Voxel51/parkseg12k_train

ParkSeg12k是一个包含11355个样本的停车场分割数据集，由UTEL-UIUC（伊利诺伊大学厄巴纳-香槟分校城市交通与经济实验室）策划，Harpreet Sahota使用FiftyOne进行了转换并计算了NDVI。数据集使用RGB和近红外(NIR)影像，覆盖了美国45个城市，标注了约35000个停车场。数据集适用于停车场检测的语义分割模型训练、城市规划与政策分析、土地利用模式分析以及支持停车改革政策讨论。

ParkSeg12k is a parking lot segmentation dataset with 11,355 samples, curated by UTEL-UIUC (Urban Traffic & Economics Lab, University of Illinois at Urbana-Champaign) and enhanced with NDVI calculations by Harpreet Sahota using FiftyOne. The dataset uses RGB and near-infrared (NIR) imagery, covering 45 US cities with approximately 35,000 annotated parking lots. It is suitable for training semantic segmentation models for parking lot detection, urban planning and policy analysis, analyzing land use patterns in US cities, and supporting discussions on parking reform policies.