OpenEM: Large-scale multi-structural 3D datasets for electromagnetic methods

OpenEM is a large-scale, multi-structural three-dimensional geoelectrical model dataset designed to support research in geophysics, particularly in the areas of electromagnetic (EM) modeling, inversion, and machine learning applications. The dataset includes nine types of geological models that cover a wide range of resistivity distributions and structural complexities. Due to file size limitations and to facilitate convenient use in deep learning tasks, the full dataset is hosted on https://huggingface.co/datasets/ws11/OpenEM. The forward modeling code developed for OpenEM is openly available at https://github.com/WAL-l/OpenEM. The 3D model generator is openly available at https://github.com/WAL-l/Geoelectrical3D.OpenEM是一个大规模多结构三维地电模型数据集，旨在支撑地球物理学领域的研究，尤其聚焦于电磁（electromagnetic，简称EM）建模、反演以及机器学习应用方向。该数据集包含九类地质模型，涵盖了广泛的电阻率分布范围与结构复杂性。 受限于文件大小，同时为便于在深度学习任务中便捷使用，完整数据集托管于https://huggingface.co/datasets/ws11/OpenEM。 为OpenEM开发的正演建模代码已开源，开源地址为https://github.com/WAL-l/OpenEM。 其三维模型生成器同样已开源，开源地址为https://github.com/WAL-l/Geoelectrical3D。