This research paper introduces a synthetic hyperspectral dataset that combines high spectral and spatial resolution imaging to achieve a comprehensive, accurate, and detailed representation of observed scenes or objects. Obtaining such desirable qualities is challenging when relying on a single camera. The proposed dataset addresses this limitation by leveraging three modalities: RGB, push-broom visible hyperspectral camera, and snapshot infrared hyperspectral camera, each offering distinct spatial and spectral resolutions. Different camera systems exhibit varying photometric properties, resulting in a trade-off between spatial and spectral resolution. RGB cameras typically offer high spatial resolution but limited spectral resolution, while hyperspectral cameras possess high spectral resolution at the expense of spatial resolution. Moreover, hyperspectral cameras themselves employ different capturing techniques and spectral ranges, further complicating the acquisition of comprehensive data. By integrating the photometric properties of these modalities, a single synthetic hyperspectral image can be generated, facilitating the exploration of broader spectral-spatial relationships for improved analysis, monitoring, and decision-making across various fields. This paper emphasizes the importance of multi-modal fusion in producing a high-quality synthetic hyperspectral dataset with consistent spectral intervals between bands.

本研究介绍了一个合成高光谱数据集，通过高光谱和高空间分辨率成像实现观测场景或物体的全面、准确、详细的表示，强调多模态融合在生成高质量合成高光谱数据集中的重要性，以及光谱-空间关系的改善对于各个领域的分析、监测和决策的影响。

多源融合合成高光谱数据集的高光谱空间分辨率提升