Organic matter and its pore distribution are closely related to the adsorbed gas content of shale and the microscopic    gas flow mechanism. Studies have shown that the organic matter distribution in shale has anisotropic characteristics at the pore    scale. However, the existing digital core reconstruction algorithms for shale cannot effectively describe this feature. Aiming at    the anisotropic characteristics of organic matter distribution in shale samples, this paper introduced a two-point cluster function    as a statistical indicator to characterize the continuity of organic matter distribution based on the traditional simulated numerical    annealing reconstruction method for digital cores. Although the two-point cluster function belongs to the two-point probability    

function, it contains high-order structural information. Therefore, the two-point cluster function is a statistical indicator that is    more sensitive to structural changes than the two-point correlation function. This paper presented the mathematical model of the    simulated annealing method coupled with a two-point cluster function. Meanwhile, taking the shale core sample of the Longmaxi    Formation of the Lower Silurian in the Sichuan Basin as an example, the systematic method flow of "sample imaging⸺image    processing⸺function extraction⸺reconstruction calculation" was discussed in detail. Through field emission scanning electron    microscopy (FESEM) and focused ion beam-field emission scanning electron microscopy (FIB-SEM) and image processing, the    shale organic matter and its pore distribution at the pore scale were obtained, and the statistical correlation of core samples in    three orthogonal directions was extracted. Functions include a single-point correlation function, two-point correlation function,    linear path function and a two-point cluster function. The paper established an effective method for three-dimensional recon   struction of organic matter aggregates, and applied the above statistical correlation functions to the reconstruction process of the    Dual-Region simulated annealing method to obtain reconstruction models of organic matter, organic matter pores and matrix    pores respectively. Combining the three coupled regions produces the final digital core 3D reconstruction model of the shale    sample. The reconstruction of the two-point cluster function distribution curve in three orthogonal directions of the shale digital    core shows that the organic matter has obvious anisotropy. By comparing organic matter equivalent volume radius distribution,    organic matter area, organic matter elongation and other parameter distributions of real shale samples and reconstructed shale    digital cores, the validity of the two-point cluster function method to characterize the anisotropic distribution of shale organic    matter has been proved. This method can be applied to related research work on shale digital core reconstruction.