Research And Application Of Digital Core Reconstruction Method Based On Image Processing

The pore structure characteristics of shale gas reservoirs are of great significance to the exploration and development of shale gas.By reconstructing the digital image of electron microscopic scanning of shale cores in a large scale,the micro-structure of these micro-fractures can be observed more intuitively,and the pore size,morphology and organic matter distribution can be obtained,thus providing basic data for the study of shale gas reservoir-forming mechanism,migration law and development.In order to make up for the shortcomings of existing methods,this article proposes a new digital core reconstruction method based on feature subgraph,which can freely adjust the proportion of features,and optimizes it by image restoration technology.The results show that this method can improve the reconstruction speed and control the weight.The proportion of ore,organic matter and other components in the core image keeps the local structural characteristics of samples well and has a good application prospect.Firstly,several feature subgraphs are selected from the original shale samples as input of training images,then a feature proportion parameter is preset for each training image.Then matching blocks are selected from the corresponding training images according to the selection criteria defined in this paper,and the overlapping regions are stitched by image stitching technology until the whole image is reconstructed.The adjustability of the feature proportion parameters of the algorithm is verified.Compared with CIQ algorithm,this algorithm reconstructs faster and retains important structural features.In order to improve the quality of reconstruction furtherly,this paper then optimizes the method,adds image restoration link before reconstruction,uses Criminisi algorithm to repair the missing part of the background area of training samples,and carries out three groups of experiments for shale samples with different resolutions.The experimental results show that the optimized method can reflect more structural features of the training image in the reconstruction results.Satisfactory results have been obtained for different shale samples.The main work of this paper includes the following sections:In the first part,the background and current situation of digital core research and the concept of digital core are introduced in detail.In the second part,several commonly used digital core reconstruction methods and texture synthesis algorithms are introduced,such as multi-point statistics and image quilting algorithms.In the third part,the digital core reconstruction method based on feature proportion control is introduced,as well as its related details and specific algorithm flow.Shale samples with large amount of feature information are reconstructed to verify the adjustability of the algorithm.The fourth part mainly introduces the process of restoration optimization before the reconstruction of the above algorithm.The Criminisi algorithm is used to repair the voids in the background area of the training image.Three groups of shale samples with different resolution and different feature information are reconstructed,which proves that the optimized method can reconstruct different shale samples effectively.