A method and system for real-time microstructure rendering based on primitives

Abstract

The application provides a kind of real-time microstructure rendering method and system based on element, it is related to the field of real-time high realistic rendering based on physics, specific scheme includes: the normal map of the material image to be rendered is discretized element representation and normal distribution function calculation, obtain the pre-computed texture of each element;According to the pre-computed texture, the space transformation and space distribution operation are carried out on the element, to obtain the enhanced element;Based on the enhanced element, the normal distribution function of the shading point is fitted, the real-time rendering of microstructure is carried out, and the final rendering image is obtained;The element microstructure is explicitly expressed, and the diversity of structure is increased by space and space distribution of single element, only a small amount of data needs to be maintained, and the corresponding normal distribution function analytical formula is proposed, so that the rendering of complex microstructure is realized with small memory and performance overhead.

Description

Technical Field

[0001] This invention belongs to the field of physical real-time high-fidelity rendering technology, and particularly relates to a primitive-based real-time microstructure rendering method and system. Background Technology

[0002] The statements in this section are merely background information related to the present invention and do not necessarily constitute prior art.

[0003] In computer graphics, the Bidirectional Reflectance Distribution Function (BRDF) is commonly used to describe the interaction between different materials and light in the real world, and the spatial distribution of energy is a crucial property for describing materials. Currently, mainstream BRDF models for materials are based on microsurface theory. This model assumes that numerous microstructures collectively constitute the macroscopic surface, and the material features observed by the human eye at the macroscopic level are the result of the combined effects of these microstructures. The complex geometric features of these microstructures make the spatial distribution of energy highly tortuous. Accurately reproducing and rendering different types of microstructures has become a crucial issue for improving the realism of materials, attracting significant attention in the field of computer graphics.

[0004] The bidirectional reflectance distribution function uses the normal distribution function (NDF) to describe the normal distribution of microstructures, thereby describing the distribution of energy in different directions. Different types of microstructures create distinctly different macroscopic visual effects, including scratches, shimmering highlights, structural highlights, and leather-like textures. Traditional microsurface theory typically uses a smooth statistical distribution to describe the NDF, which fails to accurately reproduce the material characteristics of the microstructure, resulting in the loss of some high-frequency detail information on a macroscopic level.

[0005] In recent years, in the field of offline rendering of microstructures, a method has been proposed to accurately describe the normal distribution of microstructures using high-precision normal maps. This method relies on textures to describe microstructures and can obtain highly realistic rendering results for different types of microstructures. However, this method also has problems such as excessive performance overhead and heavy storage burden, and cannot be directly ported to real-time rendering. Existing microstructure rendering methods usually model a single type of microstructure. Therefore, there is no real-time and efficient rendering method suitable for multiple types of microstructures. Summary of the Invention

[0006] To overcome the shortcomings of the prior art, this invention provides a real-time microstructure rendering method and system based on primitives, which explicitly expresses the microstructure of primitives and increases the diversity of structures by spatially distributing individual primitives, while only requiring the maintenance of a small amount of data. The invention also proposes a corresponding analytical formula for the normal distribution function, thereby achieving the rendering of complex microstructures with relatively low memory and performance overhead.

[0007] To achieve the above objectives, one or more embodiments of the present invention provide the following technical solutions: The first aspect of this invention provides a method for real-time microstructure rendering based on primitives; A primitive-based real-time microstructure rendering method includes: (1) Discretize the primitive normal map of the material image to be rendered into a mesh representation and calculate the normal distribution function to obtain the pre-computed texture of the primitive; (2) Based on the pre-calculated texture, perform spatial transformation and spatial distribution operations on the primitives to obtain multiple enhanced primitives; (3) Based on the enhanced multiple primitives, fit the normal distribution function of the shading points, and render the microstructure in real time to obtain the final rendered image.

[0008] Furthermore, the discretized mesh representation is specifically as follows: The primitive normal map of the material image to be rendered is decomposed into a regular grid composed of squares. The hierarchical structure is organized by setting different square sizes to construct primitives of different MIP-map levels.

[0009] Furthermore, the pre-computed texture is achieved through preprocessing operations, specifically: Iterate through each primitive square and generate a Gaussian function based on its stored normal information and texel position information, which serves as the normal distribution function for each square. The normal distribution function of each block is encoded as the normal distribution function of the primitive in different directions to obtain the pre-computed texture of the primitive.

[0010] Furthermore, the spatial transformation, based on the spatial transformation matrix, realizes offset transformation, mirror transformation, scaling transformation, shearing transformation, and rotation transformation.

[0011] Furthermore, the spatial distribution includes random disordered distribution and tiled splicing distribution; The random and disordered distribution is simulated by sampling random noise textures to simulate the process of primitives being randomly distributed on the material surface.

[0012] Furthermore, the normal distribution function of the fitted coloring points is specifically as follows: By iterating through all primitives within the range of the shading point and summing their normal distribution functions, the normal distribution function of the shading point can be obtained.

[0013] Furthermore, the real-time rendering of the microstructure also includes handling the case where multiple primitives overlap, specifically as follows: The overlapping region is subdivided into four equal-area sub-regions; the occlusion relationship of the overlapping region is defined using the last queried primitive.

[0014] A second aspect of the present invention provides a real-time microstructure rendering system based on primitives.

[0015] A primitive-based real-time microstructure rendering system includes a structure representation module, a spatial enhancement module, and a real-time rendering module. The structure representation module is configured to: discretize the primitive normal map of the material image to be rendered into a mesh representation and calculate the normal distribution function to obtain the pre-computed texture of the primitive; The spatial enhancement module is configured to: perform spatial transformation and spatial distribution operations on primitives based on pre-calculated textures to obtain multiple enhanced primitives; The real-time rendering module is configured to: fit the normal distribution function of the shading points based on the enhanced multiple primitives, and perform real-time rendering of the microstructure to obtain the final rendered image.

[0016] A third aspect of the present invention provides a computer-readable storage medium having a program stored thereon that, when executed by a processor, implements the steps of a primitive-based real-time microstructure rendering method as described in the first aspect of the present invention.

[0017] A fourth aspect of the present invention provides an electronic device including a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of a primitive-based real-time microstructure rendering method as described in the first aspect of the present invention.

[0018] The above one or more technical solutions have the following beneficial effects: This invention assumes that the surface of a material is composed of a large number of primitive microstructures. Since most microstructures in the real world can be abstracted and represented as primitives, the algorithm proposed in this invention is applicable to a large number of material types with independent primitive characteristics, and has a wide range of expression types and strong expressive power.

[0019] This invention provides a pre-computation-based explicit primitive microstructure representation method. In this method, the primitive microstructure is described by a small-sized normal map. The normal distribution function of the primitive microstructure in different directions is pre-computed and encoded into the pre-computation texture, thereby realizing the description of a single primitive. Unlike previous methods that preprocess the macroscopic surface, the pre-computation of this invention is only performed on the primitive, which greatly improves the processing time and reduces the amount of preprocessed data generated.

[0020] This invention provides a method to increase the diversity of structures by performing spatial transformation and spatial distribution operations on individual primitive microstructures. The method involves performing spatial transformations on the primitives, such as scaling, offsetting, rotating, shearing, and mirroring, which greatly enhances the structural diversity of the macroscopic surface composed of primitive microstructures. Furthermore, different spatial transformation operations are proposed for different types of microstructures to further enrich the visual effects.

[0021] This invention provides an analytical formula for the normal distribution function based on primitives. It obtains a pre-computed texture through rapid sampling of primitive microstructures during preprocessing, efficiently solves for the normal distribution function of shading points, completes the rendering of the microstructure, and obtains the visual effect of the microstructure at the macroscopic level. Unlike previous methods with significant memory overhead, this method only requires a small amount of data on the primitive microstructure, keeping overall performance and memory usage to a low limit, making it suitable for real-time rendering.

[0022] Advantages of additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0023] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0024] Figure 1 This is a flowchart of the method in the first embodiment; Figure 2 This is a schematic diagram of the spatial transformation of the basic microstructure in the first embodiment; Figure 3 This is a comparison chart of the microstructure rendering method of the first embodiment and the control method on the scratched material; Figures 4(a)-4(c) show the rendering results of three different types of microstructures in the first embodiment: anisotropic noise, structured type, and scratch. Figure 5 The first embodiment shows the rendering results of five different types of microstructures in the same scene: anisotropic noise, brushed metal, leather, isotropic noise, and structured type. Figures 6(a)-6(d) are comparison diagrams of the rendering results of the primitive microstructure of the first embodiment under different spatial transformations and spatial distributions. Detailed Implementation

[0025] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0026] Real-time rendering of microstructures is an important issue that urgently needs to be addressed. This invention proposes a primitive-based method for representing and rendering microstructures. The algorithm innovatively assumes that the material surface is a unified whole composed of a large number of primitive microstructures. It proposes a method for discretizing primitives, as well as methods for spatial distribution and spatial transformation of primitives to increase their diversity, and a corresponding analytical formula for the normal distribution function. This algorithm is applicable to a large number of self-similar materials in the real world and can efficiently render microstructures and restore the detailed features of materials with less memory burden, lower computational overhead, and less preprocessing time.

[0027] Example 1 This embodiment discloses a real-time microstructure rendering method based on primitives; like Figure 1 As shown, a primitive-based real-time microstructure rendering method includes: Step S1: Discretize the primitive normal map of the material image to be rendered into a mesh representation and calculate the normal distribution function to obtain the pre-computed texture of the primitive.

[0028] First, a small-sized normal map is input as a primitive. For self-similar materials, this embodiment uses 16. A smaller 16-pixel normal map is used, while a larger 32-pixel normal map is used for textures such as scratches that require higher continuity. 32 normal map.

[0029] Next, the primitives are decomposed into a regular grid of squares. The hierarchical structure is organized by setting different square sizes, similar to the organization of MIP-maps in texture space. In this embodiment, the bottom-level square size is set to 1. 1, and is defined as Level 0 (LOD), while the texture area covered by the squares of Level 1 is 2. 2. Its normal value is the average of the normals of the four level 0 squares in the tangent space, and the same applies to the squares of other levels, such as... Figure 1 As shown in (a). In this embodiment, five different levels of squares are usually sufficient to represent the microstructure at different granularities.

[0030] In this embodiment, it is assumed that the material surface is a whole composed of basic microstructures, therefore the range of the tinting point footprint is... Normal distribution function within It is the result of the joint fitting of the microstructures of all the basic elements, which can be expressed as formula (1).

[0031] (1) in, Represents a basic microstructure within the footprint range. The normal distribution function value, It's about the query location. The Gaussian function is given by s, where s is the query direction, m is the number of primitives within the footprint range, and du is the derivative with respect to the query position.

[0032] This embodiment obtains the normal distribution function by sampling pre-computed textures, instead of calculating the normal distribution function in real time during real-time rendering. Figure 1 As shown in (b), during the preprocessing process, each square in the input normal map texture is traversed. And based on its stored normal information and texel position information, a standard deviation is generated as follows: The Gaussian function is used to calculate... The normal distribution function is encoded into the pre-computed texture. In the middle, the first Hierarchical grid Pre-computed textures are defined as In the pre-computed texture, each pixel represents the query direction. normal distribution function value The calculation of the normal distribution function corresponding to the footprint can be expressed as shown in formula (2): (2) Where m is the number of primitives within the footprint range, and n is the number of squares corresponding to each primitive. It's about the query location. The Gaussian function.

[0033] In this embodiment, each square at different levels corresponds to a pre-computed texture. Generally, higher-precision pre-computed textures can more accurately reproduce the geometric features of microstructures, but this also increases the performance burden during preprocessing, increases data storage pressure, and incurs more performance overhead during real-time rendering. In this embodiment, 64 A pre-computed texture of size 64 is sufficient to reproduce most of the detailed features of the microstructure. However, for microstructures with more complex spatial distribution, such as scratches, this embodiment uses a size of 128. A pre-computed texture of size 128 is used for description; at the same time, this embodiment splices the pre-computed textures corresponding to different squares at each level into a larger pre-computed texture, and organizes the large-size pre-computed textures at different levels into a texture set, thereby optimizing the texture sampling efficiency in the real-time calculation process.

[0034] In this embodiment, preprocessing is performed only on primitives rather than globally, which greatly saves preprocessing time and memory overhead. Table 1 lists the pre-computed texture size, memory overhead, and preprocessing time for different types of microstructures in this embodiment.

[0035] Table 1. Statistics on pre-computed textures, memory overhead, and preprocessing time for different types of microstructures.

[0036] Step S2: Based on the pre-computed texture, perform spatial transformation and spatial distribution operations on the primitives to obtain multiple enhanced primitives.

[0037] This embodiment provides a method for enhancing the diversity of the appearance of primitive microstructures through spatial transformation and spatial distribution, avoiding the generation of repetitive visual features. This includes: offset transformation, mirror transformation, scaling transformation, shearing transformation, and rotation transformation, etc. Figure 1 (c) Spatial distribution includes: random distribution and tiling distribution, such as Figure 1 (d).

[0038] This embodiment describes microstructures using primitives, and these microstructures are defined by different levels of grids at different granularities. Spatial transformations of the primitive microstructures are also based on these grids. Input query direction , square In the relative spatial position of the primitive normal map and positioning level Then, query the corresponding pre-computed texture. normal distribution function The process can be represented by formula (3).

[0039] (3) in, This indicates the index of the cell within the primitive. This indicates the process of sampling the corresponding input parameters in the pre-computed texture.

[0040] In this embodiment, the transformation of the microstructure of the primitives depends on the spatial transformation matrix. Formula (4) is the spatial transformation matrix. The unfolded form: (4) By controlling the space transformation matrix In , , , Four elements can achieve Figure 2 The different types of spatial transformation operations shown in the text are as follows: (1) Scaling transformation of microstructure, the spatial transformation matrix of the scaling transformation is expressed in the form of formula (5), where , These are the scaling factors, controlled separately. direction and The scaling factor in the direction can be used to transform the scale of the microstructure through scaling transformation.

[0041] (5) (2) The offset transformation of the microstructure, the spatial transformation matrix of the offset transformation is expressed by formula (6), where , They are direction, Displacement in a direction, displacement transformation can realize the overall displacement of microstructure in two-dimensional space; (6) (3) Mirror transformation of microstructure: The new microstructure obtained by the primitive through mirror symmetry transformation is also mirror symmetric to the original microstructure in terms of specular distribution. The axisymmetric spatial transformation matrix can be expressed as formula (7).

[0042] (7) (4) The shear transformation of the microstructure can be expressed as formula (8), where , They are along shaft and The stretching coefficient of the axis can be used to control the direction of specular diffusion in the macroscopic rendering result through the shear transformation of the microstructure.

[0043] (8) (5) Rotate the basic microstructure counterclockwise The angle, its rotation transformation matrix can be expressed as formula (9), and the diverse distribution of microstructure geometric features in two-dimensional space can be achieved through rotation operation.

[0044] (9) The spatial transformation of the primitive is achieved by performing spatial transformations on all the squares that constitute the primitive. Multiple enhanced primitives are obtained under different spatial transformation matrices M.

[0045] This embodiment obtains the normal distribution function of primitives after spatial transformation during the rendering process by sampling pre-computed textures in real time. The process of sampling the normal distribution term of primitives after spatial transformation can be expressed as formula (10), through multiple basic spatial transformation matrices. By multiplying multiple primitives, a richer spatial transformation effect can be achieved. In formula (10), It is a hierarchy In the primitive normal map, the relative position is square In the query direction is The normal distribution function on.

[0046] (10) In this embodiment, the microstructure of the basic elements needs to be distributed on the surface of the material after spatial transformation. There are two ways to distribute the basic elements: random disordered distribution and flat distribution.

[0047] Users need to use prior knowledge to determine whether the material has self-similarity and is suitable for splicing. For example, for microstructures with unique structural features, such as brushed metal, randomly distributing the basic microstructures on the material surface will lead to the loss of the original structural features. For this type of structure, a flat splicing distribution method is used.

[0048] This embodiment employs a random and disordered method to distribute the microstructure of elementary elements for materials with indistinct structural features such as scratches. This random process needs to possess strong randomness, high efficiency, and the ability to be reverse-mapped. Figure 1 Step (d) shows that this embodiment simulates the random distribution of primitives by sampling random noise textures. Random numbers are encoded in the noise textures, and the sampled random numbers are used to place primitive microstructures at different spatial locations. At the same time, the scaling factor, offset factor and other parameters in the transformation matrix of the primitive microstructures described in S201-S205 are also sampled from the noise textures, thereby further enriching the diversity of the structure.

[0049] Figure 6 illustrates the impact of spatial transformation or spatial distribution operations on high-frequency information at the macroscopic level on the microstructure of primitives within the deer statue scene. Figure 6(a) shows the rendering result after tiling and splicing primitives of the brushed metal type, with a uniform rotation transformation matrix set for each primitive microstructure. The specular highlights in different diffusion directions are shown in Figure 6(b); Figure 6(c, d) shows the spatial distribution of randomly placed anisotropic noise type primitives. The rendering result below.

[0050] Step S3: Based on the enhanced multiple primitives, fit the normal distribution function of the shading points, and perform real-time rendering of the microstructure to obtain the final rendered image.

[0051] The microstructure rendering algorithm can be divided into two main stages: the preprocessing stage and the real-time rendering stage. This embodiment proposes a method for solving the normal distribution function of the primitive microstructure in the real-time rendering stage, expressed as follows: Figure 1 Step (e).

[0052] Assuming the footprints corresponding to the already colored points This allows us to obtain the set of primitives covered within the footprint area. By traversing all primitives in the set and summing their normal distribution functions, we can obtain the set of primitives covered within the footprint area. The fitted normal distribution function can be obtained. The normal distribution function of a single primitive is the sum of the values ​​of all its squares. The result of the accumulation, and the normal distribution function corresponding to the square can be obtained from the preprocessed texture, this process is expressed as formula (11).

[0053] (11) Different levels of squares correspond to pre-computed textures of the corresponding levels. In this embodiment, a coarser-grained level... Since fewer queries are needed in the pre-computed texture, the computation of the normal distribution function is more efficient. Therefore, the query strategy followed in this embodiment is to prioritize querying pre-computed data at higher levels. When it is determined that the microstructure of the primitive is completely within the footprint, the normal distribution function value will be searched for in the pre-computed texture at a higher level. However, when the primitive part is within the footprint, the level of the pre-computed texture will be gradually reduced from the highest level to determine whether the smaller-grained squares are within the footprint, until the finest-grained level 0 pre-computed texture is subdivided. Figure 1 Step (e) in the diagram illustrates this spatial subdivision process.

[0054] This embodiment needs to handle the situation where multiple primitive microstructures overlap during the real-time rendering stage. To address this, the overlapping region is first divided into four identical sub-regions, and the occlusion relationship of the overlapping region is defined by default using the last queried primitive microstructure.

[0055] This embodiment uses high-intensity point light combined with ambient lighting to illuminate the entire scene. In the real world, light comes from all directions and is reflected by objects in the scene, making it difficult to solve for integration in real time. This embodiment uses distant light probes to encode ambient lighting, which is an image-based lighting method that accelerates specular reflection and diffuse reflection in real-time rendering through preprocessing.

[0056] The platform used in this embodiment is a personal computer, with an NVIDIA RTX TITAN graphics processor, an Intel(R) i9-9900K CPU with a main frequency of 3.6GHz, and 32GB of memory. All rendering results in this embodiment are at 1920. Tested at 1080p screen space. Offline rendering typically converts high-precision normal maps to Gaussian functions and then queries and calculates the Gaussian function in real-time rendering. This embodiment uses this as a control method.

[0057] Figure 3 The image shows a comparison of the rendering results of the proposed rendering method and the control method in a metal sheet scene. The two are almost visually equivalent. The rendering method in this embodiment is applicable to different types of microstructures. Figures 4(a), 4(b), and 4(c) show the different details of three different types of microstructures, namely anisotropic noise, structural material, and scratches, in the rendering results. The upper right corner also shows the input primitive normal map. Figure 5 The demonstration showcases the rendering of various microstructures within the same scene, including anisotropic noise, isotropic noise, brushed metal, leather, and structured materials. It also compares the rendering results, memory overhead, and rendering time with control methods.

[0058] This implementation statistically analyzed the microstructure types in different scenarios, along with their rendering time and memory consumption, and listed them in Table 2. A rendering time of 1920 was achieved in all test scenarios. Real-time rendering in 1080p screen space (>=60 fps).

[0059] Table 2. Statistics on microstructure types, rendering time, and memory overhead for different scenarios.

[0060] Example 2 This embodiment discloses a real-time microstructure rendering system based on primitives; A primitive-based real-time microstructure rendering system includes a structure representation module, a spatial enhancement module, and a real-time rendering module. The structure representation module is configured to: discretize the primitive normal map of the material image to be rendered into a mesh representation and calculate the normal distribution function to obtain the pre-computed texture of the primitive; The spatial enhancement module is configured to: perform spatial transformation and spatial distribution operations on primitives based on pre-calculated textures to obtain multiple enhanced primitives; The real-time rendering module is configured to: fit the normal distribution function of the shading points based on the enhanced multiple primitives, and perform real-time rendering of the microstructure to obtain the final rendered image.

[0061] Example 3 The purpose of this embodiment is to provide a computer-readable storage medium.

[0062] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a primitive-based real-time microstructure rendering method as described in Embodiment 1 of this disclosure.

[0063] Example 4 The purpose of this embodiment is to provide an electronic device.

[0064] An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps in a primitive-based real-time microstructure rendering method as described in Embodiment 1 of this disclosure.

[0065] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for rendering real-time microstructures based on primitives, characterized in that, include: (1) Discretize the primitive normal map of the material image to be rendered into a mesh representation and calculate the normal distribution function to obtain the pre-computed texture of the primitive; (2) Based on the pre-calculated texture, perform spatial transformation and spatial distribution operations on the primitives to obtain multiple enhanced primitives; (3) Based on the enhanced multiple primitives, fit the normal distribution function of the shading points, and render the microstructure in real time to obtain the final rendered image; The pre-computed texture is achieved through preprocessing operations, specifically: Iterate through each primitive square and generate a Gaussian function based on its stored normal information and texel position information, which serves as the normal distribution function for each square. The normal distribution function of each block is encoded as the normal distribution function of the primitive in different directions to obtain the pre-computed texture of the primitive; The transformation of the microstructure of the elementary elements depends on the spatial transformation matrix. The formula is the spatial transformation matrix. The unfolded form: By controlling the spatial transformation matrix In , , , The four elements implement different types of spatial transformation operations: offset transformation, mirror transformation, scaling transformation, shearing transformation, and rotation transformation; After spatial transformation, the basic microstructure needs to be distributed on the material surface. There are two ways to distribute the basic elements: random and disordered distribution and flat distribution. For brushed metals with unique structural characteristics, a flat and spliced ​​distribution method is used. For materials with indistinct scratch structure characteristics, a random and disordered method is used to distribute the basic microstructure.

2. The method for real-time microstructure rendering based on primitives as described in claim 1, characterized in that, The discretized grid representation is specifically as follows: The primitive normal map of the material image to be rendered is decomposed into a regular grid composed of squares. The hierarchical structure is organized by setting different square sizes to construct primitives of different MIP-map levels.

3. The method for real-time microstructure rendering based on primitives as described in claim 1, characterized in that, The spatial distribution includes random and disordered distribution as well as tiled and spliced ​​distribution; The random and disordered distribution is simulated by sampling random noise textures to simulate the process of primitives being randomly distributed on the material surface.

4. The method for real-time microstructure rendering based on primitives as described in claim 1, characterized in that, The normal distribution function of the fitted coloring points is specifically as follows: By iterating through all primitives within the range of the shading point and summing their normal distribution functions, the normal distribution function of the shading point can be obtained.

5. The method for real-time microstructure rendering based on primitives as described in claim 1, characterized in that, The real-time rendering of the microstructure also includes handling the case where multiple primitives overlap, and the specific processing method is as follows: The overlapping region is subdivided into four equal-area sub-regions; the occlusion relationship of the overlapping region is defined using the last queried primitive.

6. A primitive-based real-time microstructure rendering system employs the primitive-based real-time microstructure rendering method as described in any one of claims 1-5, characterized in that, It includes a structural representation module, a spatial enhancement module, and a real-time rendering module: The structure representation module is configured to: discretize the primitive normal map of the material image to be rendered into a mesh representation and calculate the normal distribution function to obtain the pre-computed texture of the primitive; The spatial enhancement module is configured to: perform spatial transformation and spatial distribution operations on primitives based on pre-calculated textures to obtain multiple enhanced primitives; The real-time rendering module is configured to: fit the normal distribution function of the shading points based on the enhanced multiple primitives, and perform real-time rendering of the microstructure to obtain the final rendered image.

7. An electronic device, characterized in that it comprises: Memory is used to store computer-readable instructions in a non-transitory manner. as well as Processor, for executing the computer-readable instructions, When the computer-readable instructions are executed by the processor, they perform the method described in any one of claims 1-5.

8. A storage medium, characterized in that, The computer-readable instructions are stored non-transitory, wherein when the non-transitory computer-readable instructions are executed by a computer, the instructions of the method according to any one of claims 1-5 are executed.

Images