S2Looking

S2Looking 是一个建筑物变化检测数据集，其中包含以不同的偏离天底角捕获的大规模侧视卫星图像。 S2Looking 数据集由全球农村地区的 5,000 个已注册的双时间图像对（大小为 1024*1024，0.5 ~ 0.8 m/像素）和超过 65,920 个带注释的变化实例组成。我们提供了两个标签图来分别指示数据集中每个样本的新建和拆除的建筑区域。我们基于该数据集建立了一个基准任务，即识别双时图像中的像素级建筑变化。

S2Looking is a building change detection dataset that contains large-scale side-looking satellite images captured at varying off-nadir angles. The S2Looking dataset consists of 5,000 registered bi-temporal image pairs (1024×1024 in size, with a spatial resolution of 0.5 to 0.8 m/pixel) from global rural areas, along with over 65,920 annotated change instances. We provide two label maps to separately indicate the newly built and demolished building regions for each sample in the dataset. We have established a benchmark task based on this dataset, which aims to identify pixel-level building changes in bi-temporal satellite images.

搜集汇总

背景与挑战

背景概述

S2Looking是一个专注于建筑物变化检测的卫星图像数据集，包含全球农村地区的5,000个双时间图像对，分辨率高达0.5-0.8米/像素，标注了超过65,920个变化实例。该数据集以侧视角度捕获，提供新建和拆除区域的标签图，用于像素级建筑变化识别任务，适用于遥感图像分析研究。