MultiCamCows2024
收藏资源简介:
MultiCamCows2024是由布里斯托大学创建的多视角图像数据集,用于在实际农场环境中对荷斯坦-弗里西亚奶牛进行个体识别。该数据集包含101,329张图像,涵盖90头奶牛,通过三个天花板安装的视觉传感器在七天内拍摄。数据集的创建过程包括视频采集、物种检测、轨迹生成和细化,以及深度度量学习。该数据集的应用领域主要集中在智能农业、牲畜管理和行为分析,旨在通过高效的个体识别技术提升农业监测的精度和效率。
MultiCamCows2024 is a multi-view image dataset developed by the University of Bristol for individual identification of Holstein-Friesian cows in real-world farm environments. It comprises 101,329 images spanning 90 cows, captured over a seven-day period using three ceiling-mounted visual sensors. The dataset creation process includes video acquisition, species detection, trajectory generation and refinement, as well as deep metric learning. Its core application areas cover precision agriculture, livestock management and behavioral analysis, aiming to improve the accuracy and efficiency of agricultural monitoring through efficient individual identification technologies.




