武汉多应用高分辨率场景分类数据集（WH-MAVS）

WH-MAVS数据集是一个大范围、多应用、多时态、地理参考的高分辨率场景分类数据集，注释到目前为止几乎覆盖了整个特大城市。该数据集不仅可以方便场景分类和场景变化检测方法的理论研究，而且可以直接应用于土地利用/土地覆盖的实际生产 (如城市制图、景观生态分析、城市环境动态监测)。因此，可以跨越从算法到应用的 “应用差距”。

The WH-MAVS dataset is a large-scale, multi-application, multi-temporal, georeferenced high-resolution scene classification dataset. To date, its annotations cover nearly the entire megacity. This dataset not only facilitates theoretical research on scene classification and scene change detection methods, but also can be directly applied to practical production workflows for land use/land cover, including urban mapping, landscape ecological analysis, and dynamic urban environmental monitoring. Consequently, it can bridge the "application gap" between algorithm development and real-world applications.

搜集汇总

背景与挑战

背景概述

武汉多应用高分辨率场景分类数据集（WH-MAVS）是由武汉大学发布的高分辨率场景分类数据，具有大范围、多应用、多时态特点，覆盖整个特大城市区域，适用于场景分类、变化检测及土地利用分析等研究和应用。