RMSD_Transformation_Input/Output_Data

This repository contains synthetic and real seismic SEG-Y datasets generated for the development, evaluation, and reproducibility of the RMSD-based seismic waveform transformation framework. The archive includes original seismic sections together with deterministic rule-transformed, machine-learning-transformed, and ML-proxy-transformed seismic datasets. The RMSD framework operates through localized inter-peak waveform conditioning, selectively modifying short positive runs while preserving the dominant reflection peaks associated with principal impedance contrasts. Unlike globally acting filtering or deconvolution methods, the proposed approach performs sparse waveform transformation designed to enhance reflector continuity while minimizing large-scale spectral distortion. The repository includes: Synthetic 15 ms reflector-separation SEG-Y datasets RMSD rule-transformed synthetic datasets Original WS70 seismic dataset RMSD rule-transformed WS70 seismic dataset RMSD ML-transformed WS70 seismic dataset RMSD ML-proxy transformed WS70 seismic dataset Supporting metadata and README documentation The datasets are intended for: seismic waveform-conditioning research, reflector continuity enhancement studies, sparse polarity transformation analysis, benchmarking against conventional seismic enhancement methods, ML-assisted seismic transformation workflows, and reproducible geophysical research. The repository supports investigations into localized waveform conditioning, thin-bed interference behavior, feature-engineered seismic transformation, and scalable ML-assisted seismic processing. Data format:IEEE floating-point SEG-Y本仓库包含为开发、评估基于均方根偏差（Root Mean Square Deviation，RMSD）的地震波形变换框架并实现可复现性研究而生成的合成与真实地震SEG-Y数据集。该归档包含原始地震剖面，以及经确定性规则变换、机器学习变换与ML代理变换的地震数据集。 RMSD框架通过局域化峰间波形调节运作，仅选择性修改短时长正幅值段，同时保留与主阻抗反射界面相关的优势反射峰。与全局作用的滤波或反褶积方法不同，本方法采用稀疏波形变换策略，旨在增强反射层连续性的同时，最小化大规模频谱畸变。 本仓库包含以下内容： 1. 15 ms反射层分离合成SEG-Y数据集 2. 经RMSD规则变换的合成数据集 3. 原始WS70地震数据集 4. 经RMSD规则变换的WS70地震数据集 5. 经RMSD机器学习变换的WS70地震数据集 6. 经RMSD ML代理变换的WS70地震数据集 7. 配套元数据与README文档 本数据集适用于以下研究场景： - 地震波形调节研究 - 反射层连续性增强研究 - 稀疏极性变换分析 - 与常规地震增强方法的基准对比测试 - 机器学习辅助地震变换工作流研究 - 可复现地球物理研究 本仓库支持局域化波形调节、薄层干涉特性分析、特征工程地震变换以及可扩展机器学习辅助地震处理相关研究。 数据格式：IEEE浮点型SEG-Y