Google Street View Images of Milan and San Francisco

该数据集包括从谷歌街景API获取的地理标记街景图像，覆盖了两个城市区域：意大利的米兰和美国加利福尼亚州的旧金山。这些图像用于基于对比聚类的非监督土地利用映射。数据集包含了30,994张来自米兰的街景图像和30,730张来自旧金山的图像。为了确保空间连贯性并减少因观看角度造成的偏差，图像经过了筛选。研究范围覆盖了中心城区的10公里×10公里区域，任务是对城市土地利用进行分类和映射。

This dataset comprises geotagged street view images obtained through the Google Street View API, covering two urban regions: Milan in Italy and San Francisco, California, USA. These images are utilized for unsupervised land use mapping based on contrastive clustering. The dataset includes 30,994 street view images from Milan and 30,730 from San Francisco. To ensure spatial coherence and mitigate bias introduced by viewing angles, the images underwent filtering. The study covers a 10 km × 10 km area of the urban core, with the task of classifying and mapping urban land use.