A method for generating three-dimensional simulated particles based on noise algorithm

Abstract

The present invention discloses a method for generating three-dimensional simulated particles based on a noise algorithm, comprising the following steps: obtaining a basic geometric body; defining a noise space; placing the basic geometric body in the noise space, applying noise, and obtaining the basic shape of the simulated particles; based on the principle of noise superposition, superimposing several layers of high-frequency noise influence on the basic shape of the simulated particles, obtaining simulated particles with real surface features, and completing the generation of three-dimensional simulated particles. The method of the present invention can apply noise influence on the scanned real particle model to generate simulated particles of different shapes that retain the morphological characteristics of the real particles, and by artificially controlling the degree of influence of the noise, the morphological difference between the simulated particles and the original real particles can be controlled. The method of the present invention has a simple principle, efficient calculation, and fast particle generation speed, and can reduce the workload of particle scanning and reduce the cost of generating simulated particles.

Description

Technical Field

[0001] The invention belongs to the technical field of virtual simulation particle generation, and in particular relates to a three-dimensional simulation particle generation method based on a noise algorithm. Background Art

[0002] The discrete element method is a commonly used numerical simulation method, which is often used in the field of geotechnical engineering to simulate the mechanical behavior of granular soils. However, natural soil particles have different shapes, and the shape of soil particles will greatly affect the engineering properties of granular soils. Therefore, more and more scholars have begun to use the discrete element method to simulate particles with realistic shapes in their research.

[0003] How to accurately simulate the real particle morphology is a major challenge in discrete element method simulation. Some researchers use simple spheres or irregular aggregate shapes instead of particles for simulation, almost ignoring the influence of the morphological characteristics of real particles. Some researchers also use spherical balls to represent the real particle morphology in the simulation, but too many spherical units also bring significant computational load. Some researchers also use CT scanning to simulate some soil particles in the soil, but this can only obtain the morphological characteristics of some particles in the soil and cannot fully represent the behavior of the soil. If all soil particles in the soil are scanned, the workload is too large and the cost is high. In order to deal with the above problems, scholars began to study how to generate simulated particles with real particle morphological characteristics.

[0004] The current virtual simulation particle generation technology is mainly based on spherical harmonics to reconstruct the real morphological characteristics of particles. It has high requirements on the order of spherical harmonics, requires a large amount of calculation, and is difficult to achieve simple and fast production of a large number of virtual simulation particles with a wide range of morphological changes. Summary of the invention

[0005] The purpose of the present invention is to provide a method for generating three-dimensional simulated particles based on a noise algorithm with smaller calculation amount and lower cost, so as to realize artificial control of the morphological difference between the three-dimensional simulated particles and the real particles.

[0006] The present invention provides a method for generating three-dimensional simulated particles based on a noise algorithm, comprising the following steps:

[0007] S1. Get basic geometry;

[0008] S2. Define the noise space;

[0009] S3. placing the basic geometric body obtained in step S1 in the noise space obtained in step S2, applying noise, and obtaining the basic shape of the simulated particles;

[0010] S4. Based on the principle of noise superposition, several layers of high-frequency noise influence are superimposed on the basic shape of the simulated particles obtained in step S3 to obtain simulated particles with real surface features, thereby completing the generation of three-dimensional simulated particles.

[0011] Step S1 specifically includes: randomly selecting a number of soil particles that actually exist in nature, obtaining the three-dimensional morphological information of the soil particles, and establishing a soil particle model; or using three-dimensional image processing software to establish a number of simple geometric body mesh models; and using the obtained soil particle model or simple geometric body mesh model as the basic geometric body.

[0012] Step S2 specifically includes: defining a cubic noise space; uniformly spaced pixel points are distributed in the cubic noise space along the length, width and height directions, and a separate noise value is assigned to each pixel point by any noise algorithm; the noise algorithms include Value noise algorithm, Perlin noise algorithm and Worley noise algorithm.

[0013] Step S3 includes the following steps:

[0014] S3.1 places the basic geometric body obtained in step S1 in the noise space obtained in step S2, specifically: read the model grid point coordinate data of the basic geometric body obtained in step S1, and normalize its spatial position and size to the noise space obtained in step S2; the spatial position and size normalization specifically includes adjusting the model grid coordinate data of the basic geometric body to the noise space coordinate range by data processing; the data processing method includes proportional enlargement, reduction and translation;

[0015] S3.2 applying noise to the basic geometric body placed in the noise space obtained in step S3.1, so that the basic geometric body in the noise space undergoes a morphological change due to the influence of the noise value, thereby obtaining the basic shape of the simulated particle;

[0016] The basic geometric bodies in the noise space undergo morphological changes due to the influence of the noise value, which is achieved using the following formula:

[0017]

[0018] Where ρ is the distance from any grid point on the basic geometric model to the center of the basic geometric body; ρ max is the maximum value of the distance from all grid points on the basic geometric model to the center of the basic geometric body; ρ′ is the new distance from the grid point corresponding to ρ to the center of the basic geometric body under the influence of the noise value; Noise(f) is the noise value of the grid point corresponding to ρ when the frequency is f; the noise value of the grid point is calculated by using the trilinear interpolation method for the noise values ​​of the adjacent pixels of the point, and is calculated using the following formula:

[0019]

[0020] Among them, A, B, C, D, E, F, G, and H are the pixel points of the vertices of the smallest square surrounded by pixels that can enclose the model grid point. The naming rules are as follows: for the 4 pixel points below the grid point in the z-axis direction, the pixel point with the smallest coordinate value in both the x-axis and y-axis directions is A, and the remaining points are named B, C, and D in a counterclockwise direction; similarly, for the 4 pixel points above the grid point in the z-axis direction, the pixel point with the smallest coordinate value in both the x-axis and y-axis directions is E, and the remaining points are named F, G, and H in a counterclockwise direction. Noise(·) is the noise value assigned to the pixel point by the noise algorithm; Noise 0 Noise is the noise value of the grid point on the lower surface of the grid in the z-axis direction; 1 is the noise value of the projection point on the upper surface of the grid in the z-axis direction; the calculation method of the local coordinates (X, Y, Z) of the grid point (x, y, z) in the local coordinate system is as follows:

[0021]

[0022] Among them, (x 0 ,y 0 ,z 0 ) is the coordinate of pixel point A in the spatial rectangular coordinate system; (x 1 ,y 0 ,z 0 ) is the coordinate of pixel point B in the spatial rectangular coordinate system; (x 0 ,y 1 ,z 0 ) is the coordinate of pixel point D in the spatial rectangular coordinate system; (x 0 ,y 0 ,z 1 ) is the coordinate of the pixel point E in the spatial rectangular coordinate system; (x 1 ,y 0 ,z 1 ) is the coordinate of the pixel point F in the spatial rectangular coordinate system; (x 0 ,y 1 ,z 1 ) is the coordinate of pixel point H in the spatial rectangular coordinate system.

[0023] The noise superposition principle in step S4 is the fractal noise principle; the influence of superimposing several layers of high-frequency noise is achieved by the following formula:

[0024]

[0025] Among them, f i is the frequency of the noise in the i-th layer; f 0is the frequency of the lowest frequency noise among the several layers of high frequency noise, f 0 is the default setting value; Noise(f i ) is the frequency f i N is the number of noises in the operation of superimposing several layers of high-frequency noise; Fractal noise is the sum of the noise applied in step S3 and the N layers of high-frequency noise with increasing frequencies and halved amplitudes superimposed in step S4.

[0026] The present invention discloses a method for generating three-dimensional simulated particles based on a noise algorithm. The method utilizes the unique noise performance effect of the noise algorithm, applies noise influence to basic geometric bodies in the noise space, and subsequently superimposes noise, thereby generating a large number of simulated particles with noise characteristics, thereby solving the problem of simply and quickly generating a large number of particles with a wide range of morphological changes that is difficult to achieve with the current prior art. The present invention can apply noise influence to the scanned real particle model, generate simulated particles of different shapes that retain the morphological characteristics of the real particles, and control the size of the morphological difference between the simulated particles and the original real particles by artificially controlling the influence of the noise. The method of the present invention has a simple principle, high computational efficiency, and fast particle generation speed, and can reduce the workload of particle scanning and the cost of generating simulated particles. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a schematic diagram of the process of the present invention;

[0028] Figure 2 This is a schematic diagram of an embodiment of the present invention in which a spherical mesh model is used as a basic geometric body and placed in a cubic noise space;

[0029] Figure 3 This is a schematic diagram of the basic shape of simulated particles generated based on the Worley noise algorithm in an embodiment of the present invention;

[0030] Figure 4 This is a schematic diagram of obtaining three-dimensional simulated particles by superimposing noise on the basic shape in an embodiment of the present invention;

[0031] Figure 5 A schematic diagram of three-dimensional simulated particles generated based on Value noise, Perlin noise and Worley noise algorithms in an embodiment of the present invention;

[0032] Figure 6 This is a schematic diagram of generating three-dimensional simulated particles using scanned real particles as basic geometric bodies in an embodiment of the present invention. DETAILED DESCRIPTION

[0033] The present invention provides a method for generating three-dimensional simulated particles based on a noise algorithm, and its flow chart is as follows: Figure 1 As shown, the following steps are included:

[0034] S1. Obtaining basic geometric bodies, specifically: randomly selecting a number of soil particles that actually exist in nature, using CT scanning to obtain the three-dimensional morphological information of the soil particles, and establishing a soil particle model; or using three-dimensional image processing software to establish a number of simple geometric body mesh models; and using the obtained soil particle model or simple geometric body mesh model as the basic geometric body.

[0035] S2. Define a noise space, specifically: define a cubic noise space; the cubic noise space has evenly spaced pixel points distributed along the length, width and height directions, and any noise algorithm assigns a separate noise value to each pixel point; the noise algorithms include Value noise algorithm, Perlin noise algorithm and Worley noise algorithm.

[0036] S3. placing the basic geometric body obtained in step S1 in the noise space obtained in step S2, applying noise, and obtaining the basic shape of the simulated particles, including the following steps:

[0037] S3.1 places the basic geometric body obtained in step S1 in the noise space obtained in step S2, specifically: read the model grid point coordinate data of the basic geometric body obtained in step S1, and normalize its spatial position and size to the noise space obtained in step S2; the spatial position and size normalization specifically includes adjusting the model grid coordinate data of the basic geometric body to the noise space coordinate range by data processing; the data processing method includes proportional enlargement, reduction and translation;

[0038] S3.2 applying noise to the basic geometric body placed in the noise space obtained in step S3.1, so that the basic geometric body in the noise space undergoes a morphological change due to the influence of the noise value, thereby obtaining the basic shape of the simulated particle;

[0039] The basic geometric bodies in the noise space undergo morphological changes due to the influence of the noise value, which is achieved using the following formula:

[0040]

[0041] Where ρ is the distance from any grid point on the basic geometric model to the center of the basic geometric body; ρ max is the maximum value of the distance from all grid points on the basic geometric model to the center of the basic geometric body; ρ′ is the new distance from the grid point corresponding to ρ to the center of the basic geometric body under the influence of the noise value; Noise(f) is the noise value of the grid point corresponding to ρ when the frequency is f; the noise value of the grid point is calculated by using the trilinear interpolation method for the noise values ​​of the adjacent pixels of the point, and is calculated using the following formula:

[0042]

[0043] Among them, A, B, C, D, E, F, G, and H are the pixel points of the vertices of the smallest square surrounded by pixels that can enclose the model grid point. The naming rules are as follows: for the 4 pixel points below the grid point in the z-axis direction, the pixel point with the smallest coordinate value in both the x-axis and y-axis directions is A, and the remaining points are named B, C, and D in a counterclockwise direction; similarly, for the 4 pixel points above the grid point in the z-axis direction, the pixel point with the smallest coordinate value in both the x-axis and y-axis directions is E, and the remaining points are named F, G, and H in a counterclockwise direction. Noise(·) is the noise value assigned to the pixel point by the noise algorithm; Noise 0 Noise is the noise value of the grid point on the lower surface of the grid in the z-axis direction; 1 is the noise value of the projection point on the upper surface of the grid in the z-axis direction; the calculation method of the local coordinates (X, Y, Z) of the grid point (x, y, z) in the local coordinate system is as follows:

[0044]

[0045] Among them, (x 0 ,y 0 ,z 0 ) is the coordinate of pixel point A in the spatial rectangular coordinate system; (x 1 ,y 0 ,z 0 ) is the coordinate of pixel point B in the spatial rectangular coordinate system; (x 0 ,y 1 ,z 0 ) is the coordinate of pixel point D in the spatial rectangular coordinate system; (x 0 ,y 0 ,z 1 ) is the coordinate of the pixel point E in the spatial rectangular coordinate system; (x 1 ,y 0 ,z 1 ) is the coordinate of the pixel point F in the spatial rectangular coordinate system; (x 0 ,y 1 ,z 1 ) is the coordinate of pixel point H in the spatial rectangular coordinate system.

[0046] S4. Based on the principle of noise superposition, several layers of high-frequency noise are superimposed on the basic shape of the simulated particles obtained in step S3 to obtain simulated particles with real surface features, thereby completing the generation of three-dimensional simulated particles;

[0047] The noise superposition principle is the fractal noise principle; the effect of superimposing several layers of high-frequency noise is achieved by the following formula:

[0048]

[0049] Among them, f i is the frequency of the noise in the i-th layer; f 0 is the frequency of the lowest frequency noise among the several layers of high frequency noise, f 0 is the default setting value; Noise(f i ) is the frequency f i N is the number of noises in the operation of superimposing several layers of high-frequency noise; Fractal noise is the sum of the noise applied in step S3 and the N layers of high-frequency noise with increasing frequencies and halved amplitudes superimposed in step S4.

[0050] The method of the present invention is further described below in conjunction with an embodiment:

[0051] A sphere mesh model was created using the 3D image processing software Rhino and exported as an STL file. The model surface has a total of 40,962 mesh points as the basic geometry.

[0052] A cubic noise space with an edge length of 2 units is established in the range of 0-2 units of the x, y and z axes in the spatial rectangular coordinate system. In the noise space, 10×10×10 pixels are evenly distributed along the length, width and height directions. Each pixel is assigned a separate noise value by the Worley noise algorithm.

[0053] The model grid point coordinate data of the STL file obtained in step 1 is read and adjusted by proportional enlargement, reduction and translation, so as to normalize the spatial position and size of the basic geometric body to the noise space in step 2, such as Figure 2 As shown;

[0054] Apply noise to generate the basic shape of simulated particles: the basic geometric bodies in the noise space change their shape due to the influence of the noise value, and the basic shape of the simulated particles is obtained, such as Figure 3 As shown;

[0055] Noise superposition generates simulated particles with real surface features, such as Figure 4 shown.

[0056] The above steps are to generate three-dimensional simulated particles based on the Worley noise algorithm using the method of the present invention. By adjusting the algorithm parameters, a large number of particles with a wide range of morphological changes can be produced simply and quickly.

[0057] This embodiment also generates three-dimensional simulated particles based on value noise, Perlin noise and worley noise algorithms, such as Figure 5As shown. By observing Figure 5 The characteristics of different noise generating particles can be compared and analyzed.

[0058] In the above results and schematic diagrams, only the spherical grid model is used as the basic geometric body to produce three-dimensional simulated particles. In order to make the generated simulated particles closer to the morphological characteristics of soil particles that actually exist in nature, this embodiment also uses the scanned real particles as the basic geometric body to produce three-dimensional simulated particles through the method of the present invention. By adjusting the algorithm parameters, the size of the morphological difference between the simulated particles and the real particles can be artificially controlled, such as Figure 6 As shown, these simulated particles of different shapes retain the morphological characteristics of the original real particles.

Claims

1. A three-dimensional simulation particle generation method based on a noise algorithm, characterized in that: The following steps are involved: S1. Get the basic geometry, specifically: Randomly select a number of soil particles that actually exist in nature, obtain the three-dimensional morphological information of the soil particles, and establish a soil particle model; or use three-dimensional image processing software to establish a number of simple geometric body mesh models; and use the obtained soil particle model or simple geometric body mesh model as a basic geometric body; S2. Define the noise space; S3. placing the basic geometric body obtained in step S1 in the noise space obtained in step S2, applying noise, and obtaining the basic shape of the simulated particles; The basic geometric bodies in the noise space undergo morphological changes due to the influence of the noise value, which is achieved using the following formula: Where ρ is the distance from any grid point on the basic geometric model to the center of the basic geometric body; ρ max is the maximum value of the distance from all grid points on the basic geometric model to the center of the basic geometric body; ρ′ is the new distance from the grid point corresponding to ρ to the center of the basic geometric body under the influence of the noise value; Noise(f) is the noise value of the grid point corresponding to ρ when the frequency is f; the noise value of the grid point is calculated by using the trilinear interpolation method for the noise values ​​of the adjacent pixels of the point, and is calculated using the following formula: Among them, A, B, C, D, E, F, G, and H are the pixel points of the vertices of the smallest square surrounded by the pixel points that can enclose the model grid point. The naming rules are as follows: for the 4 pixel points below the grid point in the z-axis direction, the pixel point with the smallest coordinate values ​​in both the x-axis and y-axis directions is A, and the remaining points are named B, C, and D in a counterclockwise direction; similarly, for the 4 pixel points above the grid point in the z-axis direction, the pixel point with the smallest coordinate values ​​in both the x-axis and y-axis directions is E, and the remaining points are named F, G, and H in a counterclockwise direction; Noise(·) is the noise value assigned to the pixel point by the noise algorithm; Noise0 is the noise value of the lower surface projection point of the grid point in the z-axis direction; Noise1 is the noise value of the upper surface projection point of the grid point in the z-axis direction; the calculation method of the local coordinates (X, Y, Z) of the grid point (x, y, z) in the local coordinate system is as follows: Among them, (x0, y0, z0) are the coordinates of pixel point A in the spatial rectangular coordinate system; (x1, y0, z0) are the coordinates of pixel point B in the spatial rectangular coordinate system; (x0, y1, z0) are the coordinates of pixel point D in the spatial rectangular coordinate system; (x0, y0, z1) are the coordinates of pixel point E in the spatial rectangular coordinate system; (x1, y0, z1) are the coordinates of pixel point F in the spatial rectangular coordinate system; (x0, y1, z1) are the coordinates of pixel point H in the spatial rectangular coordinate system; S4. Based on the principle of noise superposition, several layers of high-frequency noise are superimposed on the basic shape of the simulated particles obtained in step S3 to obtain simulated particles with real surface features, thereby completing the generation of three-dimensional simulated particles; The noise superposition principle is the fractal noise principle; the effect of superimposing several layers of high-frequency noise is achieved by the following formula: Among them, f i is the frequency of the i-th layer of noise; f0 is the frequency of the lowest frequency noise among the several layers of high-frequency noise, and f0 is a preset setting value; Noise(f i ) is the frequency f i N is the number of noises in the operation of superimposing several layers of high-frequency noise; Fractal noise is the sum of the noise applied in step S3 and the N layers of high-frequency noise with increasing frequencies and halved amplitudes superimposed in step S4.

2. The method for generating three-dimensional simulated particles based on a noise algorithm according to claim 1, characterized in that: Step S2 specifically includes: defining a cubic noise space; uniformly spaced pixel points are distributed in the cubic noise space along the length, width and height directions, and a separate noise value is assigned to each pixel point by any noise algorithm; the noise algorithms include Value noise algorithm, Perlin noise algorithm and Worley noise algorithm.

3. The method for generating three-dimensional simulated particles based on noise algorithm according to claim 2, characterized in that: Step S3 includes the following steps: S3.1 places the basic geometric body obtained in step S1 in the noise space obtained in step S2, specifically: read the model grid point coordinate data of the basic geometric body obtained in step S1, and normalize its spatial position and size to the noise space obtained in step S2; the spatial position and size normalization specifically includes adjusting the model grid coordinate data of the basic geometric body to the noise space coordinate range by data processing; the data processing method includes proportional enlargement, reduction and translation; S3.2 applies noise to the basic geometric body placed in the noise space obtained in step S3.

1. The basic geometric body in the noise space undergoes morphological changes due to the influence of the noise value, thereby obtaining the basic shape of the simulated particles.

Images