Electrolyzers-HSI
收藏资源简介:
该数据集是一个旨在通过准确分类电解器材料来加速关键原材料恢复的多模态基准数据集。它包含了55个高分辨率RGB图像和跨越400-2500纳米光谱范围的高光谱成像(HSI)数据立方体,这些图像和数据立方体是共同注册的。该数据集包含了五种类别(网格、黑色钢材、灰色钢材、HTEL阴极和HTEL阳极)的光谱特征,并通过加权交叉熵损失方法来解决类别不平衡问题。规模上,55个共同注册的图像包含了超过420万像素向量,以及424,169个标记向量。该数据集的任务是在回收背景下,使用RGB和高光谱图像进行材料分类。
This multimodal benchmark dataset aims to accelerate the recovery of critical raw materials via accurate classification of electrolyzer materials. It comprises 55 high-resolution RGB images and hyperspectral imaging (HSI) data cubes covering the 400–2500 nm spectral range, all of which are co-registered. The dataset features spectral signatures for five distinct categories: grid, black steel, grey steel, HTEL cathode, and HTEL anode, and utilizes weighted cross-entropy loss to address the class imbalance problem. In terms of data scale, the 55 co-registered images contain over 4.2 million pixel vectors and 424,169 labeled vectors. The core task of this dataset is material classification using RGB and hyperspectral images in the scenario of recycling.



