Multisensor Hyperspectral Benchmark Dataset For Unmixing of Intimate Mixtures

本数据集名为‘多传感器高光谱基准数据集用于亲密混合物的分解’，由安特卫普大学的Imec-Visionlab创建。数据集包含325个样本，这些样本是通过均匀混合五种不同的粘土粉末（高岭土、屋顶粘土、红粘土、混合粘土和氢氧化钙）制备而成。这些样本使用13种不同的超光谱传感器在可见光至长波红外区域（350-15385 nm）获取反射光谱。数据集旨在验证高级非线性分解方法和材料成分估计，特别是在研究光谱变异性和训练监督分解方法方面。该数据集适用于近距离场景的应用，如建筑材料质量评估。

This dataset is named 'Multi-sensor Hyperspectral Benchmark Dataset for Intimate Mixture Decomposition', and was developed by Imec-Visionlab at the University of Antwerp. It contains 325 samples prepared by uniformly mixing five distinct clay powders: kaolin, roofing clay, red clay, mixed clay, and calcium hydroxide. Reflectance spectra of these samples were acquired using 13 different hyperspectral sensors across the visible to long-wave infrared range (350–15385 nm). This dataset is intended to validate advanced nonlinear decomposition methods and material composition estimation, especially for studies on spectral variability and training supervised decomposition approaches. It is suitable for close-range scene applications such as construction material quality assessment.