WO2022111400A1 - Light source sampling weight determination method for multiple light source scenario rendering, and related device - Google Patents

Abstract

A light source sampling weight determination method for multiple light source scenario rendering, and a related device. The method comprises: obtaining a scenario to be rendered in a multiple light source scenario and using a target rendering angle, and obtaining a plurality of light sources arranged in the multiple light source scenario (S201); executing single light source ray tracing rendering on the scenario to be rendered at the target rendering angle, and obtaining first rendering images corresponding to the scenario to be rendered under the sole illumination of each light source (S202); and determining a light source sampling weight corresponding to each light source at the target rendering angle according to irradiation information in each first rendering image (S203), where the light source sampling weights are used for executing multiple light source ray tracing rendering on the scenario to be rendered. The present method can accelerate the convergence speed of multiple light source scenario rendering.

Description

Light source sampling weight determination method and related equipment for multi-light source scene rendering

This application claims the priority of the Chinese patent application filed on November 27, 2020, with the application number 202011360803.X, and the application title is "Method for Determining Light Source Sampling Weights for Multi-Light Source Scene Rendering and Related Equipment", all of which are The contents are incorporated herein by reference.

technical field

The present application relates to the technical field of image processing, in particular to a method for determining a light source sampling weight for multi-light source scene rendering and related devices.

Background technique

Rendering is the process of using software to generate images from models, which can be widely used in games, film and television, animation and other fields. Rendering can include ray tracing rendering, which is a special rendering algorithm in 3D computer graphics that traces the rendering rays emitted from the virtual rendering camera, calculates the process of rendering rays propagating in the rendering scene, and finally renders the mathematics of the scene. Models are rendered as images. In order to achieve realistic rendering effects, multiple light sources are generally set in the rendering scene, and each light source is correspondingly set with its own light source sampling weight. The light source sampling weight determines whether to select the light source as an effective light source to participate in the calculation of ray tracing rendering.

In the prior art, a random determination method is adopted for the light source sampling weight of multiple light sources, that is, the light source rendering weight of each light source is the same. The light source rendering weight of each light source is the same, which makes the convergence speed of multi-light source scene rendering slow.

SUMMARY OF THE INVENTION

The present application provides a method for determining a light source sampling weight for multi-light source scene rendering and the correlation, which can speed up the convergence speed of multi-light source scene rendering.

In a first aspect, an embodiment of the present application discloses a method for determining a light source sampling weight for multi-light source scene rendering. The method includes:

Obtain the scene to be rendered in the multi-light source scene from the target rendering perspective, and obtain the multiple light sources set in the multi-light source scene;

From the target rendering perspective, perform ray tracing rendering of a single light source on the scene to be rendered, to obtain each first rendered image corresponding to the scene to be rendered under the lighting of each light source alone;

According to the irradiance information in each first rendered image, the light source sampling weight corresponding to each of the above-mentioned light sources in the above-mentioned target rendering perspective is determined, and the light source sampling weight is used to perform multi-light source ray tracing rendering on the above-mentioned scene to be rendered.

In this embodiment of the present application, the light source sampling weight corresponding to each light source is determined according to the irradiance information of each first rendered image obtained after each light source performs single light source ray tracing rendering on the scene to be rendered, and the light source sampling weight can make the multi-light source ray tracing rendering process The probability of selecting a light source corresponding to a larger light source sampling weight increases. By implementing the embodiments of the present application, a more suitable light source can be selected during the multi-light source ray tracing rendering process, thereby accelerating the convergence speed of multi-light source scene rendering.

With reference to the first aspect, in a first possible implementation manner, the above method further includes: in the above-mentioned target rendering perspective, based on the light source sampling weights corresponding to the above-mentioned light sources in the above-mentioned target rendering perspective, performing multiple executions on the above-mentioned scene to be rendered. The ray tracing rendering of the light source obtains a second rendered image of the scene to be rendered under the illumination of multiple light sources. Implementing the embodiments of the present application, based on the light source sampling weight of each light source, increases the probability of selecting a light source corresponding to a larger light source sampling weight during the ray tracing rendering process, which can speed up the convergence speed of multi-light source scene rendering.

With reference to the first possible implementation manner of the first aspect, in a second possible implementation manner, the resolution of each of the foregoing first rendered images is smaller than the resolution of the foregoing second rendered images. By implementing the embodiment of the present application, in the single light source ray tracing rendering process, the first rendered image is set as a smaller resolution image, and the light source sampling weight of each light source is calculated, which can reduce the amount of calculation and improve the rendering speed.

With reference to any one of the above possible implementation manners of the first aspect, in a third possible implementation manner, the foregoing ray tracing rendering includes bidirectional path tracing rendering. By implementing the present application, multiple rays can be quickly generated, and the convergence speed can be further accelerated.

With reference to the first aspect, in a fourth possible implementation manner, the irradiation information in each of the above-mentioned first rendered images is determined based on the brightness information corresponding to each of the above-mentioned first rendered images, wherein the The luminance information is obtained by conversion based on the color information in each of the first rendered images.

In combination with any one of the possible implementation manners of the first aspect, in a fifth possible implementation manner, the plurality of light sources include a first light source; the scene to be rendered includes a first intersection and at least one second intersection, the first The intersection is the intersection of the first rendering ray and the first object in the scene to be rendered on the surface of the first object, and the second intersection is the reflected ray of the first intersection and the second object in the scene to be rendered. The intersection point on the surface of the second object and/or the light emitted by each of the above-mentioned light sources and each of the third objects in the scene to be rendered are intersected on the surface of each of the third objects; the first rendering ray is a virtual Sent from the rendering camera to the scene to be rendered, the virtual rendering camera is used to determine the target rendering angle of view;

Performing ray tracing rendering of a single light source on the scene to be rendered from the perspective of target rendering, and obtaining each first rendered image corresponding to the scene to be rendered under the illumination of each light source is specifically implemented as follows:

Under the above-mentioned target rendering perspective, lighting sampling is performed on the above-mentioned first light source based on the above-mentioned first intersection point, so as to obtain the first intersection point of the first intersection point projected on the above-mentioned first rendered image by the above-mentioned first rendering light ray under the single illumination of the first light source. a radiance;

Under the above target rendering perspective, lighting sampling is performed on the first light source based on each second intersection point to obtain the projection of each second intersection point on the first rendered image by the first rendering light under the single illumination of the first light source. each second radiance;

According to the sum of the first radiance and each of the second radiances, obtain the first irradiance of the first intersection point projected on the first rendered image by the first rendering light under the single illumination of the first light source;

Based on the first irradiance and the object attribute of the object where the first intersection is located, determining the rendering display information of the first intersection under the single illumination of the first light source;

Based on the rendering display information of all the first intersection points in the scene to be rendered under the illumination of the first light source, a first rendered image corresponding to the scene to be rendered under the single illumination of the first light source is acquired.

In combination with the first possible implementation manner of the first aspect, in a sixth possible implementation manner, the above-mentioned scene to be rendered includes a target intersection point, and the target intersection point is the intersection of the target rendering ray and any object in the above-mentioned scene to be rendered. An intersection point on the surface of any object; the target rendering light is sent by a virtual rendering camera to the scene to be rendered, and the virtual rendering camera is used to determine the target rendering angle of view;

In the above-mentioned target rendering perspective, based on the light source sampling weights corresponding to the above-mentioned respective light sources in the above-mentioned target rendering perspective, ray tracing rendering of multiple light sources is performed on the above-mentioned scene to be rendered, and a second ray tracing rendering of the to-be-rendered scene under the illumination of multiple light sources is obtained. The rendered image is specifically implemented as:

In the above-mentioned target rendering perspective, multiple illumination samplings are performed on the above-mentioned light sources based on the above-mentioned target intersection points, wherein each illumination sampling corresponds to selecting each target light source from the above-mentioned plurality of light sources, and each target light source is used to perform multiple illumination operations on the above-mentioned target intersection points. The ray tracing rendering of the light source is to obtain the second irradiance of the target intersection point projected by the target rendering light on the second rendering image under each illumination sampling; wherein, the probability of selecting each target light source is proportional to the above The light source sampling weight of the light source in the above-mentioned target rendering perspective;

According to each second irradiance of the target intersection under the multiple illumination sampling and the object attribute of the object where the target intersection is located, determine the rendering display information of the target intersection under the illumination of multiple light sources;

The second rendered image is acquired according to the rendering display information of all target intersections in the scene to be rendered under the illumination of multiple light sources.

With reference to the sixth possible implementation manner of the first aspect, in a seventh possible implementation manner, the light source sampling weight of each of the above target light sources in the above target rendering perspective is greater than a preset threshold. By setting the light source sampling weight threshold in the embodiments of the present application, the number of light source samples can be reduced, thereby reducing the amount of computation.

In a second aspect, an embodiment of the present application provides an apparatus for determining a light source sampling weight for multi-light source scene rendering, the apparatus including:

an acquisition module, configured to acquire the scene to be rendered in the multi-light source scene from the target rendering perspective, and acquire the multiple light sources set in the above-mentioned multi-light source scene;

A preprocessing module, configured to perform ray tracing rendering of a single light source on the scene to be rendered acquired by the acquisition module from the perspective of the target rendering, and obtain each first rendered image corresponding to the scene to be rendered under the separate illumination of each light source ;

A determination module, configured to determine the light source sampling weights corresponding to the above-mentioned light sources in the above-mentioned target rendering perspective according to the irradiation information in the above-mentioned first rendered images obtained by the above-mentioned preprocessing module, and the above-mentioned light source sampling weights are used for the above-mentioned to-be-rendered. The scene performs ray-traced rendering of multiple lights.

In a third aspect, an embodiment of the present application provides an electronic device, the electronic device includes: a processor and a memory, the processor and the memory are connected through a bus system;

the memory is used to store instructions;

The processor is configured to invoke the instructions stored in the above-mentioned memory to execute the first aspect or to execute the method steps in any possible implementation manner in combination with the first aspect.

In a fourth aspect, the embodiments of the present application provide a computer program product, including computer-readable instructions, when the computer-readable instructions are executed by one or more processors, the above-mentioned first aspect or any combination of the first aspect is performed. Method steps in one possible implementation.

In a fifth aspect, embodiments of the present application provide a computer storage medium, including computer-readable instructions, when the computer-readable instructions are executed by one or more processors, the first aspect is executed or any one of the first aspects is executed in combination method steps in a possible implementation.

It should be understood that the implementation and beneficial effects of the above-mentioned aspects of the present application may refer to each other.

Description of drawings

FIG. 1 is a structural block diagram of an electronic device provided by an embodiment of the present application;

2 is a schematic flowchart of a method for determining a light source sampling weight for multi-light source scene rendering according to an embodiment of the present application;

3 is a schematic diagram of a ray of bidirectional path tracing rendering;

4-5 are schematic diagrams of human-computer interaction provided by the embodiments of the present application;

6-8 are schematic diagrams of rendering result diagrams provided by embodiments of the present application;

9-14 are schematic diagrams of rendering details of rendering result graphs provided by embodiments of the present application;

FIG. 15 is a schematic structural diagram of an apparatus for determining a light source sampling weight for rendering a multi-light source scene according to an embodiment of the present application.

Detailed ways

The implementation of the technical solutions of the present application will be further described in detail below with reference to the accompanying drawings.

The method for determining the light source sampling weight for multi-light source scene rendering provided in the embodiments of the present application may be executed by an electronic device. The electronic device may be a mobile terminal (for example, a smart phone, a notebook, a tablet computer, etc.), a vehicle-mounted device, an Internet of Things device, or other devices capable of rendering a scene. The electronic device may be a device running an Android system, an IOS system, a windows system, and other systems.

Exemplarily, the specific structure of the electronic device may be shown in FIG. 1 , and the specific structure of the electronic device will be described in detail below with reference to FIG. 1 .

In some possible implementations, the electronic device 10 may include a processor 101 and a memory 102 . The processor 101 may include a central processing unit (Central Processing Unit, CPU) and a graphics processing unit (Graphics Processing Unit, GPU). Further, the electronic device 10 may further include a display 103 . The CPU, GPU, memory 102, display 103, etc. in the electronic device 10 may establish communication connections through a bus system (not shown in the figure).

Optionally, the CPU and the GPU may be located on the same chip, or may be separate chips.

The functions of the processor 101 and the memory 102 are briefly introduced below.

The processor 101 is used to run the operating system 1012 and application programs 1011 . The application 1011 may be a graphics application, such as a game, a video player, 3D modeling, and the like. The operating system 1012 provides a system interface, and the application 1011 generates an instruction stream for rendering the scene through the system interface and system drivers provided by the operating system 1012, such as user mode drivers and/or kernel mode drivers. Further, the processor 101 may also generate rendering data for rendering the scene. The processor 101 generates a rendering target through a rendering pipeline, caches the rendering target, and displays the rendering target on the display 103 . In other words, the processor may be configured to execute the method for determining a light source sampling weight for rendering a multi-light source scene provided in this embodiment of the present application.

Exemplarily, the processor 101 may include, but is not limited to, an application processor, one or more microprocessors, a digital signal processor (Digital Signal Processing, DSP), a microcontroller (microcontroller unit, MCU) or an artificial intelligence processor. Wait.

In some feasible implementations, the operating system 1012 and the application 1011 are run by the CPU, the rendering pipeline runs in the GPU, and the GPU obtains the instruction stream generated by the CPU, generates a rendering target through the rendering pipeline, caches the rendering target, and displays the rendering target. on display 103 .

The display 103 is used to display various rendered images generated by the electronic device 10, and the rendered images may be a graphical user interface (graphical user interface, GUI) of an operating system or image data processed by the processor 101. Alternatively, display 103 may comprise any suitable type of display screen. For example, liquid crystal display (liquid crystal display, LCD) or plasma display or organic light-emitting diode (organic light-emitting diode, OLED) display and so on.

Memory 102 is used to store instructions and rendering data. The memory 102 may be a double data rate synchronous dynamic random access memory (DDR SDRAM) or other types of cache.

It should be noted that the rendering pipeline is a series of operations sequentially performed by the processor in the process of rendering the scene, for example, the ray tracing rendering provided by the embodiments of the present application. The rendering pipeline may run in the CPU and/or the GPU. In other words, the method for determining the light source sampling weight for multi-light source scene rendering provided by the embodiments of the present application may be implemented in the CPU and/or the GPU.

Referring to FIG. 2 , FIG. 2 is a schematic flowchart of a method for determining a light source sampling weight for multi-light source scene rendering according to an embodiment of the present application. As shown in Figure 2, the specific execution steps of the embodiment of the present application are as follows:

S201. The processor acquires a scene to be rendered in a multi-light source scene from a target rendering perspective, and acquires a plurality of light sources set in the multi-light source scene.

The target rendering perspective is determined by the position of the virtual rendering camera in the multi-light scene. In a specific implementation, the processor may obtain the above-mentioned target rendering angle of view according to the position of the virtual rendering camera and the performance parameters of the virtual rendering camera (for example, the focal length, etc.). The target rendering perspective determines the scene to be rendered, and different target rendering perspectives correspond to different scenes to be rendered. It should be noted that the virtual rendering camera may be understood as a device that renders an image and finally presents an image, such as a display. The virtual rendering camera may implement functions including, but not limited to, determining the above-mentioned target rendering angle of view, and may also determine the rendering focal length so as to determine the field of view of the scene to be rendered.

In some feasible implementations, the multi-light source scene can be embodied as a 3D model in a rendering engine.

Multiple light sources are set in a multi-light source scene. Obtaining multiple light sources can be understood as obtaining the light source properties of the light sources, such as light source position, luminous intensity, luminous color, and light source shape. Exemplarily, the position of the light source is identified by (x, y, z) three-dimensional coordinates, and a light source list may be generated from the acquired light source properties of multiple light sources, as shown in Table 1.

Table 1

S202. From the target rendering perspective, the processor performs ray tracing rendering of a single light source for the scene to be rendered, and obtains each first rendered image corresponding to the scene to be rendered under the separate illumination of each light source.

Ray tracing rendering of a single light source means that only one light source is turned on in a multi-light source scene. In other words, the light source properties of only one light source are involved in the calculation of ray tracing rendering.

It should be noted that ray tracing rendering includes path tracing rendering, wherein path tracing rendering includes unidirectional path tracing rendering and bidirectional path tracing rendering. In the embodiment of the present application, both the one-way path tracing rendering and the two-way path tracing rendering can obtain each first rendered image corresponding to the to-be-rendered image under the illumination of each light source alone. Bidirectional path tracing rendering is to emit rendering rays from the virtual rendering camera and each light source, which can quickly generate multiple rays and further accelerate the convergence speed. The convergence speed in this embodiment of the present application includes rendering speed and rendering effect. For example, a fast convergence speed can be understood as rendering the scene more realistic in the same time, or it takes less time to achieve the same rendering effect.

In order to better understand ray tracing rendering, the following takes the specific implementation of ray tracing rendering as bidirectional path tracing rendering as an example, and describes the process of bidirectional path tracing rendering in conjunction with FIG. 3 .

Referring to FIG. 3 , FIG. 3 is a schematic diagram of a ray of bidirectional path tracing rendering. As shown in FIG. 3 , the plurality of light sources include light source 1 , light source 2 , light source 3 , and the like. The scene to be rendered includes a first intersection i and at least one second intersection (for example, a second intersection j, a second intersection k), and the first intersection i is the intersection of the first rendering ray A1 and the first object 1 in the scene to be rendered at the first intersection. An intersection on the surface of an object 1. The second intersection point j is the intersection point where the reflected ray (eg, ray ④ ) of the first intersection point i and the second object 2 in the scene to be rendered intersect on the surface of the second object 2 . It can be understood that in the bidirectional path tracing rendering, the second intersection j may also be the intersection point on the surface of the second object 2 where the light emitted by the light source 1 (that is, the light ②) intersects with the second object 2 in the scene to be rendered, At this time, the second object 2 can be regarded as a third object, that is, the second object and the third object can be the same object. The second intersection point k is the intersection point on the surface of the third object 3 where the light (eg, light ③ ) emitted by the light source 1 intersects with the third object 3 in the scene to be rendered.

The first rendering ray A1 is sent to the scene to be rendered by the virtual rendering camera. It can be understood that the virtual rendering camera can send multiple rendering rays to the scene to be rendered. In FIG. 3, the first rendering ray A1 is sent as an example, and It should not be construed as limiting the application.

For example, it is assumed that the resolution of the first rendered image is 1*1, the number of samples of pixel x (Samples Per Pixel, SPP) is 1, and the first light source (ie, light source 1 ) is in an on state as an example.

In the target rendering perspective, the processor performs illumination sampling on the light source 1 based on the first intersection point i to obtain the first radiance of the first intersection point i projected by the first rendering ray A1 on the first rendered image under the single illumination of the light source 1 .

In the specific implementation, based on the position information of the first intersection i and the light source position of the light source 1, it is determined whether the light ① is visible. For example, the intersection operation can be performed based on the position information of the light source 1 and the position information of the first intersection i. The lines between them do not intersect with other objects in the scene to be rendered, such as the second object 2, the third object 3, etc., which means that light ① is visible. The light source 1 can directly illuminate the first intersection i through the light ①. At this time, the first radiance L ₁ of the first intersection i projected on the first rendered image by the first rendering ray A1 under the single illumination of the light source 1 can be expressed as :

L ₁ =bsdf ₁ .cosα ₁ .L _i1 (x,w _i1 ) Equation 1

Where bsdf ₁ is the bidirectional scattering distribution function, which is determined by the material of the first object 1; α ₁ is the angle between the ray ① and the normal of the first intersection i, and L _i1 represents the incident radiance of the ray ①.

Further, the first intersection i can reflect the illumination of the light source 1. If the first intersection i is reflected to the second object 2 through the light ④, it intersects at the second intersection j at the surface of the second object 2. Lighting sampling is performed on light source 1 based on the second intersection point j, that is, the position information of the second intersection point j and the light source position of light source 1 are intersected. If the line between the two does not intersect with other objects in the scene to be rendered , which means that light ② is visible. The light source 1 can illuminate the first intersection point i through light rays ② and ray ④, and the second intersection point j is projected by the first rendering ray A1 on the first rendering image under the single illumination of the light source 1. The _second radiance L2 can be expressed as :

bsdf ₂ is the bidirectional scattering distribution function, which is determined by the material of the second object 2; α ₂ is the angle between the ray ② and the normal of the second intersection j, β ₁ is the difference between the ray ④ and the normal of the first intersection i The included angle between , L _i2 represents the incident radiance of light ②,

is the _first probability density function, related to β1.

In some possible implementations, based on bidirectional path tracing, light ③ may be emitted from the light source 1 to the third object 3 , and intersect at the second intersection point k at the surface of the third object 3 . Light sampling is performed on the virtual rendering camera based on the second intersection point k. It should be noted that since the light path is reversible, lighting sampling can be understood as sampling the light source from the rendering light emitted by the virtual rendering camera, or sampling the virtual rendering camera from the light emitted by the light source. . The second intersection point k samples the virtual rendering camera through the first rendering ray A1, that is, the position information of the second intersection point k and the position information of the first intersection point i are intersected. The intersection of other objects in the scene means that light ⑤ is visible. At this time, light source 1 can illuminate the first intersection i through light ③ and light ⑤, and the second intersection k is illuminated by light source 1 alone by the first rendering light A1 The second radiance L3 projected _on the first rendered image can be expressed as:

bsdf ₃ is the bidirectional scattering distribution function, which is determined by the material of the third object 3; α ₃ is the angle between the ray ③ and the normal of the second intersection k, and β ₂ is the angle between the ray ⑤ and the normal of the first intersection i The included angle between , L _i3 represents the incident radiance of light ③,

is the _second probability density function, related to β2.

Further, the intersection operation is performed between the position information of the second intersection k and the position information of the second intersection j. If the connection between the two does not intersect with other objects in the scene to be rendered, it means that the light ⑥ is visible. , at this time, the light source 1 can illuminate the first intersection i through light ③, light ⑥ and light ④, and the second intersection k and the second intersection j are projected on the first rendered image by the first rendering ray A1 under the single illumination of the light source 1 The second radiance L4 _on can be expressed as:

β ₃ is the angle between the ray ⑥ and the normal of the second intersection j;

is the _third probability density function, related to β3.

According to the first radiance L ₁ and the respective second radiances such as L ₂ , L ₃ , L ₄ , obtain the first intersection point i projected on the first rendered image by the first rendering ray A1 under the single illumination of the light source 1 The irradiance L is (regardless of object luminescence):

L=L ₁ +L ₂ +L ₃ +L ₄ Equation 5

Based on the first irradiance L and the object attributes (such as color, material, angle, etc.) of the object where the first intersection i is located (ie, the first object 1 ), the rendering display information of the first intersection i under the single illumination of the light source 1 is determined. In a specific implementation, the rendering display information includes color information (ie, RGB representation), and the first irradiance L is multiplied by the RGB representation of the first object 1 to obtain the RGB representation of the first intersection i under the single illumination of the light source 1 .

Since the preset resolution of the first rendered image is 1*1 and the number of samples is 1SPP, the RGB representation of the first intersection i under the single illumination of light source 1 is the RGB representation of the first rendered image. The RGB representation of , that is, the first rendered image can be obtained.

It should be noted that, if the preset number of samples of the first rendered image is greater than 1SPP. In other words, in order to obtain the color information of pixel x, the virtual rendering camera sends out multiple rendering rays to the image to be rendered, and the number of rendering rays is equal to the number of samples. For example, if the sampling number of the first rendered image is preset to 10SPP, the virtual camera sends 10 rendering rays to the image to be rendered, and generates 10 first intersection points with the objects in the scene to be rendered. The calculation of the irradiance of each first intersection point Similar to the irradiance calculation of the first intersection i, the irradiances of all the first intersections in the scene to be rendered are multiplied by the object attributes of their respective objects, respectively, and all the first intersections in the scene to be rendered are obtained. 1 Rendering display information under lighting. Similarly, the rendering display information includes color information (ie, RGB representation). The RGB representations of the multiple first intersection points under the single illumination of the light source 1 are added and the average is taken as the RGB representation of the pixel x.

In some possible implementations, if the preset resolution of the first rendered image is 1*1, the RGB representation of the pixel x is the RGB representation of the first rendered image, and the RGB representation of the first rendered image can be Get the first rendered image.

Optionally, in some feasible implementation manners, if the preset resolution of the first rendered image is greater than 1*1, that is, the first rendered image includes multiple pixels x, the RGB of the multiple pixels x may be represented. Addition is made as the RGB representation of the first rendered image. Optionally, the RGB representations of multiple pixels x may also be added and then averaged, and used as the RGB representation of the first rendered image. The first rendered image can be obtained from the RGB representation of the first rendered image.

It can be understood that the above example illustrates the ray tracing rendering process of a single light source by taking only the light source 1 on as an example. In this embodiment of the present application, each light source in the scene to be rendered can be turned on in sequence, and only one light source is turned on at a time. For example, Only light source 2 is turned on, and all other light sources except light source 2 in the multi-light source scene are turned off. For the specific implementation of the single light source ray tracing rendering of the light source 2, refer to the implementation of only turning on the light source 1, so that each first rendered image corresponding to the scene to be rendered under the independent illumination of each light source can be obtained. At this time, the processor may acquire image information of the first rendered image, such as color information (ie, RGB representation), luminance information (ie, YUV representation), and the like.

S203: The processor determines, according to the irradiation information in each first rendered image, the light source sampling weight corresponding to each light source in the target rendering perspective.

In some feasible implementations, the irradiation information in each first rendered image is determined based on brightness information corresponding to each first rendered image, wherein the brightness information of each first rendered image is determined based on the brightness information in each first rendered image obtained from the color information. Exemplarily, the processor may acquire the RGB representation of each first rendered image, convert the RGB representation of each first rendered image into a YUV representation, and use the brightness information (ie, the Y value) of each first rendered image as each light source in the The corresponding light source sampling weight in the target rendering perspective. The brightness information of each first rendered image is in one-to-one correspondence with each light source, and the brightness information of each first rendered image is used as the light source sampling weight of the corresponding light source. Further, the light source sampling weight of each light source may be normalized. , that is, convert the light source sampling weight of each light source to the range of 0 to 1. Exemplarily, the light source sampling weight of each light source may be associated with the light source attribute in Table 1, and a new light source list is obtained as shown in Table 2.

Form 2

In this embodiment of the present application, the light source sampling weight corresponding to each light source is determined according to the irradiance information of each first rendered image obtained after each light source performs single light source ray tracing rendering on the scene to be rendered, and the light source sampling weight can make the multi-light source ray tracing rendering process The probability of selecting a light source corresponding to a larger light source sampling weight increases. By implementing the embodiments of the present application, a more suitable light source can be selected during the multi-light source ray tracing rendering process, thereby accelerating the convergence speed of multi-light source scene rendering.

Further, based on step S203 to obtain the light source sampling weights corresponding to each light source in the target rendering perspective, the processor may perform ray tracing rendering of multiple light sources in the scene to be rendered in the target rendering perspective, and obtain the first ray tracing rendering of the scene to be rendered under the illumination of multiple light sources. 2. Render the image.

In a specific implementation, multiple light sources in the multi-light source scene are in an on state. The virtual rendering camera emits a target rendering ray to the scene to be rendered, and the intersection of the target rendering ray and any object in the to-be-rendered scene on the surface of any object is the target intersection point. From the target rendering perspective, multiple lighting samples are performed on each light source based on the target intersection, and the irradiance of the target intersection under each lighting sampling is obtained. It can be understood that, unlike the single light source illumination sampling performed on the first light source based on the first intersection point, more than one light source is turned on at this time, and each light source is subjected to multiple illumination sampling based on the target intersection point. Each target light source is selected from each light source, and each target light source selected for each lighting sampling is not necessarily the same combination. Select light source 2 and light source 3 or select light source 1, light source 2 and light source 3, etc. Since the light source sampling weight of light source 2 is the largest, the probability of selecting light source 2 is greater than the probability of selecting light source 1 and light source 3. In other words, the probability of selecting each target light source is proportional to the light sampling weight of each light source in the target rendering perspective. When the processor obtains the second irradiance of the target intersection point projected by the target rendering ray on the second rendered image under each illumination sampling, reference may be made to the acquisition method of the first irradiance.

Optionally, in the embodiment of the present application, the ray tracing rendering of multiple light sources may be unidirectional path tracing rendering, or may be bidirectional path tracing rendering.

According to each second irradiance of the target intersection under multiple illumination sampling and the object attributes of the object where the target intersection is located, determine the rendering display information of the target intersection under multi-light source illumination, where the rendering display information may include color information (ie, RGB express). Exemplarily, the irradiance of the target intersection point under each illumination sampling is added and averaged, as the irradiance of the target intersection point under the illumination of multiple light sources. Multiply the irradiance of the target intersection point under multi-light source illumination by the RGB representation of the object where the target intersection point is located, so as to obtain the RGB representation of the target intersection point under multi-light source illumination.

The rendering display information (such as RGB representation) of all target intersections in the scene to be rendered under the illumination of multiple light sources is added and averaged to obtain the RGB representation of the second rendered image. The RGB representation of the second rendered image can be Get the second rendered image.

In the embodiment of the present application, based on the light source sampling weight of each light source, the probability of selecting a light source corresponding to a larger light source sampling weight increases during the ray tracing rendering process, which can speed up the convergence speed of multi-light source scene rendering.

Further, the light source sampling weight of each target light source in the target rendering perspective is greater than the preset threshold. In other words, the processor culls each light source by setting a preset threshold. Exemplarily, the preset threshold is 0.3, and the light source whose light source sampling weight is not greater than 0.3 (for example, light source 1 in Table 3) is not rendered by ray tracing. Within the light source selection range, that is, light source 1 does not participate in the calculation of ray tracing rendering. By setting the light source sampling weight threshold in the embodiments of the present application, the number of light source samples can be reduced, thereby reducing the computational complexity of the processor.

In some feasible implementation manners, the resolution of each of the above-mentioned first rendered images is smaller than the resolution of the second rendered images. In this embodiment of the present application, the first rendered image is set to be a smaller resolution image, and the light source sampling weight of each light source is calculated before the official rendering, which can reduce the calculation amount of the processor and improve the rendering speed.

Some human-computer interaction embodiments of the present application are described below with reference to FIGS. 4 to 14 .

In some embodiments of the present application, the electronic device may be set with a rendering preprocessing mode, and the user may enable the rendering preprocessing mode as required. Turning on the rendering preprocessing mode is to implement the light source sampling weight method for multi-light source scene rendering provided by the embodiment of the present application. After the electronic device enables the rendering preprocessing mode, the convergence speed of the multi-light source scene rendering can be accelerated.

Exemplarily, FIG. 4 shows a possible human-computer interaction diagram in which the user turns on the rendering preprocessing mode. As shown in FIG. 4 , the screen of the electronic device can display a rendering interface of 3D animation production software (such as Blender software), the rendering interface includes a multi-light source scene 40 and a setting interface 41 , and the multi-light source scene 40 is a residential building as an example , there are light sources on each floor of the residential building. It can be understood that the walls, floors, light sources, etc. of the residential building are set during the 3D modeling process, and the embodiment of the present application renders the model obtained by the 3D modeling (ie, the multi-light source scene 40 ) based on the Blender software. . The electronic device may acquire the placement position of the virtual rendering camera in response to the user's click operation, drag operation, etc. on the virtual rendering camera in the setting interface 41 . Based on the placement position of the virtual rendering camera and performance parameters (eg, focal length, etc.) of the virtual rendering camera, the target rendering angle of view is determined. The processor obtains the scene to be rendered in the multi-light source scene 40 from the target rendering perspective. For example, the placement position of the virtual rendering camera is a room in a residential building, and the screen of the electronic device can display the scene to be rendered 50 as shown in FIG. 5 . In some feasible implementation manners, the setting interface 41 may further include other information, such as shape information, attributes, rendering time, etc. of each object in the multi-light source scene.

The electronic device may also respond to the resolution setting input by the user in the setting interface 41, where the resolution set here is the resolution of the second image obtained after the scene to be rendered is rendered by ray tracing under the illumination of multiple light sources.

The setting interface 41 further includes an option of whether to enable the rendering preprocessing mode. Exemplarily, the electronic device may also enable the rendering preprocessing mode in response to the user checking the selection of light source sampling weight preprocessing.

After the rendering preprocessing mode is turned on, the electronic device will first perform ray tracing rendering of a single light source on the scene 50 to be rendered. Exemplarily, the electronic device turns on the single light source in turn, and respectively performs bidirectional path tracing rendering on the scene to be rendered 50 to obtain the scene to be rendered. 50 Determine the light source sampling weight of each light source based on the irradiation information in each first rendered image (eg, the Y value of each first rendered image) for each first rendered image corresponding to each light source individually illuminated. After the electronic device obtains the light source sampling weight of each light source, it turns on all the light sources in the multi-light source scene 40, and performs ray tracing rendering of the multi-light source in the scene to be rendered 50, for example, bidirectional path tracing rendering, and obtains the scene to be rendered under the illumination of multiple light sources. of the second rendered image. Exemplarily, the electronic device may respond to the rendering time set by the user, for example, the rendering time is 60s. Or, based on the default rendering time, the electronic device performs ray tracing rendering of multiple light sources on the scene 50 to be rendered, and obtains the rendering result diagram (ie, the second rendered image) as shown in FIG. 6 .

In order to illustrate that the method for determining the light source sampling weight for multi-light source scene rendering provided by the embodiment of the present application can speed up the convergence speed of multi-light source scene rendering, the inventor of the present application has conducted a comparative experiment, turning off the rendering preprocessing mode (that is, not checking light source sampling weight options). In the same case as the rendering time when the rendering preprocessing mode is turned on, for example, the rendering time is also 60s. All light sources in the multi-light source scene 40 are turned on, and the electronic device performs ray tracing rendering of the multi-light source in the scene 50 to be rendered based on the default light source sampling weight, and the rendering result diagram as shown in FIG. 7 is obtained. Further, the rendering preprocessing mode can be turned off, but when the rendering time is long enough, for example, the rendering time is 2h, the electronic device turns on all the light sources in the multi-light source scene 40 based on the default light source sampling weight, and executes multiple functions in the scene 50 to be rendered. The ray tracing rendering of the light source, the rendering result shown in Figure 8 is obtained. At this time, it can be considered that the rendering result graph shown in FIG. 8 is an effect close to the real image, that is, FIG. 8 can be used as a reference rendering result graph of the scene to be rendered.

It can be seen from the comparison of the rendering result graphs shown in FIG. 6 and FIG. 7 that, by implementing the method for determining the light source sampling weight for multi-light source scene rendering provided by the embodiment of the present application, within the same rendering time, the second rendered image obtained by the embodiment of the present application is obtained. More realistic and less noisy. And it can be seen from the comparison diagrams of rendering results shown in FIG. 6 and FIG. 8 that implementing the method for determining the light source sampling weight for multi-light source scene rendering provided by the embodiment of the present application can achieve the same rendering as the reference rendering result diagram in a relatively short period of time. Effect. That is, by implementing the embodiments of the present application, the convergence speed of multi-light source scene rendering can be accelerated.

Further, in order to more clearly and intuitively present the rendering effect achieved by implementing the present application, the present application also provides a schematic diagram of rendering details of the rendering result graph. See Figures 9 to 14.

It should be noted first that Figures 9 and 12 are schematic diagrams of rendering details of the rendering result diagram shown in Figure 6; Figures 10 and 13 are schematic diagrams of rendering details of the rendering result diagram shown in Figure 7; Figures 11 and 14 It is a schematic diagram of rendering details of the rendering result graph shown in FIG. 8 . Figure 9, Figure 10, and Figure 11 constitute a set of comparison combinations, and Figures 12, 13, and 14 constitute another set of comparison combinations. It can be understood that a set of comparison combinations indicates that all the rendering results shown in the combination are Rendered for the same detail (ie the same object). According to the comparison of the rendering result graphs shown in FIG. 9 with FIG. 10 and FIG. 12 with FIG. 13 , it can be more clearly seen that within the same rendering time, the second rendered image obtained by implementing the embodiment of the present application is more realistic.

According to FIG. 9 and FIG. 11 , FIG. 12 and FIG. 14 , it can also be more clearly seen that the rendering result obtained by implementing the present application is not much different from the reference rendering result graph. to achieve the same rendering effect as the reference rendering result image within the time. That is, by implementing the embodiments of the present application, the convergence speed of multi-light source scene rendering can be accelerated.

Exemplarily, in addition to using the fidelity of the above rendering result graph to measure the convergence speed of multi-light source scene rendering, quantitative data may also be used to measure the convergence speed of multi-light source scene rendering. The quantized data may include but not limited to peak signal-to-noise ratio (peak signal-to-noise ratio, PSNR), structural similarity (structural similarity, SSIM), root-mean-square error (root-mean-square error, RMSE) and the like. When other rendering conditions are the same (for example, the rendering time is the same, the rendering angle of view is the same, etc.), the comparison of the quantitative data between the embodiment of the present application (that is, the rendering preprocessing mode is turned on) and the rendering preprocessing mode not turned on can be shown in Table 3. .

Form 3

It should be noted that the larger the value of PSNR is, the closer the rendered image is to the reference rendering result image; the larger the value of SSIM, the closer the rendered image is to the reference rendering result image; the smaller the value of RMSE, the closer the rendered image is to the reference rendering image. The resulting graph error is smaller. The reference rendering images are shown in Figure 8, Figure 11, and Figure 14, respectively, which can be obtained from the quantified data in Table 3. The rendered image obtained by enabling the rendering preprocessing mode is closer to the reference rendering than the rendered image obtained by not enabling the rendering preprocessing mode. image. That is, by implementing the embodiments of the present application, the convergence speed of multi-light source scene rendering can be accelerated.

Referring to FIG. 15 , FIG. 15 is a schematic structural diagram of an apparatus for determining a light source sampling weight for rendering a multi-light source scene according to an embodiment of the present application. As shown in FIG. 15 , the device 150 for determining light source sampling weights for multi-light source scene rendering includes:

an acquisition module 1500, configured to acquire the scene to be rendered in the multi-light source scene from the target rendering perspective, and acquire a plurality of light sources set in the multi-light source scene;

The preprocessing module 1501 is configured to perform ray tracing rendering of a single light source on the scene to be rendered acquired by the acquisition module 1500 under the target rendering perspective, and obtain each first rendered image corresponding to the scene to be rendered under the separate illumination of each light source;

The determining module 1502 is configured to determine the light source sampling weight corresponding to each light source in the target rendering perspective according to the irradiance information in each first rendered image obtained in the preprocessing module 1501, and the light source sampling weight is used to perform multi-processing on the scene to be rendered. Raytraced rendering of light sources.

Further, the device 150 for determining the light source sampling weight for multi-light source scene rendering further includes a processing module 1503;

The processing module 1503 is configured to perform ray tracing rendering of multiple light sources in the scene to be rendered based on the light source sampling weights corresponding to each light source in the target rendering perspective determined by the determining module 1502 under the target rendering perspective;

The obtaining module 1500 is further configured to obtain a second rendered image of the scene to be rendered under the illumination of multiple light sources.

In some feasible implementation manners, the determining module 1502 is further configured to determine irradiation information in each first rendered image based on brightness information corresponding to each first rendered image, wherein the brightness information of each first rendered image is based on each first rendered image. The color information in a rendered image is converted.

In some feasible implementation manners, the plurality of light sources include a first light source; the scene to be rendered includes a first intersection and at least one second intersection, where the first intersection is the first rendering ray and the first object in the scene to be rendered The intersection point on the surface of the first object, the second intersection point is the intersection point where the reflected ray of the first intersection point and the second object in the scene to be rendered intersect on the surface of the second object and/or the above-mentioned light sources emit The ray of light intersects each third object in the scene to be rendered at the intersection on the surface of each third object; the first rendering ray is sent by the virtual rendering camera to the above scene to be rendered, and the virtual rendering camera is used to determine the above target rendering perspective;

The acquiring module 1500 is further configured to perform illumination sampling on the first light source based on the first intersection point under the target rendering perspective, to obtain the first intersection point projected on the above-mentioned first rendered image by the first rendering light under the single illumination of the first light source. first radiance;

The preprocessing module 1501 is further configured to perform illumination sampling on the first light source based on each second intersection point under the target rendering perspective, and obtain each second intersection point under the single illumination of the first light source by the first rendering light projected on the above-mentioned first rendering each second radiance on the image;

The obtaining module 1500 is further configured to obtain, according to the sum of the first radiance and each second radiance, the first irradiance of the first intersection point projected on the above-mentioned first rendered image by the first rendering light under the single illumination of the first light source Spend;

The determining module 1502 is further configured to determine the rendering display information of the first intersection under the single illumination of the first light source based on the above-mentioned first irradiance and the object attribute of the object where the first intersection is located;

The obtaining module 1500 is further configured to obtain a first rendered image corresponding to the scene to be rendered under the single illumination of the first light source based on the rendering display information of all the first intersection points in the scene to be rendered under the illumination of the first light source.

In some feasible implementations, the scene to be rendered includes a target intersection point, and the target intersection point is the intersection point between the target rendering ray and any object in the scene to be rendered on the surface of any object; the target rendering ray is virtual rendering Sent by the camera to the above scene to be rendered, the virtual rendering camera is used to determine the above target rendering perspective;

The processing module 1503 is further configured to perform multiple illumination sampling on each light source based on the target intersection point under the target rendering perspective; wherein each illumination sampling corresponds to selecting each target light source from multiple light sources, and each target light source is used to perform execution on the target intersection point Ray tracing rendering of multiple light sources, wherein the probability of selecting each target light source is proportional to the light source sampling weight of each light source in the target rendering perspective;

The obtaining module 1500 is further configured to obtain the second irradiance of the target intersection point projected by the target rendering light on the above-mentioned second rendered image under each illumination sampling;

The determining module 1502 is further configured to determine the rendering display information of the target intersection under multi-light source illumination according to each second irradiance of the target intersection under multiple illumination sampling and the object attributes of the object where the target intersection is located;

The obtaining module 1500 is further configured to obtain the above-mentioned second rendered image according to the rendering display information of all target intersections in the scene to be rendered under the illumination of multiple light sources.

It should be noted that the above-mentioned terms "first" and "second" are only used for descriptive purposes, and should not be construed as indicating or implying relative importance.

In the several embodiments provided in this application, it should be understood that the disclosed method, apparatus and system may be implemented in other manners. The above-described embodiments are only illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components may be combined, or Integration into another system, or some features can be ignored, or not implemented. In addition, the coupling, or direct coupling, or communication connection between the components shown or discussed may be through some interfaces, and the indirect coupling or communication connection of devices or units may be electrical, mechanical or other forms. of.

Those of ordinary skill in the art can understand that all or part of the steps of implementing the above method embodiments can be completed by program instructions related to hardware, the aforementioned program can be stored in a computer-readable storage medium, and when the program is executed, execute Including the steps of the above method embodiment; and the aforementioned storage medium includes: a mobile storage device, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk and other various A medium on which program code can be stored.

Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software function module and sold or used as an independent product, it may also be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the embodiments of the present invention may be embodied in the form of software products in essence or the parts that make contributions to the prior art. The computer software products are stored in a storage medium and include several instructions for A computer device (which may be a personal computer, a server, or a network device, etc.) is caused to execute all or part of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes: a removable storage device, a ROM, a RAM, a magnetic disk or an optical disk and other mediums that can store program codes.

Claims (12)

1. A method for determining light source sampling weights for multi-light source scene rendering, characterized in that the method includes:

   acquiring the scene to be rendered in the multi-light source scene from the target rendering perspective, and acquiring the multiple light sources set in the multi-light source scene;

   From the target rendering perspective, perform ray tracing rendering of a single light source on the scene to be rendered, to obtain each first rendered image corresponding to the scene to be rendered under the separate illumination of each light source;

   According to the irradiance information in each first rendered image, the light source sampling weight corresponding to each light source in the target rendering perspective is determined, and the light source sampling weight is used to perform multi-light source lighting on the scene to be rendered Trace rendering.
2. The method according to claim 1, wherein the method further comprises:

   In the target rendering perspective, based on the light source sampling weights corresponding to the respective light sources in the target rendering perspective, ray tracing rendering of multiple light sources is performed on the scene to be rendered, and the multi-light source illumination of the scene to be rendered is obtained. The second rendered image below.
3. The method according to claim 2, wherein the resolution of each first rendered image is smaller than the resolution of the second rendered image.
4. The method according to any one of claims 1-3, wherein the ray tracing rendering comprises bidirectional path tracing rendering.
5. The method according to claim 1, wherein the irradiation information in each of the first rendered images is determined based on luminance information corresponding to the respective first rendered images, wherein the respective first rendered images The brightness information of is obtained by conversion based on the color information in each of the first rendered images.
6. The method according to any one of claims 1-5, wherein the plurality of light sources comprises a first light source; the scene to be rendered comprises a first intersection and at least one second intersection, and the first intersection is The intersection of the first rendering ray and the first object in the scene to be rendered on the surface of the first object, and the second intersection is the reflected ray of the first intersection and the scene to be rendered. The intersection of the second object on the surface of the second object and/or the intersection of the light emitted by the light source and the third object in the scene to be rendered on the surface of the third object respectively; The first rendering light is sent by a virtual rendering camera to the scene to be rendered, and the virtual rendering camera is used to determine the target rendering angle of view;

   The performing ray tracing rendering of a single light source on the scene to be rendered from the target rendering perspective, and obtaining each first rendered image corresponding to the scene to be rendered under the lighting of each light source alone includes:

   Under the target rendering perspective, lighting sampling is performed on the first light source based on the first intersection point, so as to obtain that the first intersection point is projected on the a first radiance on the first rendered image;

   Under the target rendering perspective, lighting sampling is performed on the first light source based on each second intersection point, so as to obtain that each second intersection point is projected by the first rendering light on the respective second radiances on the first rendered image;

   According to the sum of the first radiance and each of the second radiances, obtain the projection of the first intersection point on the first rendered image by the first rendering light under the single illumination of the first light source first irradiance;

   determining, based on the first irradiance and an object attribute of the object where the first intersection is located, rendering display information of the first intersection under the single illumination of the first light source;

   Based on the rendering display information of all the first intersection points in the scene to be rendered under the illumination of the first light source, a first rendered image corresponding to the scene to be rendered under the single illumination of the first light source is acquired.
7. The method according to claim 2, wherein the scene to be rendered includes a target intersection point, and the target intersection point is that the target rendering ray intersects with any object in the scene to be rendered on the surface of any object. The target rendering ray is sent to the scene to be rendered by a virtual rendering camera, and the virtual rendering camera is used to determine the target rendering angle of view;

   In the target rendering perspective, based on the light source sampling weights corresponding to the respective light sources in the target rendering perspective, ray tracing rendering of multiple light sources is performed on the scene to be rendered, and the number of the scene to be rendered is obtained. The second rendered image under the light source includes:

   Under the target rendering perspective, multiple illumination samplings are performed on each light source based on the target intersection point, wherein each illumination sampling corresponds to selecting each target light source from the plurality of light sources, and each target light source is used for the Performing ray tracing rendering of multiple light sources at the target intersection to obtain the second irradiance projected by the target rendering rays on the second rendered image at the target intersection under each illumination sampling; wherein, selecting the respective The probability of the target light source is proportional to the light source sampling weight of each light source in the target rendering perspective;

   According to each second irradiance of the target intersection under the multiple illumination sampling and the object attribute of the object where the target intersection is located, determine the rendering display information of the target intersection under the illumination of multiple light sources;

   The second rendered image is acquired according to rendering display information of all target intersections in the scene to be rendered under the illumination of multiple light sources.
8. The method according to claim 7, wherein a light source sampling weight of each target light source in the target rendering perspective is greater than a preset threshold.
9. A device for determining light source sampling weights for multi-light source scene rendering, characterized in that the device comprises:

   an acquisition module, configured to acquire a scene to be rendered in a multi-light source scene from a target rendering perspective, and acquire a plurality of light sources set in the multi-light source scene;

   A preprocessing module, configured to perform ray tracing rendering of a single light source on the scene to be rendered acquired by the acquisition module from the perspective of the target rendering, so as to obtain each corresponding scene of the scene to be rendered under the separate illumination of each light source. the first rendered image;

   A determination module, configured to determine, according to the irradiation information in the respective first rendered images obtained by the preprocessing module, the light source sampling weights corresponding to the respective light sources in the target rendering perspective, and the light source sampling weights are determined by using performing ray tracing rendering of multiple light sources on the scene to be rendered.
10. An electronic device, characterized in that the electronic device comprises: a processor and a memory, and the processor and the memory are connected through a bus system;

    the memory is used to store instructions;

    The processor is configured to invoke the instructions stored in the memory to perform the method of any one of the preceding claims 1-8.
11. A computer program product comprising computer readable instructions which, when executed by one or more processors, perform the method of any of the preceding claims 1-8.
12. A computer storage medium comprising computer readable instructions which, when executed by one or more processors, perform the method of any one of the preceding claims 1-8.

Images