计量经济学系、基尔大学动物育种与繁殖研究所、德国科学数据处理公司 · 2025年



