红外像元的光谱与性能关系数据

采用基于HDF5底层的 NeXus 标准格式构建数据体系，所有文件均遵循 NXentry 根组下包含 NXdata 数据子组的结构，通过signal和axes属性定义数据映射，确保兼容性。
核心数据分为四部分。首先是预处理光谱，存储于Spectra_Derivative_NXS 目录下的81920 个独立文件中，对应单像素响应。关键字段derivative存储归一化并对波长求导后的光电流强度（Y轴），wavelength 字段存储波长（X轴），并附带波长与波数转换说明。其次是机器学习特征矩阵文件 features_for_ml.nxs，专为模型训练设计。
主字段 features是维度为(81920, 20)的二维数组，作为输入矩阵X；response字段存储目标响应率向量y。同时feature_names字符串数组用于标识特征物理含义，pixel_index用于空间回溯。模型配置存储于rf_model_config.nxs，采用NXcollection结构组织。n_estimators、max_depth等超参数，字段均附带description属性提供功能描述。最后是分析结果部分。

The data system is constructed using the NeXus standard format based on the underlying HDF5 framework. All files follow the structure where the NXentry root group contains the NXdata data subgroup, and data mapping is defined via the `signal` and `axes` attributes to ensure compatibility.

The core data is divided into four parts. First is the preprocessed spectra, stored as 81920 independent files under the `Spectra_Derivative_NXS` directory, corresponding to single-pixel responses. The key field `derivative` stores the photocurrent intensity (Y-axis) after normalization and wavelength derivation, while the `wavelength` field stores wavelengths (X-axis), with additional instructions for wavelength and wavenumber conversion.

Second is the machine learning feature matrix file `features_for_ml.nxs`, specially designed for model training. The main field `features` is a 2D array with dimensions (81920, 20) serving as the input matrix X; the `response` field stores the target response rate vector y. Additionally, the `feature_names` string array is used to identify the physical meanings of the features, and `pixel_index` is for spatial backtracking.

The model configuration is stored in `rf_model_config.nxs`, organized using the NXcollection structure. Hyperparameters such as `n_estimators` and `max_depth` are all accompanied by a `description` attribute to provide functional descriptions. Finally, there is the analysis result section.

搜集汇总

背景与挑战

背景概述

该数据集聚焦于红外像元的光谱特性与其性能之间的关联，旨在提供关键参数以支持光学传感器或相关技术的研究与开发。其内容可能涉及光谱数据分析和性能评估，适用于红外成像、材料科学或工业检测等领域。数据集的特点在于专业性和应用导向，有助于优化红外系统的设计和性能提升。