Aerial Image Segmentation

该数据集用于航空图像分割研究。该数据集包含80个高分辨率航空图像，其空间分辨率范围从0.3到1.0。图像集包含不同的场景，包括学校，住宅，城市，仓库和发电厂。每个图像的大小是512 512像素。分配了四个人类受试者来手动分割图像。下图显示了数据集中的五幅图像以及相应的人工分割。

This dataset is designed for aerial image segmentation research. It contains 80 high-resolution aerial images with spatial resolutions ranging from 0.3 to 1.0. The image set encompasses various scenarios including schools, residential areas, urban areas, warehouses, and power plants. Each image has a resolution of 512 × 512 pixels. Four human annotators were assigned to manually segment these images. The figure below displays five images from the dataset along with their corresponding manual segmentations.

搜集汇总

背景与挑战

背景概述

该数据集包含80张512x512像素的高分辨率航空图像，覆盖学校、住宅等不同场景，用于图像分割研究。图像由四人手动标注，由橡树岭国家实验室于2013年发布。