Objects in Public Open Spaces (OPOS)

OPOS数据集是由密歇根大学创建的，旨在通过深度学习技术量化公共开放空间中的人类活动。该数据集包含7826张图像，这些图像是从18个监控摄像头在不同光照条件下收集的，涵盖了多种交通类型和对象类别。数据集的创建过程包括初始的手动标注和后续的模型辅助标注，确保了数据的高质量和准确性。OPOS数据集主要用于城市规划和设计领域，帮助研究人员和城市管理者更好地理解和优化公共空间的使用，从而提升城市居民的生活质量。

The OPOS Dataset was developed by the University of Michigan, with the goal of quantifying human activities in public open spaces via deep learning technologies. It contains 7,826 images collected from 18 surveillance cameras under varying lighting conditions, covering a wide range of traffic types and object categories. The dataset creation process includes initial manual annotation followed by model-assisted annotation, which ensures high data quality and accuracy. The OPOS Dataset is primarily applied in the field of urban planning and design, aiding researchers and urban managers to better understand and optimize the utilization of public spaces, thus enhancing the quality of life for urban residents.