HSOD-BIT-V1 高光谱显着目标检测基准数据集

HSOD-BIT 是第一个大规模、高质量的高光谱显着目标检测基准数据集，旨在利用光谱信息的优势在显着目标检测任务中实现更高的精度。该数据集针对当代深度学习模型的数据需求，为 319 个高光谱数据立方体提供像素级手动标注，并生成相应的伪彩色图像。每个数据立方体包含 200 个波段，涵盖从可见光到近红外波段的光谱信息，空间分辨率高达 1240×1680 像素。除了常规场景外，该数据集还专门收集了具有挑战性的数据来反映现实世界的复杂性，例如相似的背景干扰、光照不均匀、过度曝光等具有挑战性的场景。这进一步增强了数据集的实用性和评估能力。

HSOD-BIT is the first large-scale, high-quality benchmark dataset for hyperspectral salient object detection, designed to leverage the advantages of spectral information to achieve higher accuracy in salient object detection tasks. To meet the data demands of contemporary deep learning models, it provides pixel-level manual annotations for 319 hyperspectral data cubes and generates corresponding pseudo-color images. Each data cube contains 200 spectral bands covering the spectrum from visible light to near-infrared, with a spatial resolution of up to 1240×1680 pixels. In addition to conventional scenarios, this dataset also specially collects challenging data to reflect real-world complexities, such as similar background interference, non-uniform illumination, overexposure and other challenging scenes, which further enhances the practicality and evaluation capability of the dataset.

搜集汇总

背景与挑战

背景概述

HSOD-BIT-V1是首个大规模、高质量的高光谱显着目标检测基准数据集，包含319个高光谱数据立方体，每个立方体有200个波段（覆盖可见光到近红外）和高达1240×1680像素的空间分辨率，并提供像素级手动标注和伪彩色图像。该数据集旨在利用光谱信息提升检测精度，并专门收集了具有挑战性的场景（如相似背景干扰、光照不均匀），以增强数据集的实用性和评估能力，适用于深度学习模型训练和测试。