HSRS-SC Dataset

The HSRS-SC dataset comprises a total of 1,385 hyperspectral scene images, each with a standardized dimension of 256×256 pixels, which are categorized into five semantic classes. The number of samples per class in this dataset ranges from 154 to 485. Each hyperspectral scene image is characterized by high spatial resolution (1 meter) and a broad spectral range (visible to near-infrared, 380–1050 nm), enabling the reflection of detailed spatial information of ground objects and rich spectral features.

HSRS-SC 数据集共涵盖1385幅高光谱场景图像，所有图像均采用256×256像素的标准尺寸，并被划分为5个语义类别。该数据集各分类的样本数量介于154至485之间。每幅高光谱场景图像均具备1米的高空间分辨率与380–1050 nm的宽光谱范围（可见光至近红外波段），可有效呈现地物的精细空间信息与丰富的光谱特征。

搜集汇总

背景与挑战

背景概述

HSRS-SC数据集包含1,385个256×256像素的高光谱场景图像，分为5个语义类别，每类样本数154-485个。图像具有1米高空间分辨率和380-1050纳米宽光谱范围，能提供地面物体的详细空间和丰富光谱信息。