isp-uv-es/rtm_emulation

大气辐射传输模拟挑战数据集是一个旨在促进大气辐射传输模型（RTM）模拟的替代方法研究的挑战数据集。该数据集提供了多个训练数据集，这些数据集是通过拉丁超立方采样生成的RTM模拟，涵盖了不同的样本大小、输入参数和光谱范围/分辨率。挑战包括两个主要任务：训练机器学习模型以预测光谱值，并根据准确性和运行时评估模拟器性能。数据集分为两个应用测试场景：大气校正和二氧化碳柱状检索，并为每个场景提供了插值和 extrapolation的测试数据集。

The Atmospheric Radiative Transfer Emulation Challenge dataset is designed to promote research on alternative methods for simulating Atmospheric Radiative Transfer Models (RTM). The dataset provides multiple training datasets of RTM simulations generated using Latin Hypercube Sampling, covering different sample sizes, input parameters, and spectral range/resolution. The challenge includes two main tasks: training machine learning models to predict spectral values and evaluating emulator performance based on accuracy and runtime. The dataset is divided into two application test scenarios: atmospheric correction and CO2 column retrieval, with interpolation and extrapolation test datasets provided for each scenario.