Traffic scene dynamic object separation method and device based on gaussian spanning algorithm

Abstract

The invention relates to an image processing technology, and discloses a traffic scene dynamic object separation method based on a gaussian spanning algorithm, which comprises the following steps: acquiring a dynamic object point cloud and a static background point cloud in a traffic scene; constructing a dynamic object model according to the dynamic object point cloud, and constructing a static background model according to the static background point cloud; and combining the dynamic object model and the static background model, and rendering a combined scene image by using a preset scene rendering algorithm based on Gaussian patterning. The invention further provides a traffic scene dynamic object separation device based on the gaussian spanning algorithm, electronic equipment and a storage medium. According to the method, the accuracy of traffic scene dynamic object separation based on the gaussian spanning algorithm can be improved, and the image reconstruction effect of the dynamic objects in the traffic scene can be improved.

Description

Technical Field

[0001] The invention belongs to the technical field of image processing, and in particular relates to a method and a device for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm. Background Art

[0002] With the widespread application and advancement of computer vision technology, reconstructing new views is a long-standing challenge in the field of computer vision and a basic requirement for many applications. In complex traffic scenes, scene image reconstruction of moving objects can effectively separate moving objects and complete the reconstruction of scene backgrounds in complex traffic scenes.

[0003] Traditional technology usually uses the DrivingGaussian framework to perform three-dimensional reconstruction of scene images of moving objects in complex traffic scenes. This method first uses incremental 3D Gaussians to gradually model the complex static background of the entire scene, and then constructs a combined dynamic Gaussian map to characterize each dynamic foreground object and restore their accurate position and occlusion relationship in the scene. However, this method may cause visual artifacts and blur when representing fast-moving dynamic objects, resulting in poor image reconstruction of dynamic objects in traffic scenes. Summary of the invention

[0004] The present invention provides a method and device for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm, which can improve the accuracy of separating dynamic objects in traffic scenes and improve the image reconstruction effect of dynamic objects in traffic scenes.

[0005] To achieve the above object, the present invention provides a method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm, comprising:

[0006] Obtain point clouds of dynamic objects and static background in traffic scenes;

[0007] constructing a dynamic object model according to the dynamic object point cloud, and constructing a static background model according to the static background point cloud;

[0008] The dynamic object model and the static background model are combined, and a combined scene image is rendered using a preset scene rendering algorithm based on Gaussian splatting.

[0009] Optionally, before obtaining the point cloud of dynamic objects and the point cloud of static background in the traffic scene, it also includes: obtaining camera internal and external parameters of the static background, wherein the camera external parameters of the static background convert the point cloud of the static background in the world coordinate system into a point cloud in the camera coordinate system, and the camera internal parameters of the static background are projected and transformed to obtain the background part in the static background picture.

[0010] Optionally, before obtaining the point cloud of dynamic objects and the static background point cloud in the traffic scene, it also includes: obtaining the camera internal and external parameters of the dynamic object, wherein the camera external parameters of the dynamic object convert the point cloud of the dynamic object in the dynamic object's own coordinate system into a point cloud in the camera coordinate system, and the camera internal parameters of the dynamic object are projected and transformed to obtain the object image.

[0011] Optionally, combining the dynamic object model and the static background model includes:

[0012] The dynamic object model, the static background model, the camera extrinsic parameters of the dynamic object and the camera extrinsic parameters of the static background are combined to obtain a combined model.

[0013] Optionally, the combined model includes:

[0014] In the combined model, for the point cloud of the dynamic object itself, the dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the external parameters of the dynamic object, and then the point cloud is transformed to the world coordinate system through the external parameters of the static background. The transformation process is as follows:

[0015] The dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the dynamic object's external parameters:

[0016]

[0017] Among them, P c is the coordinate of the point cloud of the dynamic object in the camera coordinate system, P o is the coordinate of the point cloud of the dynamic object in the dynamic object's own coordinate system, It is the transformation matrix from the dynamic object's own coordinate system to the camera coordinate system;

[0018] Then transform the point cloud to the world coordinate system through the external parameters of the static background:

[0019]

[0020] Among them, P w is the coordinate of the point cloud of the dynamic object in the world coordinate system, is the transformation matrix from the camera coordinate system to the world coordinate system;

[0021] The merge transformation process is as follows:

[0022]

[0023] Optionally, rendering the combined scene image using a preset Gaussian splatting-based scene rendering algorithm includes:

[0024] Randomly select a camera view of a static background, and query the dynamic object model corresponding to the camera view in the complete picture of the dynamic object and the camera internal and external parameters of the dynamic object corresponding to the current frame;

[0025] The dynamic object model corresponding to the camera perspective in the complete picture of the dynamic object and the internal and external parameters of the camera of the dynamic object corresponding to the current frame are input into the fusion renderer, and the initial rendering image is rendered using the scene rendering algorithm preset in the fusion renderer;

[0026] Calculating the loss value of the initial rendered image and the complete image of the dynamic object, optimizing the point cloud position and Gaussian sphere attributes of the static background and the point cloud position and Gaussian sphere attributes of the dynamic object according to the loss value, and iteratively optimizing the fusion renderer to obtain the target fusion renderer;

[0027] The combined model corresponding to the current frame is rendered using a target fusion renderer to obtain a combined scene image.

[0028] Optionally, rendering the combined model corresponding to the current frame using a target fusion renderer to obtain a combined scene image includes:

[0029] All dynamic objects in the dynamic object list involving coordinate transformation attributes are converted to the world coordinate system through the camera internal and external parameters of the static background participating in the dynamic objects, and the coordinate attributes of the dynamic objects are obtained;

[0030] Combining the rendering related attributes of all dynamic objects with the rendering related attributes of the static background to obtain a combined rendering attribute;

[0031] The dynamic object coordinate attributes and the combined rendering attributes are summarized to obtain the combined rendering parameters, the combined model and the combined rendering parameters are input into the target fusion renderer, and the differentiable Gaussian rasterization module is used to render the combined scene image based on the Gaussian splatting scene rendering algorithm.

[0032] In order to solve the above problems, the present invention also provides a traffic scene dynamic object separation device based on the Gaussian splatting algorithm, the device comprising:

[0033] Point cloud acquisition module, used to acquire point clouds of dynamic objects and static background in traffic scenes;

[0034] A model building module, used to build a dynamic object model according to the dynamic object point cloud, and to build a static background model according to the static background point cloud;

[0035] The scene rendering module is used to combine the dynamic object model and the static background model, and render a combined scene image using a preset scene rendering algorithm based on Gaussian splatting.

[0036] In order to solve the above problem, the present invention further provides an electronic device, the electronic device comprising:

[0037] at least one processor; and,

[0038] a memory communicatively connected to the at least one processor; wherein,

[0039] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the above-mentioned method for separating dynamic objects in traffic scenes based on the Gaussiansplatting algorithm.

[0040] In order to solve the above problems, the present invention also provides a computer-readable storage medium, in which at least one computer program is stored. The at least one computer program is executed by a processor in an electronic device to implement the above-mentioned method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm.

[0041] The present invention constructs a dynamic object model based on the dynamic object point cloud and a static background model based on the static background point cloud, so as to separate the dynamic objects and the static background in the traffic scene, and avoid the influence of the motion state of the dynamic object on the extraction of the dynamic object in the traffic scene. In addition, by combining the dynamic object model and the static background model, the dynamic object and the static background in the traffic scene can be combined, so that the combined scene image can be rendered subsequently using the preset Gaussian splatting-based scene rendering algorithm, and the visual artifacts and blur in the rendered combined scene image can be effectively reduced, so as to improve the accuracy of separation of dynamic objects in the traffic scene and enhance the image reconstruction effect of dynamic objects in the traffic scene. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 A schematic diagram of a flow chart of a method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm provided in one embodiment of the present invention;

[0043] Figure 2An algorithm structure flow chart of a method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm provided in one embodiment of the present invention;

[0044] Figure 3 A complete picture schematic diagram of a method for separating dynamic objects in a traffic scene based on a Gaussian splatting algorithm provided by an embodiment of the present invention;

[0045] Figure 4 A schematic diagram of a split static background image of a method for separating dynamic objects in a traffic scene based on a Gaussian splatting algorithm provided in an embodiment of the present invention;

[0046] Figure 5 A schematic diagram of a split dynamic object picture of a traffic scene dynamic object separation method based on a Gaussian splatting algorithm provided in an embodiment of the present invention;

[0047] Figure 6 A schematic diagram of a combined scene image obtained by rendering using a target fusion renderer in a traffic scene dynamic object separation method based on a Gaussian splatting algorithm provided in an embodiment of the present invention;

[0048] Figure 7 A functional module diagram of a traffic scene dynamic object separation device based on a Gaussian splatting algorithm provided by an embodiment of the present invention;

[0049] Figure 8 A schematic diagram of the structure of an electronic device for implementing a method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm provided in one embodiment of the present invention.

[0050] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0051] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0052] The embodiment of the present application provides a method for separating dynamic objects in traffic scenes based on the gaussian splatting algorithm. The execution subject of the method for separating dynamic objects in traffic scenes based on the gaussian splatting algorithm includes but is not limited to at least one of the electronic devices that can be configured to execute the method provided in the embodiment of the present application, such as a server, a terminal, etc. In other words, the method for separating dynamic objects in traffic scenes based on the gaussian splatting algorithm can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms.

[0053] Reference Figure 1 FIG. 1 is a flow chart of a method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm according to an embodiment of the present invention. In this embodiment, the method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm includes:

[0054] S1. Obtain the point cloud of dynamic objects and static background in the traffic scene.

[0055] In the embodiment of the present invention, a dynamic object refers to an object that can move in a traffic scene, such as a moving vehicle and a walking pedestrian.

[0056] In the embodiment of the present invention, the static background refers to a background formed by objects that cannot move in a traffic scene.

[0057] The embodiment of the present invention obtains the dynamic object point cloud and the static background point cloud in the traffic scene through the laser radar equipment.

[0058] As an embodiment of the present invention, before obtaining the point cloud of dynamic objects and the point cloud of static background in the traffic scene, it also includes: obtaining the camera internal and external parameters of the static background, wherein the camera external parameters of the static background convert the point cloud of the static background in the world coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the static background are projected and transformed to obtain the background part in the static background picture.

[0059] As an embodiment of the present invention, before obtaining the point cloud of dynamic objects and the static background point cloud in the traffic scene, it also includes: obtaining the camera internal and external parameters of the dynamic object, wherein the camera external parameters of the dynamic object convert the point cloud of the dynamic object in the dynamic object's own coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the dynamic object are projected and transformed to obtain the object image.

[0060] In an embodiment of the present invention, camera intrinsic and extrinsic parameters refer to key parameters for projecting points in a three-dimensional world onto a two-dimensional image plane, wherein the camera intrinsic and extrinsic parameters include camera intrinsic parameters and camera extrinsic parameters. Camera intrinsic parameters refer to the inherent properties of the camera itself, describing how the camera maps points in three-dimensional space to a two-dimensional image plane; camera extrinsic parameters refer to the position and direction of the camera in the world coordinate system, and the camera extrinsic parameters convert points in the world coordinate system to the camera coordinate system.

[0061] S2. Constructing a dynamic object model according to the dynamic object point cloud, and constructing a static background model according to the static background point cloud.

[0062] S3, combining the dynamic object model and the static background model, and rendering a combined scene image using a preset Gaussian splatting-based scene rendering algorithm.

[0063] As an embodiment of the present invention, the dynamic object model and the static background model are combined, including:

[0064] The dynamic object model, the static background model, the camera extrinsic parameters of the dynamic object and the camera extrinsic parameters of the static background are combined to obtain a combined model.

[0065] In the embodiment of the present invention, the combined model is equivalent to a model in which the dynamic object point cloud and the static background are in the same scene and fixed in the current frame.

[0066] Furthermore, the combined model includes:

[0067] In the combined model, for the point cloud of the dynamic object itself, the dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the external parameters of the dynamic object, and then the point cloud is transformed to the world coordinate system through the external parameters of the static background. The transformation process is as follows:

[0068] The dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the dynamic object's external parameters:

[0069]

[0070] Among them, P c is the coordinate of the point cloud of the dynamic object in the camera coordinate system, P ois the coordinate of the point cloud of the dynamic object in the dynamic object's own coordinate system, It is the transformation matrix from the dynamic object's own coordinate system to the camera coordinate system;

[0071] Then transform the point cloud to the world coordinate system through the external parameters of the static background:

[0072]

[0073] Among them, P ω is the coordinate of the point cloud of the dynamic object in the world coordinate system, is the transformation matrix from the camera coordinate system to the world coordinate system;

[0074] The merge transformation process is as follows:

[0075]

[0076] As an embodiment of the present invention, a combined scene image is rendered using a preset scene rendering algorithm based on Gaussian splatting, including:

[0077] Randomly select a camera view of a static background, and query the dynamic object model corresponding to the camera view in the complete picture of the dynamic object and the camera internal and external parameters of the dynamic object corresponding to the current frame;

[0078] The dynamic object model corresponding to the camera perspective in the complete picture of the dynamic object and the internal and external parameters of the camera of the dynamic object corresponding to the current frame are input into the fusion renderer, and the initial rendering image is rendered using the scene rendering algorithm preset in the fusion renderer;

[0079] Calculating the loss value of the initial rendered image and the complete image of the dynamic object, optimizing the point cloud position and Gaussian sphere attributes of the static background and the point cloud position and Gaussian sphere attributes of the dynamic object according to the loss value, and iteratively optimizing the fusion renderer to obtain the target fusion renderer;

[0080] The combined model corresponding to the current frame is rendered using a target fusion renderer to obtain a combined scene image.

[0081] In the embodiment of the present invention, the fusion renderer mainly defines a render function, whose function is to input a static background model of a certain frame and the internal and external parameters of the static background camera, a dynamic object list, the internal and external parameters of the dynamic object camera, and other rendering related parameters, render the picture corresponding to the above parameters and output some auxiliary variables.

[0082] Further, the combined model corresponding to the current frame is rendered using a target fusion renderer to obtain a combined scene image, including:

[0083] All dynamic objects in the dynamic object list involving coordinate transformation attributes are converted to the world coordinate system through the camera internal and external parameters of the static background participating in the dynamic objects, and the coordinate attributes of the dynamic objects are obtained;

[0084] Combining the rendering related attributes of all dynamic objects with the rendering related attributes of the static background to obtain a combined rendering attribute;

[0085] The dynamic object coordinate attributes and the combined rendering attributes are summarized to obtain the combined rendering parameters, the combined model and the combined rendering parameters are input into the target fusion renderer, and the differentiable Gaussian rasterization module is used to render the combined scene image based on the Gaussian splatting scene rendering algorithm.

[0086] In the embodiment of the present invention, the dynamic object list refers to a list obtained by summarizing multiple dynamic objects in a traffic scene.

[0087] In the embodiment of the present invention, the dynamic object coordinate attributes include "point cloud coordinates", "rotation attributes", and "spherical harmonic function coefficients".

[0088] In the embodiment of the present invention, the combined rendering attributes include attributes such as transparency, scales, etc. that do not involve coordinate transformation.

[0089] Exemplarily, the following implementation steps may be used to optimize the point cloud position and Gaussian sphere properties of the static background and the point cloud position and Gaussian sphere properties of the dynamic object:

[0090] The coordinate transformation of the rotation attribute in the Gaussian sphere of a dynamic object follows the following formula:

[0091]

[0092] Among them, q ′ is the quaternion after coordinate transformation, q is the quaternion before coordinate transformation, is the rotation matrix from the dynamic object's own coordinate system to the world coordinate system, is the quaternion corresponding to the rotation matrix.

[0093] For example, the spherical harmonic coefficients in the Gaussian sphere attributes may be calculated using the following implementation steps:

[0094] For the spherical harmonic function coefficients, the relationship between the coefficients before and after the coordinate transformation is: the 0th-order coefficients remain unchanged, and the principle of high-order coefficient transformation is to ensure that the calculated color remains unchanged after the point cloud coordinates are transformed.

[0095] For the first-order coefficient, the formula for calculating the color affected by this part is:

[0096]

[0097] Among them, sh1 is the coefficient of the first spherical harmonic function, sh2 is the coefficient of the second spherical harmonic function, sh3 is the coefficient of the third spherical harmonic function, sh1, sh2 and sh3 are all three-dimensional column vectors, C is a constant, is the point cloud coordinate of the dynamic object, and RGB is the color component.

[0098] In the embodiment of the present invention, the preset scene rendering algorithm refers to a static gaussian-splatting algorithm.

[0099] The present invention constructs a dynamic object model based on the dynamic object point cloud and a static background model based on the static background point cloud, so as to separate the dynamic objects and the static background in the traffic scene, and avoid the influence of the motion state of the dynamic object on the extraction of the dynamic object in the traffic scene. In addition, by combining the dynamic object model and the static background model, the dynamic object and the static background in the traffic scene can be combined, so that the combined scene image can be rendered using the preset Gaussian splatting-based scene rendering algorithm, and the visual artifacts and blur in the rendered combined scene image can be effectively reduced, so as to improve the accuracy of separation of dynamic objects in the traffic scene and improve the image reconstruction effect of dynamic objects in the traffic scene.

[0100] Reference Figure 2 , which is a flowchart of the algorithm structure of a method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm provided by an embodiment of the present invention.

[0101] Reference Figure 3 , which is a complete schematic diagram of a method for separating dynamic objects in a traffic scene based on a Gaussian splatting algorithm according to an embodiment of the present invention;

[0102] Reference Figure 4 , which is a schematic diagram of a split static background picture of a method for separating dynamic objects in a traffic scene based on a Gaussian splatting algorithm provided by an embodiment of the present invention;

[0103] Reference Figure 5 , which is a schematic diagram of a split dynamic object picture of a traffic scene dynamic object separation method based on a Gaussian splatting algorithm provided by an embodiment of the present invention;

[0104] Reference Figure 6, which is a schematic diagram of a combined scene image obtained by rendering using a target fusion renderer in a method for separating dynamic objects in a traffic scene based on a Gaussian splatting algorithm provided by an embodiment of the present invention;

[0105] like Figure 7 , which is a functional module diagram of a traffic scene dynamic object separation device based on a Gaussian splatting algorithm provided by an embodiment of the present invention.

[0106] The traffic scene dynamic object separation device 100 based on the Gaussian splatting algorithm of the present invention can be installed in an electronic device. According to the functions implemented, the traffic scene dynamic object separation device 100 based on the Gaussian splatting algorithm can include a point cloud acquisition module 101, a model construction module 102 and a scene rendering module 103.

[0107] The module described in the present invention may also be referred to as a unit, which refers to a series of computer program segments that can be executed by a processor of an electronic device and can complete fixed functions, and is stored in a memory of the electronic device.

[0108] In this embodiment, the functions of each module / unit are as follows:

[0109] The point cloud acquisition module 101 is used to acquire the point cloud of dynamic objects and the point cloud of static background in the traffic scene.

[0110] In the embodiment of the present invention, a dynamic object refers to an object that can move in a traffic scene, such as a moving vehicle and a walking pedestrian.

[0111] In the embodiment of the present invention, the static background refers to a background formed by objects that cannot move in a traffic scene.

[0112] The embodiment of the present invention obtains the dynamic object point cloud and the static background point cloud in the traffic scene through the laser radar equipment.

[0113] As an embodiment of the present invention, before obtaining the point cloud of dynamic objects and the point cloud of static background in the traffic scene, it also includes: obtaining the camera internal and external parameters of the static background, wherein the camera external parameters of the static background convert the point cloud of the static background in the world coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the static background are projected and transformed to obtain the background part in the static background picture.

[0114] As an embodiment of the present invention, before obtaining the point cloud of dynamic objects and the static background point cloud in the traffic scene, it also includes: obtaining the camera internal and external parameters of the dynamic object, wherein the camera external parameters of the dynamic object convert the point cloud of the dynamic object in the dynamic object's own coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the dynamic object are projected and transformed to obtain the object image.

[0115] In an embodiment of the present invention, camera intrinsic and extrinsic parameters refer to key parameters for projecting points in a three-dimensional world onto a two-dimensional image plane, wherein the camera intrinsic and extrinsic parameters include camera intrinsic parameters and camera extrinsic parameters. Camera intrinsic parameters refer to the inherent properties of the camera itself, describing how the camera maps points in three-dimensional space to a two-dimensional image plane; camera extrinsic parameters refer to the position and direction of the camera in the world coordinate system, and the camera extrinsic parameters convert points in the world coordinate system to the camera coordinate system.

[0116] In an embodiment of the present invention, camera intrinsic and extrinsic parameters refer to key parameters for projecting points in a three-dimensional world onto a two-dimensional image plane, wherein the camera intrinsic and extrinsic parameters include camera intrinsic parameters and camera extrinsic parameters. Camera intrinsic parameters refer to the inherent properties of the camera itself, describing how the camera maps points in three-dimensional space to a two-dimensional image plane; camera extrinsic parameters refer to the position and direction of the camera in the world coordinate system, and the camera extrinsic parameters convert points in the world coordinate system to the camera coordinate system.

[0117] Furthermore, before obtaining the point cloud of dynamic objects and the point cloud of static background in the traffic scene, it also includes: obtaining the camera internal and external parameters of the static background, wherein the camera external parameters of the static background convert the point cloud of the static background in the world coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the static background are projected and transformed to obtain the background part in the static background picture.

[0118] Furthermore, before obtaining the point cloud of dynamic objects and the static background point cloud in the traffic scene, it also includes: obtaining the camera internal and external parameters of the dynamic object, wherein the camera external parameters of the dynamic object convert the point cloud of the dynamic object in the dynamic object's own coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the dynamic object are projected and transformed to obtain the object image.

[0119] The model building module 102 is used to build a dynamic object model according to the dynamic object point cloud, and to build a static background model according to the static background point cloud.

[0120] The scene rendering module 103 is used to combine the dynamic object model and the static background model, and render a combined scene image using a preset Gaussian splatting-based scene rendering algorithm.

[0121] As an embodiment of the present invention, the dynamic object model and the static background model are combined, including:

[0122] The dynamic object model, the static background model, the camera extrinsic parameters of the dynamic object and the camera extrinsic parameters of the static background are combined to obtain a combined model.

[0123] In the embodiment of the present invention, the combined model is equivalent to a model in which the dynamic object point cloud and the static background are in the same scene and fixed in the current frame.

[0124] Furthermore, the combined model includes:

[0125] In the combined model, for the point cloud of the dynamic object itself, the dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the external parameters of the dynamic object, and then the point cloud is transformed to the world coordinate system through the external parameters of the static background. The transformation process is as follows:

[0126] The dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the dynamic object's external parameters:

[0127]

[0128] Among them, P c is the coordinate of the point cloud of the dynamic object in the camera coordinate system, P o is the coordinate of the point cloud of the dynamic object in the dynamic object's own coordinate system, It is the transformation matrix from the dynamic object's own coordinate system to the camera coordinate system;

[0129] Then transform the point cloud to the world coordinate system through the external parameters of the static background:

[0130]

[0131] Among them, P ω is the coordinate of the point cloud of the dynamic object in the world coordinate system, is the transformation matrix from the camera coordinate system to the world coordinate system;

[0132] The merge transformation process is as follows:

[0133]

[0134] As an embodiment of the present invention, a combined scene image is rendered using a preset scene rendering algorithm based on Gaussian splatting, including:

[0135] Randomly select a camera view of a static background, and query the dynamic object model corresponding to the camera view in the complete picture of the dynamic object and the camera internal and external parameters of the dynamic object corresponding to the current frame;

[0136] The dynamic object model corresponding to the camera perspective in the complete picture of the dynamic object and the internal and external parameters of the camera of the dynamic object corresponding to the current frame are input into the fusion renderer, and the initial rendering image is rendered using the scene rendering algorithm preset in the fusion renderer;

[0137] Calculating the loss value of the initial rendered image and the complete image of the dynamic object, optimizing the point cloud position and Gaussian sphere attributes of the static background and the point cloud position and Gaussian sphere attributes of the dynamic object according to the loss value, and iteratively optimizing the fusion renderer to obtain the target fusion renderer;

[0138] The combined model corresponding to the current frame is rendered using a target fusion renderer to obtain a combined scene image.

[0139] In the embodiment of the present invention, the fusion renderer mainly defines a render function, whose function is to input a static background model of a certain frame and the internal and external parameters of the static background camera, a dynamic object list, the internal and external parameters of the dynamic object camera, and other rendering related parameters, render the picture corresponding to the above parameters and output some auxiliary variables.

[0140] Furthermore, the combined model corresponding to the current frame is rendered using a target fusion renderer to obtain a combined scene image, including:

[0141] All dynamic objects in the dynamic object list involving coordinate transformation attributes are converted to the world coordinate system through the camera internal and external parameters of the static background participating in the dynamic objects, and the coordinate attributes of the dynamic objects are obtained;

[0142] Combining the rendering related attributes of all dynamic objects with the rendering related attributes of the static background to obtain a combined rendering attribute;

[0143] The dynamic object coordinate attributes and the combined rendering attributes are summarized to obtain the combined rendering parameters, the combined model and the combined rendering parameters are input into the target fusion renderer, and the differentiable Gaussian rasterization module is used to render the combined scene image based on the Gaussian splatting scene rendering algorithm.

[0144] In the embodiment of the present invention, the dynamic object list refers to a list obtained by summarizing multiple dynamic objects in a traffic scene.

[0145] In the embodiment of the present invention, the dynamic object coordinate attributes include "point cloud coordinates", "rotation attributes", and "spherical harmonic function coefficients".

[0146] In the embodiment of the present invention, the combined rendering attributes include attributes such as transparency, scales, etc. that do not involve coordinate transformation.

[0147] Exemplarily, the following implementation steps may be used to optimize the point cloud position and Gaussian sphere properties of the static background and the point cloud position and Gaussian sphere properties of the dynamic object:

[0148] The coordinate transformation of the rotation attribute in the Gaussian sphere of a dynamic object follows the following formula:

[0149]

[0150] Among them, q ′ is the quaternion after coordinate transformation, q is the quaternion before coordinate transformation, is the rotation matrix from the dynamic object's own coordinate system to the world coordinate system, is the quaternion corresponding to the rotation matrix.

[0151] For example, the spherical harmonic coefficients in the Gaussian sphere attributes may be calculated using the following implementation steps:

[0152] For the spherical harmonic function coefficients, the relationship between the coefficients before and after the coordinate transformation is: the 0th-order coefficients remain unchanged, and the principle of high-order coefficient transformation is to ensure that the calculated color remains unchanged after the point cloud coordinates are transformed.

[0153] For the first-order coefficient, the formula for calculating the color affected by this part is:

[0154]

[0155] Among them, sh1 is the coefficient of the first spherical harmonic function, sh2 is the coefficient of the second spherical harmonic function, sh3 is the coefficient of the third spherical harmonic function, sh1, sh2 and sh3 are all three-dimensional column vectors, C is a constant, is the point cloud coordinate of the dynamic object, and RGB is the color component.

[0156] In the embodiment of the present invention, the preset scene rendering algorithm refers to a static gaussian-splatting algorithm.

[0157] like Figure 8 , which is a schematic diagram of the structure of an electronic device for implementing a method for separating dynamic objects in a traffic scene based on a Gaussian splatting algorithm, provided in one embodiment of the present invention.

[0158] The electronic device may include a processor 10, a memory 11, a communication bus 12 and a communication interface 13, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as a method program for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm.

[0159] In some embodiments, the processor 10 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 10 is the control core (ControlUnit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes or executes programs or modules stored in the memory 11 (for example, executing a traffic scene dynamic object separation method program based on a gaussian splatting algorithm, etc.), and calls data stored in the memory 11 to execute various functions of the electronic device and process data.

[0160] The memory 11 includes at least one type of readable storage medium, and the readable storage medium includes a flash memory, a mobile hard disk, a multimedia card, a card-type memory (for example, an SD or DX memory, etc.), a magnetic memory, a disk, an optical disk, etc. In some embodiments, the memory 11 may be an internal storage unit of an electronic device, such as a mobile hard disk of the electronic device. In other embodiments, the memory 11 may also be an external storage device of an electronic device, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device. Further, the memory 11 may also include both an internal storage unit of the electronic device and an external storage device. The memory 11 may not only be used to store application software and various types of data installed in the electronic device, such as a code of a traffic scene dynamic object separation method program based on a gaussian splatting algorithm, but may also be used to temporarily store data that has been output or is to be output.

[0161] The communication bus 12 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The bus is configured to realize connection and communication between the memory 11 and at least one processor 10, etc.

[0162] The communication interface 13 is used for communication between the above-mentioned electronic device and other devices, including a network interface and a user interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the electronic device and other electronic devices. The user interface may be a display (Display), an input unit (such as a keyboard (Keyboard)), and optionally, the user interface may also be a standard wired interface, a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, etc. Among them, the display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the electronic device and to display a visual user interface.

[0163] Figure 8 Only an electronic device with components is shown, and those skilled in the art will understand that Figure 8 The structure shown does not constitute a limitation on the electronic device, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0164] For example, although not shown, the electronic device may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 10 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0165] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0166] The program of a method for separating dynamic objects in traffic scenes based on a Gaussian splatting algorithm stored in the memory 11 of the electronic device is a combination of multiple instructions. When running in the processor 10, the following can be achieved:

[0167] Obtain point clouds of dynamic objects and static background in traffic scenes;

[0168] constructing a dynamic object model according to the dynamic object point cloud, and constructing a static background model according to the static background point cloud;

[0169] The dynamic object model and the static background model are combined, and a combined scene image is rendered using a preset scene rendering algorithm based on Gaussian splatting.

[0170] Specifically, the specific implementation method of the processor 10 for the above instructions can refer to the description of the relevant steps in the corresponding embodiment of the accompanying drawings, which will not be repeated here.

[0171] Furthermore, if the module / unit integrated in the electronic device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

[0172] The present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of an electronic device, the computer program can implement:

[0173] Obtain point clouds of dynamic objects and static background in traffic scenes;

[0174] constructing a dynamic object model according to the dynamic object point cloud, and constructing a static background model according to the static background point cloud;

[0175] The dynamic object model and the static background model are combined, and a combined scene image is rendered using a preset scene rendering algorithm based on Gaussian splatting.

[0176] In the several embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0177] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0178] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0179] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0180] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0181] The blockchain referred to in this invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0182] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0183] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.

[0184] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for separating dynamic objects in traffic scenes based on Gaussian splatting algorithm, characterized in that: The method comprises: Obtain point clouds of dynamic objects and static background in traffic scenes; constructing a dynamic object model according to the dynamic object point cloud, and constructing a static background model according to the static background point cloud; The dynamic object model and the static background model are combined, and a combined scene image is rendered using a preset gaussiansplatting-based scene rendering algorithm.

2. The method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm according to claim 1, characterized in that: Before obtaining the point cloud of dynamic objects and the point cloud of static background in the traffic scene, it also includes: obtaining the camera internal and external parameters of the static background, wherein the camera external parameters of the static background convert the point cloud of the static background in the world coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the static background are projected and transformed to obtain the background part in the static background picture.

3. The method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm according to claim 1 or 2, characterized in that: Before obtaining the point cloud of dynamic objects and the static background point cloud in the traffic scene, it also includes: obtaining the camera internal and external parameters of the dynamic object, wherein the camera external parameters of the dynamic object convert the point cloud of the dynamic object in the dynamic object's own coordinate system into the point cloud in the camera coordinate system, and the camera internal parameters of the dynamic object are projected and transformed to obtain the object image.

4. The method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm as claimed in claim 3, characterized in that: The combining the dynamic object model and the static background model comprises: The dynamic object model, the static background model, the camera extrinsic parameters of the dynamic object and the camera extrinsic parameters of the static background are combined to obtain a combined model.

5. The method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm as claimed in claim 4, characterized in that: The combined model includes: In the combined model, for the point cloud of the dynamic object itself, the dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the external parameters of the dynamic object, and then the point cloud is transformed to the world coordinate system through the external parameters of the static background. The transformation process is as follows: The dynamic object point cloud is transformed from the dynamic object's own coordinate system to the camera coordinate system through the dynamic object's external parameters: Among them, P c is the coordinate of the point cloud of the dynamic object in the camera coordinate system, P o is the coordinate of the point cloud of the dynamic object in the dynamic object's own coordinate system, It is the transformation matrix from the dynamic object's own coordinate system to the camera coordinate system; Then transform the point cloud to the world coordinate system through the external parameters of the static background: Among them, P w is the coordinate of the point cloud of the dynamic object in the world coordinate system, is the transformation matrix from the camera coordinate system to the world coordinate system; The merge transformation process is as follows:

6. The method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm according to claim 1, characterized in that: The method of rendering a combined scene image by using a preset scene rendering algorithm based on Gaussian splatting includes: Randomly select a camera view of a static background, and query the dynamic object model corresponding to the camera view in the complete picture of the dynamic object and the camera internal and external parameters of the dynamic object corresponding to the current frame; The dynamic object model corresponding to the camera perspective in the complete picture of the dynamic object and the internal and external parameters of the camera of the dynamic object corresponding to the current frame are input into the fusion renderer, and the initial rendering image is rendered using the scene rendering algorithm preset in the fusion renderer; Calculating the loss value of the initial rendered image and the complete image of the dynamic object, optimizing the point cloud position and Gaussian sphere attributes of the static background and the point cloud position and Gaussian sphere attributes of the dynamic object according to the loss value, and iteratively optimizing the fusion renderer to obtain the target fusion renderer; The combined model corresponding to the current frame is rendered using a target fusion renderer to obtain a combined scene image.

7. The method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm as claimed in claim 6, characterized in that: The combining model corresponding to the current frame is rendered using a target fusion renderer to obtain a combined scene image, including: All dynamic objects in the dynamic object list involving coordinate transformation attributes are converted to the world coordinate system through the camera internal and external parameters of the static background participating in the dynamic objects, and the coordinate attributes of the dynamic objects are obtained; Combining the rendering related attributes of all dynamic objects with the rendering related attributes of the static background to obtain a combined rendering attribute; The dynamic object coordinate attributes and the combined rendering attributes are summarized to obtain the combined rendering parameters, the combined model and the combined rendering parameters are input into the target fusion renderer, and the differentiable Gaussian rasterization module is used to render the combined scene image based on the Gaussian splatting scene rendering algorithm.

8. A dynamic object separation device for traffic scenes based on Gaussian splatting algorithm, characterized in that: The device can implement the method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm as claimed in any one of claims 1 to 7, and the device comprises: Point cloud acquisition module, used to acquire point clouds of dynamic objects and static background in traffic scenes; A model building module, used to build a dynamic object model according to the dynamic object point cloud, and to build a static background model according to the static background point cloud; The scene rendering module is used to combine the dynamic object model and the static background model, and render a combined scene image using a preset scene rendering algorithm based on Gaussian splatting.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method for separating dynamic objects in traffic scenes based on the Gaussian splatting algorithm as described in any one of claims 1 to 7 is implemented.

Images