基于压缩感知的金属矿地震最小二乘逆时偏移方法研究

地震勘探方法在深部金属矿资源勘探中发展潜力巨大，同时也面临挑战。由于金属矿地质体的非成层性、非均匀性、尺度小、陡倾构造的特征，常规地震偏移方法对复杂小尺度矿体的成像分辨率提高有限。本项目基于压缩感知理论，改进传统最小二乘逆时偏移方法，联合随机震源编码和稀疏促进反演，减小地震数据维数，同时保持有效信息，极大减少正演和反偏移计算中求解波动方程的次数；并将小尺度、陡倾角、非均匀地质体模型作为先验信息正则化约束成像结果，提高金属矿地质体的成像效果；研究在模型空间和数据空间分别稀疏化的条件下，模型和数据的正演和反偏移对应关系，缩小稀疏化模型空间的未知参量；研究随机震源编码技术，压制小尺度散射体引起的强散射能量，降低串扰噪声引起的成像误差。此外，项目研究的成像方法将应用于实际金属矿地震资料中，包括甘肃金昌镍铜矿区二维地震数据、安徽庐枞金属矿集区二维地震数据以及新疆阿舍勒铜锌矿的三维地震勘探区域等。

Seismic exploration methods have great potential for development in deep metal ore resources exploration, but also face challenges. Due to the characteristics of non-stratification, non-uniformity, small scale and steep dip structure of metal ore geological bodies, conventional seismic migration methods have limited improvement in the imaging resolution of complex small-scale ore bodies. This project is based on the theory of compressed sensing, improving the traditional least-squares inverse time migration method, combining random source coding and sparsity-promoting inversion, reducing the dimension of seismic data, while maintaining effective information, and greatly reducing forward calculation and inverse migration calculations, the number of times to solve the wave equation. We use small-scale, steep dip, non-uniform geological body models as a priori information to regularize constrained imaging results to improve the imaging effect of metal ore geological bodies. We study the sparseness in model space and data space, and under those conditions, the correspondence between the model and data forward and reverse migration reduces the unknown parameters of the sparse model space. We study the random source coding technology to suppress the strong scattering energy caused by small-scale scatters and reduce the imaging error caused by crosstalk noise. In addition, the imaging methods studied in the project will be applied to the field seismic data of metal mines, including 2D seismic data from the Jinchang nickel-copper deposit in Gansu, 2D seismic data from the Luzong mineral deposit in Anhui, and 3D seismic data from the Ashele copper-zinc deposit in Xinjiang.