The morphological characteristics (size and shape) of sand particles are critical for the evaluation of the mechanical properties of sand. Accurate quantification of these morphological characteristics is a prerequisite to investigating the mechanical properties of sand. Compared with the three-dimensional (3-D) test method, the two-dimensional (2-D) counterpart has the advantages of rich test types, simple procedures, and low cost.

To investigate the feasibility of 2-D images for 3-D morphological characterization of sand particles, 3-D laser scanning technology is employed in the present study to obtain the point cloud data of sand particles. Furthermore, a computer algorithm is developed to reconstruct the 3-D models of the sand particles. Thus, the 3-D morphological characteristics of the particles are accurately and quickly obtained. Three 2-D methods, namely the random single projection method, maximum and minimum projected area method (selected from 1 000 projections), and three orthogonal direction projection methods, are proposed for obtaining the 2-D morphological parameters of sand particles.

Computer algorithms for reconstructing the 3-D model of sand particles based on a point cloud were proposed. The algorithms were used to automatically acquire the 3-D and 2-D morphological parameters. Thereafter, the correlations between the 3-D and 2-D parameters were analyzed. A practical method was proposed to obtain the 3-D morphological characteristics of the granular material based on the corresponding 2-D morphological characteristics.

The results reveal that there is a large gap between the 2-D parameters obtained from various angle projections, especially the area parameters in various projections. Notably, the error between the maximum and minimum area is as high as 40%. The random single angle projection method does not provide accurate 3-D information. A strong linear correlation exists between the 3-D and 2-D size parameters of the particles, and a moderate correlation exists between the shape parameters. An empirical relationship for estimating the 3-D morphology of sand particles using 2-D parameters is established. The 3-D parameters, namely the length, width, and thickness of the granular material, can be estimated using the corresponding 2-D parameters (R²=0.735-0.994). The fitting effects of the particle size parameters are better than those of the shape parameters. Additionally, the fitting effects of the maximum and minimum projected area and three orthogonal directional projection methods are better than those of the random single projection method. Among the three 2-D methods, the maximum and minimum projected area and three orthogonal direction projection methods can more accurately estimate the 3-D morphological characteristics of sand particles; thus, these are practical methods for the morphological quantification of granular materials.