中国农村地区建筑物样本及标注无人机影像数据集

农村建筑物是观察农村土地变化和经济发展的基础资料。中国作为农业大国，从高空间分辨率遥感影像上及时、准确提取农村建筑物，对于农村发展至关重要。近年来，随着计算机视觉和运算能力的迅速发展，深度学习以其自动学习特征、适用性强等优点，已在建筑物自动提取等领域取得较好效果。深度学习通常需要大量的训练数据。目前，深度学习提取建筑物常用的数据集以国际上开源建筑物数据集为主，包括Massachusetts, INRIA, WHU等。这些数据集大多基于国外建筑物，缺乏开源、高精度、覆盖范围广、贴切我国农村地区建筑主体结构的建筑物样本数据。为此，本研究基于2017–2020年在陕西渭南、江苏淮安、四川康定、广东汕尾、广东惠州、新疆阿图什、吉林松原等多个中国农村地区采集的无人机航拍图像，制作并开放共享本数据集。本数据集空间分辨率高，基本涵盖我国农村地区房屋建筑的主体结构类型，可应用深度学习方法进行建筑物提取，并可进一步结合具体研究目标进行空间分析和研究，对于国土部门统筹城乡发展和美丽乡村建设具有重要意义和应用价值。

Rural buildings are fundamental data for observing rural land changes and economic development. As a major agricultural country, timely and accurate extraction of rural buildings from high-spatial-resolution remote sensing images is crucial for rural development in China. In recent years, with the rapid advancement of computer vision and computing power, deep learning has achieved excellent performance in fields such as automatic building extraction, owing to its advantages including automatic feature learning and strong general applicability. Deep learning typically requires a large volume of training data. Currently, the most widely used datasets for building extraction via deep learning are primarily international open-source building datasets, such as Massachusetts, INRIA, WHU, etc. However, most of these datasets are based on foreign building samples, and lack open-source, high-precision, wide-coverage building sample data that match the main structural types of buildings in China's rural areas. To address this gap, this study compiled and openly shared this dataset using UAV aerial images collected in multiple rural areas of China from 2017 to 2020, including Weinan (Shaanxi), Huai'an (Jiangsu), Kangding (Sichuan), Shanwei (Guangdong), Huizhou (Guangdong), Atushi (Xinjiang), Songyuan (Jilin) and other regions. This dataset features high spatial resolution and basically covers the main structural types of residential buildings in China's rural areas. It can be applied for building extraction using deep learning methods, and further utilized for spatial analysis and research combined with specific research objectives. It holds important significance and application value for land management departments to coordinate urban-rural development and advance beautiful countryside construction.