Spacenet7

该数据集名为Spacenet7，包含了全球60个地区的RGB镶嵌图像，每个地区的时间序列包含24张图像，这些图像在两年时间内收集，覆盖了4公里乘4公里的区域。该数据集主要用于处理卫星图像，以识别城市结构，处理过程包括图像的标准化、直方图计算以及使用K均值聚类的分析。目前，已有720张图像得到处理（来自30个地区），任务重点在于为城市发展进行的图像处理与分析。

The dataset named Spacenet7 encompasses RGB mosaic imagery collected from 60 global regions. Each region spans a 4 km × 4 km area, with its time series consisting of 24 images gathered over a two-year period. This dataset is primarily designed for satellite image processing aimed at urban structure recognition, with its processing pipeline including image standardization, histogram computation, and analysis via K-means clustering. As of now, 720 images from 30 regions have been processed, and the key tasks of this dataset center on image processing and analysis for urban development.

搜集汇总

背景与挑战

背景概述

Spacenet7是一个全球卫星图像数据集，包含60个地区的时间序列RGB图像，用于城市结构识别和分析。数据集覆盖4公里×4公里区域，每个地区有24张图像，已有部分图像经过标准化和聚类处理。