Leonard Alycia, Wheeler Scot, McCulloch Malcolm
Department of Engineering Science, University of Oxford, Parks Road, Oxford, OX1 3PJ, UK.
Data Brief. 2022 May 11;42:108262. doi: 10.1016/j.dib.2022.108262. eCollection 2022 Jun.
This article presents a geolocated dataset of rural home annotations on very high resolution satellite imagery from Uganda, Kenya, and Sierra Leone. This dataset was produced through a citizen science project called "Power to the People", which mapped rural homes for electrical infrastructure planning and computer-vision-based mapping. Additional details on this work are presented in "Power to the People: Applying citizen science to home-level mapping for rural energy access" [1]. 578,010 home annotations were made on approximately 1,267 km of imagery over 179 days by over 6,000 volunteers. The bounding-box annotations produced in this work have been anonymized and georeferenced. These raw annotations were found to have a precision of 49% and recall of 93% compared to a researcher-generated set of gold standard annotations. Data on roof colour and shape were also collected and are provided. Metadata about the sensors used to capture the original images and the annotation process are also attached to data records. While this dataset was collected for electrical infrastructure planning research, it can be useful in diverse sectors, including humanitarian assistance and public health.
本文展示了一个基于地理位置的数据集，该数据集包含乌干达、肯尼亚和塞拉利昂非常高分辨率卫星图像上的农村房屋标注。这个数据集是通过一个名为“人民力量”的公民科学项目生成的，该项目为电力基础设施规划和基于计算机视觉的测绘绘制农村房屋地图。这项工作的更多细节见《人民力量：将公民科学应用于农村能源获取的家庭层面测绘》[1]。6000多名志愿者在179天内，在约1267公里的图像上进行了578,010次房屋标注。这项工作中产生的边界框标注已进行匿名化处理并地理配准。与一组由研究人员生成的金标准标注相比，发现这些原始标注的精度为49%，召回率为93%。还收集并提供了有关屋顶颜色和形状的数据。用于捕获原始图像和标注过程的传感器的元数据也附加到数据记录中。虽然这个数据集是为电力基础设施规划研究而收集的，但它在包括人道主义援助和公共卫生在内的不同领域可能会有用。
