RGB-MS Dataset

RGB-MS数据集是由意大利博洛尼亚大学计算机科学与工程系CVLAB团队创建的，专注于室内环境中的RGB-MS图像配准问题。该数据集包含34对同步的RGB和多光谱图像，每对图像均配有半密集、高分辨率的真实标签，以视差图形式提供。数据集的创建过程涉及使用两个高分辨率RGB相机和多个投影仪，构建了一个精确的时空立体设置，从而生成了超过1.25亿个真实对应点。该数据集主要用于解决RGB和多光谱图像之间的跨模态匹配问题，支持在移动设备上实现多光谱成像应用的研究。

The RGB-MS dataset was created by the CVLAB team from the Department of Computer Science and Engineering, University of Bologna, Italy, and focuses on RGB-MS image registration in indoor environments. This dataset includes 34 pairs of synchronized RGB and multispectral images, each accompanied by semi-dense, high-resolution ground truth labels provided in the form of disparity maps. The dataset's development involved utilizing two high-resolution RGB cameras and multiple projectors to construct a precise spatiotemporal stereo setup, yielding over 125 million ground truth correspondences. This dataset is primarily designed to address cross-modal matching challenges between RGB and multispectral images, enabling research on multispectral imaging applications on mobile devices.