Method of separating, identifying and characterizing cracks in 3D space

A method of separating, identifying and characterizing cracks in three-dimensional (3D) space for a digital volumetric image, the method comprising:

• preprocessing the digital volumetric image, the digital volumetric image comprised of voxels;
• statistically analyzing basic information of the digital volumetric image to determine a plurality of void structures in the digital volumetric image, wherein the basic information includes porosity, connectivity, and position, size, orientation and anisotropy of each of the plurality of void structures;
• determining one or more unnecessary void structures of the plurality of void structures in the digital volumetric image based on a filtering condition, and removing the unnecessary void structures from the digital volumetric image;
• smoothening and mending the digital volumetric image;
• determining three extension lengths in three directions x, y and z of a 3D coordinate system for each of the plurality of void structures, wherein each of the extension length is a length, in voxels, from one to another void-solid boundary of the corresponding void structure in the corresponding direction, and thinning each of the plurality of void structures in the direction x, y, or z in which the extension length is the shortest within the three extension lengths of the corresponding void structure;
• analyzing the voxels of the digital volumetric image to separate each of the plurality of thinned void structures into a plurality of cracks, the plurality of cracks corresponding to an intra-connected structure of cracks in a crack network;
• and
• combining disconnected cracks of the plurality of cracks to form a first crack within the plurality of cracks, combining fragmentary cracks of the plurality of cracks to form a second crack within the plurality of cracks, and restoring a thickness of each crack of the plurality of cracks.