国图遥感影像建筑物语义分割样本数据集

遥感影像建筑物语义分割数据集包含高分辨率遥感影像（.tiff）和9类建筑物标注数据(,png)，是依靠国图信息自行研发的样本管理系统进行标注和生成的，原始影像数据来源于Google earth公开免费数据，分辨率为 0.5m，样本尺寸包括512×512或256×256，可用于训练影像解译深度学习模型，开展建筑物的识别分割任务，能够有效应用于自然资源、林业、农业、环保、水利、海洋、城市建设等业务领域。

Remote Sensing Image Building Semantic Segmentation Dataset encompasses high-resolution remote sensing images (stored in .tiff format) and 9-category building annotation data (stored in .png format). It was annotated and generated via a sample management system independently developed by Guotu Information. The original imagery is sourced from publicly and freely available data on Google Earth, with a spatial resolution of 0.5 meters. The available sample sizes are 512×512 and 256×256. This dataset can be used to train deep learning models for image interpretation, perform building recognition and segmentation tasks, and can be effectively applied in business fields including natural resources, forestry, agriculture, environmental protection, water conservancy, ocean affairs, and urban construction.

搜集汇总

背景与挑战

背景概述

该数据集包含0.5米分辨率的Google earth遥感影像及9类建筑物标注数据，样本尺寸为512×512或256×256，适用于深度学习模型训练及建筑物识别分割任务，可广泛应用于自然资源、农业等多个领域。