A method, system, and terminal device for displaying a 3D geographic information system based on a multi-screen display.

Abstract

This invention discloses a method, system, and terminal device for displaying a 3D geographic information system based on multiple screens, relating to the field of 3D display technology. The method includes: acquiring 3D geographic information data; constructing a 3D geographic scene in a 3D rendering engine; determining the observer's viewpoint within a physical space, which is enclosed by multiple planar display screens forming a multi-screen display; deploying a corresponding number of virtual cameras facing the corresponding screens, with the observer's viewpoint as the origin; configuring the cameras in a physical trapezoidal combination based on the physical size and position of the screens relative to the observer and calculating the asymmetric view frustum projection matrix; allocating rendering tasks to pre-built distributed rendering cluster nodes, and controlling the synchronous output of each camera's image to the corresponding screen through a hardware synchronization mechanism. This invention achieves naked-eye stereoscopic display on multiple screens, eliminates perspective distortion at screen junctions, ensures seamless synchronization of multi-screen images without tearing, and enhances the immersiveness and accuracy of 3D geographic information display.

Description

Technical Field

[0001] This invention relates to the field of 3D display technology, and in particular to a method, system, and terminal device for displaying 3D geographic information systems based on multi-screen displays. Background Technology

[0002] Against the backdrop of digital transformation and smart city construction, the integration of high-precision GIS (Geographic Information System) and high-fidelity visual presentation technology has become an important tool for natural resource management, urban planning simulation, and macro-level decision support. Mega-cities demand highly immersive and interactive presentations of planning and land use results, requiring natural resource simulation experiments to support massive amounts of real-world 3D data and provide deeply immersive spaces with high rendering quality.

[0003] Current mainstream immersive display solutions all have significant shortcomings. VR (Virtual Reality) headsets are prone to causing dizziness and cannot support multi-person collaborative rendering. CAVE (Cave Automatic Virtual Environment) is costly and prone to image gaps. WebGL (Web Graphics Library) 3D GIS is limited by browser performance, making it difficult to achieve large-scale high-fidelity rendering. Conventional large-screen splicing lacks a sense of depth, and multi-machine rendering is prone to screen tearing. Furthermore, existing technologies suffer from incompatibility between GIS and rendering engines such as UE4 (Unreal Engine 4) coordinate systems, and insufficient synchronization accuracy in distributed rendering.

[0004] Therefore, there is an urgent need for a naked-eye stereoscopic display method for 3D geographic information based on multi-screen displays to fill the gaps in existing technologies. Summary of the Invention

[0005] The technical problem this invention aims to solve is that, in the field of GIS (Geographic Information System) 3D display, existing multi-screen solutions suffer from perspective distortion at screen junctions and poor synchronization and tearing issues in distributed rendering. Therefore, an effective solution is urgently needed to address these technical problems.

[0006] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: In a first aspect, the present invention provides a method for displaying a three-dimensional geographic information system based on a multi-screen display, the method comprising: Acquire 3D geographic information data and construct 3D geographic scenes in a 3D rendering engine; Determine the observer's viewpoint within a physical space, wherein a multi-sided screen is provided within the physical space, and the multi-sided screen is formed by at least three planar display screens that form an angle with each other; In the 3D rendering engine, virtual cameras are deployed with the observer's viewpoint as the origin, and the number of virtual cameras is the same as the number of multi-faceted screens. Each virtual camera faces the corresponding flat display screen. Based on the physical size and spatial position of each planar display screen relative to the observer's viewpoint, a physical trapezoidal combination configuration is performed for each virtual camera, and the projection matrix of the corresponding asymmetric view frustum is calculated. The rendering task of each virtual camera is assigned to the corresponding computing node of the pre-built distributed rendering cluster, and the rendering images of all virtual cameras are synchronously output to the corresponding flat display screen at the same time through a hardware synchronization mechanism. Based on the physical dimensions and spatial positions of each planar display screen relative to the observer's viewpoint, a physical trapezoidal combination configuration is performed for each virtual camera, and the projection matrix of the corresponding asymmetric view frustum is calculated, including: For any flat display screen, the projection boundary of the screen relative to the observer's viewpoint is determined based on the coordinates of the four corner points of the flat display screen in physical space. Configure the projection surface of the virtual camera corresponding to the flat display screen as a trapezoid that perfectly matches the projection boundary; The vertical distance from the observer's viewpoint to the plane of the flat display screen, the near clipping plane distance of the virtual camera, and the horizontal and vertical offsets of the observer's viewpoint relative to the center of the flat display screen are obtained as raw parameters. Based on the original parameters, the left and right boundary parameters and the top and bottom boundary parameters of the near-cut surface are calculated, and the projection matrix of the asymmetric view frustum is generated.

[0007] In one implementation, acquiring three-dimensional geographic information data and constructing a three-dimensional geographic scene in a three-dimensional rendering engine includes: Obtain raw geospatial data in a geodetic coordinate system, wherein the raw geospatial data includes oblique photogrammetry model, terrain data, and image data; Convert the latitude and longitude coordinates of the original geospatial data into geocentric rectangular coordinates; The geocentric rectangular coordinates are converted into local tangent plane coordinates with the physical location of the observer's viewpoint as the origin, thus constructing the mapping of 3D geographic information data in the 3D rendering engine.

[0008] In one implementation, determining the observer's viewpoint within the physical space includes: A predetermined fixed position in physical space is used as the observer's viewpoint; Alternatively, visual capture devices can be used to obtain the real-time position of the observer's head or eyeballs in the physical space as the observer's viewpoint.

[0009] In one implementation, the step of allocating the rendering task of each virtual camera to the corresponding computing node of a pre-built distributed rendering cluster includes: For any virtual camera, based on the current position of the virtual camera and the coverage of the asymmetric view frustum, calculate the tile detail level and obtain the tile data of the corresponding level; The visibility of the tile data is determined, and tiles that are below the observer's horizon or are obscured by other objects are removed from the rendering task. The white models of buildings generated from vector data are overlaid on the three-dimensional geographic scene, and display materials are assigned according to the vector attributes; Based on the removed tile data and the white model of the building, the portion of the three-dimensional geographic scene corresponding to the virtual camera is allocated to the corresponding computing nodes of the distributed rendering cluster for rendering.

[0010] In one implementation, the distributed rendering cluster includes at least one master computing node, and the graphics cards of all computing nodes in the distributed rendering cluster are daisy-chained together via synchronization cables. The step of controlling the rendering images from all virtual cameras to be synchronously output to the corresponding flat-panel display screens at the same time via a hardware synchronization mechanism includes: The graphics card of the master computing node emits a synchronization pulse signal as the global clock source for the distributed rendering cluster. Align the refresh rate of the graphics cards on all computing nodes with the vertical synchronization signal of the corresponding flat-panel display screen; Configure a swap barrier strategy so that after all computing nodes have completed all rendering tasks for the current frame, they simultaneously perform a buffer swap operation and output the rendered screen.

[0011] In one implementation, after controlling the rendered images of all virtual cameras to be synchronously output to the corresponding flat display screen at the same time through the hardware synchronization mechanism, the method further includes: The rendered image of the 3D geographic scene is encoded into a video stream, and the video stream is pushed to the mobile control terminal. Receive operation instructions sent by the mobile control terminal and convert the operation instructions into logical instructions that can be recognized by the 3D rendering engine; Based on the aforementioned logical instructions, the position, rotation angle, and scaling of all virtual cameras are updated synchronously.

[0012] In one implementation, the multi-sided screen is a four-sided screen, including a front screen, a left screen, a right screen and a bottom screen. The front screen, left screen and right screen are connected to each other in pairs, and the bottom screen is horizontally set at the bottom of the front screen, left screen and right screen, together enclosing a semi-enclosed display space. The four virtual cameras are facing directly forward, left, right, and down. The pre-built distributed rendering cluster includes one master computing node and at least three slave computing nodes. The master computing node is responsible for the rendering tasks of the virtual camera corresponding to the front screen and the issuance of global scheduling synchronization instructions. The slave computing nodes are respectively responsible for the rendering tasks of the virtual cameras corresponding to the left screen, right screen and ground screen.

[0013] Secondly, embodiments of the present invention also provide a multi-screen-based 3D geographic information system display system, applied to the steps of implementing any one of the above-described solutions for displaying a multi-screen-based 3D geographic information system, the system comprising: The 3D geographic scene construction module is used to acquire 3D geographic information data and construct 3D geographic scenes in the 3D rendering engine. An observer viewpoint determination module is used to determine the observer viewpoint within a physical space, wherein a multi-sided screen is provided within the physical space, and the multi-sided screen is formed by at least three planar display screens that form an angle with each other. The virtual camera deployment module is used in the 3D rendering engine to deploy the same number of virtual cameras as the multi-faceted screen, with the observer's viewpoint as the origin, and each virtual camera facing the corresponding planar display screen; The physical trapezoidal combination calculation module is used to configure the physical trapezoidal combination of each virtual camera based on the physical size and spatial position of each planar display screen relative to the observer's viewpoint, and calculate the projection matrix of the corresponding asymmetric view frustum. The distributed rendering module is used to allocate the rendering tasks of each virtual camera to the corresponding computing nodes of the pre-built distributed rendering cluster, and control the rendering images of all virtual cameras to be synchronously output to the corresponding flat display screen at the same time through a hardware synchronization mechanism.

[0014] Thirdly, embodiments of the present invention also provide a terminal device, the terminal device including a memory, a processor, and a multi-screen-based three-dimensional geographic information system display program stored in the memory and executable on the processor. When the processor executes the multi-screen-based three-dimensional geographic information system display program, it implements the steps of the multi-screen-based three-dimensional geographic information system display method described in any of the above schemes.

[0015] Beneficial Effects: This invention discloses a method, system, and terminal device for displaying a 3D geographic information system based on multiple screens, relating to the field of 3D display technology. The method first acquires 3D geographic information data, constructs a 3D geographic scene in a 3D rendering engine, and determines the observer's viewpoint within the physical space. The physical space contains multiple screens, each formed by at least three planar display screens forming an angle with each other. Then, in the 3D rendering engine, with the observer's viewpoint as the origin, the same number of virtual cameras as the number of multiple screens are deployed, each virtual camera facing its corresponding planar display screen. Next, based on the physical size and spatial position of each planar display screen relative to the observer's viewpoint, a physical trapezoidal combination configuration is performed for each virtual camera, and the projection matrix of the corresponding asymmetric view frustum is calculated. Finally, the rendering task of each virtual camera is assigned to the corresponding computing node of a pre-constructed distributed rendering cluster, and a hardware synchronization mechanism controls the simultaneous output of the rendered images from all virtual cameras to the corresponding planar display screens. This invention eliminates perspective distortion and image misalignment at the junction of multiple screens by using physical trapezoidal combination configuration and asymmetric frustum projection matrix calculation, achieving a naked-eye stereoscopic display effect without the need for wearing devices. Furthermore, it employs a distributed rendering cluster combined with a hardware synchronization mechanism to ensure that all screen images are output synchronously at the millisecond level, solving the tearing problem of high-speed dynamic images and significantly improving the immersiveness, smoothness, and geographic coordinate accuracy of 3D geographic information display. Attached Figure Description

[0016] Figure 1 A flowchart illustrating a specific implementation method for a multi-screen-based 3D geographic information system display method provided in this invention.

[0017] Figure 2 This is a schematic diagram illustrating the display effect of a multi-screen-based three-dimensional geographic information system display method provided in an embodiment of the present invention.

[0018] Figure 3 This is a schematic diagram of the distributed rendering cluster network topology for the multi-screen-based 3D geographic information system display method provided in this embodiment of the invention.

[0019] Figure 4 This is a logical diagram of the physical trapezoidal combination projection of a four-camera system for displaying a three-dimensional geographic information system based on a multi-screen display method provided in an embodiment of the present invention.

[0020] Figure 5 This is a data interaction flow rendering sequence diagram for the multi-screen-based 3D geographic information system display method provided in the embodiments of the present invention.

[0021] Figure 6 This is a schematic diagram of the principle of a three-dimensional geographic information system display device based on a multi-screen provided in an embodiment of the present invention.

[0022] Figure 7 This is a block diagram illustrating the internal structure of the terminal device provided in an embodiment of the present invention. Detailed Implementation

[0023] To make the objectives, technical solutions, and effects of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0024] The flowchart shown in the attached diagram is for illustrative purposes only and does not necessarily include all content, operations, or steps, nor does it require execution in the described order. For example, some operations or steps can be broken down, combined, or partially merged, so the actual execution order may change depending on the actual situation.

[0025] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0026] It should be understood that, in order to clearly describe the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. For example, "first control information" and "second control information" are only used to distinguish different control information and do not limit their order.

[0027] Those skilled in the art will understand that the words "first" and "second" do not limit the quantity or the order of execution, and that the words "first" and "second" do not necessarily imply that they are different.

[0028] It should also be understood that the term “and / or” as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0029] Against the backdrop of digital transformation and smart city construction, the integration of high-precision 3D geographic information systems and high-fidelity visual presentation technologies has become an important tool for natural resource management, urban planning simulation, and macro-level decision support. Mega-cities place extremely high demands on the immersive and interactive nature of their planning and land use results, particularly in the field of natural resource simulation experiments, which requires deeply immersive spaces capable of handling massive amounts of real-world 3D data, possessing high-fidelity rendering quality, and supporting multi-device collaborative operation.

[0030] Currently, the mainstream immersive or stereoscopic display solutions in the industry mainly include virtual reality headsets, traditional cave-like automated virtual display systems, and 3D GIS based on web-based graphics libraries. However, these solutions have problems in specific business scenarios.

[0031] While virtual reality headsets offer a strong sense of personal immersion and panoramic views, their weight can cause dizziness with prolonged use. Furthermore, the screen-door effect results in noticeable graininess on the screen and hinders collaborative discussions and presentations among multiple users in the same physical space. Traditional cave-like automated virtual display systems use room-based projection, eliminating the need for heavy equipment and providing wide field of view. However, these systems are extremely large and complex, incurring high construction costs. Imaging is also limited by the physical constraints of the projector or display screen, often resulting in overlapping or gaps at the edges of the image. Multi-channel synchronization is also difficult, making it challenging to achieve high-precision image consistency. Web-based 3D GIS solutions, implemented through a browser, offer convenient deployment and strong cross-platform capabilities. However, limited by the browser kernel and graphics card performance, their performance in large-scale realistic 3D scenes, highly dynamic lighting effects, and complex physical rendering is far inferior to desktop rendering engines. Conventional large-screen splicing solutions, while offering high resolution and large display area, lack true depth and cannot adjust for perspective based on the observer's viewpoint. They are also prone to severe screen tearing during multi-machine rendering.

[0032] When constructing a naked-eye stereoscopic space based on multiple screens, such as a multi-screen environment consisting of three vertical walls and a horizontal floor, three technical challenges are faced.

[0033] First, there's the compatibility issue between GIS data and high-fidelity simulation engines. The geographic information industry generally uses right-handed coordinate systems, such as WGS84 and the Earth-Centered, Earth-Fixed (ECEF) coordinate system, and has strict requirements for the Earth's curvature and the Coordinate Reference System (CRS). However, mainstream high-performance game engines like Unreal Engine 4 (UE4) use left-handed coordinate systems, and their native design tends to favor small-scene rendering in local coordinate systems. This coordinate system conflict, along with the lack of direct support for spatial indexing standards such as 3D tiles, makes it difficult to maintain accurate physical scale and geographic positioning of realistic 3D city floor plans within game engines.

[0034] Secondly, there's the issue of distributed rendering synchronization at ultra-large resolutions. To cover all areas of a multi-screen display and maintain extremely high clarity, rendering tasks typically need to be distributed across a master-slave cluster of multiple high-performance computing nodes. However, in fast-paced dynamic scenes, if the frame swapping (swap buffer) between nodes cannot be synchronized with millisecond precision, noticeable screen tearing or gaps will appear at the screen connections due to differences in graphics card rendering load. This significantly diminishes the realism of the immersive experience.

[0035] Finally, there's the issue of real-time and collaborative multi-device interaction. In immersive labs, users often need to interact with large-screen scenes in real time via mobile devices or external capture equipment. Traditional client / server (C / S) architectures place excessive demands on mobile hardware, making it impossible to render high-fidelity images. Meanwhile, conventional cloud rendering solutions often suffer from significant lag between interactive operations and visual feedback in ultra-high-resolution multi-screen environments due to long command return paths and high video encoding latency.

[0036] The aforementioned existing technologies cannot simultaneously solve problems such as perspective distortion, insufficient synchronization accuracy of distributed rendering, and difficulty in compatibility between GIS data and high-fidelity simulation engines in multi-screen naked-eye stereoscopic displays, making it difficult to meet the demand for high-precision, highly immersive, and highly interactive 3D geographic information displays.

[0037] This embodiment provides a method for displaying a 3D geographic information system based on a multi-screen display, such as... Figure 1 As shown, the specific steps include the following: Step S100: Obtain 3D geographic information data and construct a 3D geographic scene in the 3D rendering engine.

[0038] In this embodiment, 3D geographic information data refers to digital information describing the spatial location and morphological features of the Earth's surface, specifically including oblique photogrammetry models, digital elevation models, high-resolution satellite imagery, vector terrain data, etc. The 3D rendering engine is a software platform with real-time 3D graphics rendering capabilities, specifically including Unreal Engine, Unity Engine, or other high-performance rendering engines. This embodiment uses a 3D rendering engine to integrate discrete geospatial data into a continuous, interactive 3D virtual scene.

[0039] When constructing a 3D geographic scene, the acquired raw geographic data is first preprocessed, including data format conversion, topological relationship checking, texture compression, and model simplification. The preprocessed geographic data is organized according to a spatial index structure for rapid retrieval and loading. Based on a preset geographic origin and coordinate system mapping, the 3D rendering engine places the geographic data in corresponding positions in virtual space, constructing a 3D geographic scene with realistic geographic coordinates. This enables the visualization of massive amounts of geographic data and provides the content foundation for multi-screen, glasses-free 3D display.

[0040] In one implementation, the step of acquiring 3D geographic information data and constructing a 3D geographic scene in a 3D rendering engine specifically includes the following steps: Step S110: Obtain raw geospatial data in the geodetic coordinate system, wherein the raw geospatial data includes oblique photogrammetry model, terrain data and image data; Step S120: Convert the latitude and longitude coordinates of the original geospatial data into geocentric rectangular coordinates; Step S130: Convert the geocentric rectangular coordinates into local tangent plane coordinates with the physical location of the observer's viewpoint as the origin, and construct the mapping of the three-dimensional geographic information data in the three-dimensional rendering engine.

[0041] In this embodiment, the geodetic coordinate system is a geographic coordinate system established based on the Earth's ellipsoid, used to represent the latitude and longitude coordinates of any point on Earth. Specifically, it can be the WGS84 coordinate system, CGCS2000 coordinate system, UTM coordinate system, etc. Raw geospatial data is usually stored and exchanged using the geodetic coordinate system, and its coordinate form is longitude, latitude, and elevation. The geocentric rectangular coordinate system is a three-dimensional Cartesian coordinate system with the Earth's center of mass as the origin. The X-axis points to the intersection of the Prime Meridian and the equator, the Y-axis points to the intersection of 90 degrees east longitude and the equator, and the Z-axis points to the Earth's North Pole. The local tangent plane coordinate system is an East-North-Sky coordinate system with the observer's physical location as the origin, and it is the working coordinate system used for scene rendering in the 3D rendering engine.

[0042] This embodiment uses a three-level precision transformation to achieve coordinate adaptation between geographic data and the rendering engine.

[0043] First, obtain the raw geospatial data in the geodetic coordinate system, whose coordinate form is (longitude) ,latitude ellipsoidal height ).

[0044] Subsequently, the latitude and longitude coordinates of the original geospatial data are converted to geocentric rectangular coordinates. The radius of curvature of the circumpolar region is then calculated. The calculation formula is:

[0045] in, For the semi-major axis of the ellipsoid, For the first eccentricity, Latitude.

[0046] Calculate the geocentric rectangular coordinates based on the radius of curvature of the circumpolar region. The calculation formula is:

[0047]

[0048]

[0049] in, Longitude This represents the height of the ellipsoid.

[0050] Finally, the geocentric rectangular coordinates are converted to local tangent plane coordinates with the physical location of the observer's viewpoint as the origin, and the East-North-Sky coordinate system is adopted as the local coordinate system. Transformation matrix It is calculated from the latitude and longitude of the observer's viewpoint, using the following formula:

[0051] This transformation matrix converts geocentric rectangular coordinates to local tangent plane coordinates, achieving a mapping from the right-handed coordinate system of the Geographic Information System (GIS) to the left-handed coordinate system of the 3D rendering engine. This step resolves the coordinate system conflict between GIS data and the game engine, ensuring the accuracy of geographic coordinates and spatial location in the 3D geographic scene.

[0052] Step S200: Determine the observer's viewpoint within the physical space, wherein the physical space is provided with a multi-sided screen, which is formed by at least three planar display screens that are angled to each other.

[0053] In this embodiment, the observer's viewpoint is a reference point in physical space used to determine the position and orientation of the virtual camera. It serves as the benchmark for constructing the naked-eye stereoscopic visual effect. For example, it could be the viewpoint position of a user's eyes when interacting with the multi-screen. The physical space is a closed or semi-closed indoor space used to deploy the multi-screen and accommodate the observer, specifically such as an exhibition hall, laboratory, or conference room. The multi-screen is a display system formed by multiple planar display screens arranged at a certain angle, including at least three planar display screens that form an angle with each other. The angle between adjacent screens is greater than 0 degrees and less than 180 degrees, and they collectively face the observer to form a semi-closed visual space.

[0054] Each flat display screen in the multi-screen system can use an ultra-high-definition LED display or an LCD display, featuring high brightness, high contrast, and fast response. The screens are fixedly connected by physical supports, with narrow bezels at the joints to reduce visual gaps in the image stitching. This embodiment provides a common reference point for the deployment and parameter configuration of all virtual cameras by establishing a unified observer's viewpoint, ensuring that the images on each screen form a unified perspective relationship to achieve a naked-eye stereoscopic effect.

[0055] In one implementation, determining the observer's viewpoint within the physical space specifically includes the following steps: Step S210a: Use a preset fixed position in the physical space as the observer's viewpoint; Step S210b, or, using a visual capture device, the position of the observer's head or eyeballs in the physical space is acquired in real time as the observer's viewpoint.

[0056] In this embodiment, the preset fixed position is a location pre-set in physical space. It can be an optimal viewing position that has been precisely measured and calibrated to achieve the best naked-eye stereoscopic visual effect. In implementations using the preset fixed position as the observer's viewpoint, all virtual camera parameters are configured once based on this fixed position, and the viewpoint remains unchanged during operation. This method is suitable for display scenarios where multiple people are watching simultaneously, ensuring that all viewers receive a consistent visual experience within the preset area.

[0057] Visual capture devices are sensor devices used to acquire observer position information in real time. Specifically, they can be monocular cameras, binocular cameras, infrared motion capture systems, optical tracking systems, etc. In implementations using visual capture devices, the system tracks the observer's head or eye position in real time, using the tracked 3D coordinates as a dynamically updated observer viewpoint. Based on the dynamically changing observer viewpoint, the asymmetric frustum projection matrix of all virtual cameras is recalculated in real time, adjusting the rendering of each screen. This method allows the observer to move freely in physical space while consistently achieving a good naked-eye stereoscopic effect, enhancing the interactivity and immersion of the display.

[0058] Step S300: In the 3D rendering engine, with the observer's viewpoint as the origin, deploy virtual cameras in the same number as the multi-faceted screen, with each virtual camera facing the corresponding planar display screen.

[0059] In this embodiment, the virtual camera is a virtual object in the 3D rendering engine used to simulate the human eye's viewing perspective. Its parameters determine the field of view, perspective, and visual effects of the rendered image. The realization of naked-eye stereoscopic vision is based on the topological layout of the virtual cameras in the 3D scene. This embodiment uses a camera configuration scheme based on the viewpoint center of the audience platform. This embodiment uses the observer's viewpoint as a common origin and deploys the same number of virtual cameras as the multi-screen display in the 3D rendering engine. Each virtual camera corresponds to a planar display screen, and the camera's optical axis points towards the center of the corresponding screen. All virtual cameras do not use a simple orthographic projection method but are configured according to the actual topological structure of the physical screens. By adjusting the camera's rotation and translation parameters, it is ensured that its field of view can completely cover the geometric range of the corresponding physical screen, while simultaneously ensuring that the view frustums of each camera form a continuous and non-overlapping stitching relationship in space. This topological layout fundamentally guarantees the perspective consistency of 3D objects when displayed across screens, avoiding image breaks or distortions.

[0060] The virtual camera's initial position coincides with the observer's viewpoint, and the camera's orientation is adjusted according to the spatial position of the corresponding screen. Figure 4 The demonstration showcased a multi-screen environment with four screens. For the screen directly in front of the observer, the camera faces forward; for the screen to the left, the camera faces left; for the screen to the right, the camera faces right; and for the ground screen below the observer, the camera faces downward. The front-facing camera serves as the primary visual guide, presenting the core 3D geographical scene content. The side cameras facing left and right are angled outward to simulate peripheral vision, enhancing the sense of immersion. The ground camera, facing downward, can be set vertically downward or slightly tilted depending on the scene's requirements, bridging the visual extension between the walls and the ground, providing the observer with a truly immersive experience of the scene beneath their feet. This deployment ensures that each virtual camera's field of view covers the entire area of ​​its corresponding physical screen, while maintaining seamless transitions between camera perspectives to create a panoramic view surrounding the observer, achieving a natural, naked-eye stereoscopic visual effect.

[0061] Step S400: Based on the physical size and spatial position of each planar display screen relative to the observer's viewpoint, perform physical trapezoidal combination configuration for each virtual camera and calculate the projection matrix of the corresponding asymmetric view frustum.

[0062] In this embodiment, the physical trapezoidal configuration is a method of asymmetrically adjusting the projection surface of the virtual camera based on the actual geometry and relative position of the physical screen. An asymmetrical view frustum refers to a view frustum structure where the near and far clipping planes are not centrally symmetrical rectangles, but rather offset according to the position of the physical screen and the observer's viewpoint. Unlike traditional symmetrical view frustum projection, this embodiment uses a physical trapezoidal configuration to configure the projection surface of each virtual camera into a trapezoidal shape that perfectly matches the corresponding physical screen.

[0063] When configuring the physical trapezoidal combination, the physical dimensions and coordinate positions in physical space of each planar display screen are first measured, including the three-dimensional coordinates of the four corner points of the screen. Then, based on this coordinate information, the spatial positional relationship of the screen relative to the observer's viewpoint and the projection boundary are calculated. Based on the calculated projection boundary, the projection matrix of the virtual camera is asymmetrically corrected to generate the corresponding asymmetric frustum projection matrix. This step solves the perspective distortion problem at the corners of multi-screen displays, ensuring that the lines of three-dimensional objects remain straight at the screen connections, achieving a naked-eye stereoscopic visual effect without the need for any additional equipment.

[0064] In one implementation, the step of physically configuring each virtual camera based on the physical size and spatial position of each planar display screen relative to the observer's viewpoint, and calculating the projection matrix of the corresponding asymmetric view frustum, specifically includes the following steps: Step S410: For any planar display screen, determine the projection boundary of the screen relative to the observer's viewpoint based on the coordinates of the four corner points of the planar display screen in physical space. Step S420: Configure the projection surface of the virtual camera corresponding to the flat display screen as a trapezoid that perfectly matches the projection boundary; Step S430: Obtain the vertical distance from the observer's viewpoint to the plane of the flat display screen, the near clipping plane distance of the virtual camera, and the horizontal and vertical offsets of the observer's viewpoint relative to the center of the flat display screen as raw parameters; Step S440: Based on the original parameters, calculate the left and right boundary parameters and the top and bottom boundary parameters of the near-cut surface, and generate the projection matrix of the asymmetric view frustum.

[0065] In this embodiment, for any planar display screen, the three-dimensional coordinates of the four corner points of the screen in physical space are first obtained. These coordinates are precisely measured using a laser rangefinder or a 3D scanner. Based on the coordinates of the four corner points, the plane equation of the plane containing the screen and the projection boundary of the screen relative to the observer's viewpoint are calculated. The projection boundary refers to the spatial region enclosed by rays originating from the observer's viewpoint and passing through the four corner points of the screen.

[0066] The projection surface of the virtual camera is configured as a trapezoid that perfectly matches the projection boundary. Traditional virtual camera projection surfaces are centrally symmetrical rectangles, while in this embodiment, the projection surface is an asymmetrical trapezoid, with its four vertices perfectly coinciding with the projection positions of the four corner points of the screen on the projection plane. This configuration ensures that the field of view of the virtual camera precisely corresponds to the actual display area of ​​the physical screen.

[0067] Obtain the vertical distance from the observer's viewpoint to the plane where the display screen is located. Near clipping distance of virtual camera And the offset of the observer's viewpoint relative to the center of the flat display screen. ,in For horizontal offset and This is the vertical offset. The horizontal distance from the observer's viewpoint to the left edge of the screen. The horizontal distance from the observer's viewpoint to the right edge of the screen. The vertical distance from the observer's viewpoint to the bottom edge of the screen. This is the vertical distance from the observer's viewpoint to the top edge of the screen.

[0068] Based on the above original parameters, calculate the left and right boundary parameters and the top and bottom boundary parameters of the near-cut surface using the following formula:

[0069]

[0070]

[0071]

[0072] Based on the calculated near-clipping plane boundary parameters, an asymmetric view frustum projection matrix is ​​generated. This projection matrix is ​​used to convert the object coordinates in the 3D scene into 2D pixel coordinates on the screen. Through the calculation of this asymmetric projection matrix, the system can accurately correct perspective distortion caused by screen folds and the observer's non-direct viewing angle, ensuring the correct display ratio and perspective relationship of 3D objects on each screen. In this embodiment, the projection matrix can be calculated inversely using nDisplay's asymmetric view frustums technology. nDisplay is a distributed multi-screen rendering plugin provided by Unreal Engine.

[0073] Step S500: Assign the rendering task of each virtual camera to the corresponding computing node of the pre-built distributed rendering cluster, and control the rendering screen of all virtual cameras to be synchronously output to the corresponding flat display screen at the same time through the hardware synchronization mechanism.

[0074] In this embodiment, the pre-built distributed rendering cluster is a parallel computing system composed of multiple high-performance computing nodes connected through a high-speed network, used to handle rendering tasks of large-scale 3D scenes. Each computing node is equipped with a high-performance graphics card, large-capacity memory, and high-speed storage devices, and has independent rendering output capabilities. The hardware synchronization mechanism achieves precise synchronization of the rendering images of multiple computing nodes through hardware devices, controlling the frame synchronization error between nodes to within milliseconds.

[0075] Specifically, the distributed rendering cluster can be deployed using a master-slave architecture, with nodes deployed based on load balancing. The master and slave computing nodes can be connected via both high-speed Ethernet and synchronization cables to ensure stable data transmission rates and synchronization signals. Figure 3 A distributed computing architecture for four screens was demonstrated, which is based on a local area network environment and uses a high-performance workstation cluster as hardware. Figure 3 The demonstration specifically showcases the node deployment architecture of the distributed rendering cluster, the master-slave node connection relationship, and the resource synchronization link, illustrating the division of labor and data interaction paths between the master node and each slave node. After node deployment, a static resource synchronization operation is performed. Static resources in the cluster include the 3D rendering engine installation package, preprocessed 3D geographic information data, virtual camera configuration files, and display parameter configurations. These static resources are uniformly managed and distributed by the master compute node. A file verification mechanism ensures that the resources obtained by each slave compute node are complete and consistent, avoiding rendering deviations due to resource differences. Furthermore, because the distributed launcher Switchboard uses absolute path references when issuing commands, any path deviation will cause node asset loading failures. Therefore, all node servers need to store project resources under completely consistent physical drive letters and paths, such as E:\CityRender\. Switchboard is a distributed cluster startup and management tool provided by Unreal Engine.

[0076] This embodiment distributes the rendering tasks corresponding to each virtual camera to different computing nodes in a distributed rendering cluster, with each node rendering its own image in parallel. After rendering is complete, a hardware synchronization mechanism controls all computing nodes to perform buffer swapping operations simultaneously, outputting the rendered image to the corresponding flat display screen. This distributed rendering architecture effectively alleviates the rendering pressure of ultra-large-scale resolution scenes, ensuring the system's rendering frame rate and image smoothness. At the same time, the hardware synchronization mechanism completely eliminates screen tearing and stuttering at multi-screen splicing points, ensuring visual consistency for immersive displays.

[0077] In one implementation, the step of assigning the rendering task of each virtual camera to the corresponding computing node of a pre-built distributed rendering cluster specifically includes the following steps: Step S510: For any virtual camera, based on the current position of the virtual camera and the coverage of the asymmetric view frustum, calculate the tile detail level and obtain the tile data of the corresponding level; Step S520: Perform visibility judgment on the tile data and remove tiles that are below the observer's horizon or obscured by other objects from the rendering task; Step S530: Overlay the white model of the building generated by the vector data onto the three-dimensional geographic scene, and assign display materials according to the vector attributes; Step S540: Based on the removed tile data and the white model of the building, the part corresponding to the virtual camera in the three-dimensional geographic scene is allocated to the corresponding computing node of the distributed rendering cluster for rendering.

[0078] In this embodiment, for any virtual camera, the required level of detail (LOD) for the tiles is first calculated based on the virtual camera's current position and the coverage area of ​​the asymmetric view frustum, thus achieving hierarchical level of detail (LOD) loading. The 3D geographic information data is tiled and hierarchically processed according to the 3D Tiles open standard, dividing the entire geographic scene into tile data of different levels of detail. Areas closer to the virtual camera are loaded with higher-level-of-detail tiles, while areas farther away are loaded with lower-level-of-detail tiles. This hierarchical level of detail loading strategy effectively controls memory usage and loading time while ensuring image accuracy.

[0079] Subsequently, the system requests and retrieves tile data of the corresponding level from the data server on demand. The data server can be a local storage server or a cloud data server, connected to the distributed rendering cluster via a high-speed network. The system adopts a streaming loading method, rendering while loading, and can display the scene without waiting for all data to be loaded, ensuring that buildings are detailed up close and the terrain is macroscopic in the distance, while keeping the video memory usage within a controllable range.

[0080] After acquiring the tile data, a horizon culling algorithm is used to assess its visibility, removing tiles that are below the observer's horizon or obscured by other objects from the rendering task. The horizon culling algorithm calculates the farthest horizon visible to the observer based on the virtual camera's position and the Earth's curvature; tiles below this horizon are automatically removed. The occlusion culling algorithm detects the spatial relationship between tiles and other objects, removing tiles obscured by mountains, buildings, etc. This culling operation significantly reduces the number of triangles that need to be rendered, improving the system's rendering frame rate.

[0081] Optionally, a white model of a building generated from vector data can be overlaid on the 3D geographic scene, and different display materials can be assigned according to the vector attributes. Specifically, the vector data can be building outline data in SHP (Shapefile) format or GeoJSON (Geographic JavaScript Object Notation) format. Based on the attribute information in the vector data, such as building height, construction year, and purpose, the system assigns different colors, transparency, or textures to the white model of the building to intuitively display different types of geographic information.

[0082] Based on the processed tile data and optional overlaid white models of buildings, the 3D geographic scene corresponding to the virtual camera is allocated to the corresponding computing nodes in the distributed rendering cluster for rendering. Each computing node is only responsible for rendering the scene data within the field of view of its corresponding virtual camera, further improving rendering efficiency.

[0083] In one implementation, the distributed rendering cluster includes at least one master computing node. The graphics cards of all computing nodes in the distributed rendering cluster are daisy-chained together via synchronization cables. The step of controlling the rendering images of all virtual cameras to be synchronously output to the corresponding flat display screens at the same time through a hardware synchronization mechanism specifically includes the following steps: Step S550: The graphics card of the main computing node sends out a synchronization pulse signal as the global clock source of the distributed rendering cluster. Step S560: Align the refresh rate of the graphics cards of all computing nodes with the vertical synchronization signal of the corresponding flat-panel display screen. Step S570: Configure the swap barrier strategy. After all computing nodes have completed all rendering tasks for the current frame, perform a buffer swap operation and output the rendered screen.

[0084] In this embodiment, the distributed rendering cluster includes at least one master compute node and multiple slave compute nodes, with all compute nodes' graphics cards daisy-chained together via synchronization cables. The master compute node acts as the cluster's control center, responsible for global scheduling and issuing synchronization commands, while the slave compute nodes execute specific rendering tasks. The graphics cards are professional graphics cards with hardware synchronization capabilities, specifically NVIDIA RTX series graphics cards or other graphics cards.

[0085] The master compute node's graphics card emits a synchronization pulse signal, serving as the global clock source for the entire distributed rendering cluster. All graphics cards on slave compute nodes synchronously receive this pulse signal, aligning their internal clocks with the global clock source. This global clock synchronization mechanism ensures that the rendering rhythm of all compute nodes remains consistent. When using NVIDIA RTX series graphics cards, the synchronization pulse signal is achieved by connecting the graphics cards of each node in series via a Quadro Sync II synchronization card and a physical synchronization cable.

[0086] Align the refresh rates of all graphics cards on the computing nodes with the vertical sync signal of the corresponding flat-panel display screen. Employing generator locking technology and introducing an external video lock signal, this forces the graphics card's output refresh rate to be perfectly synchronized with the monitor's vertical sync signal. This step eliminates screen tearing caused by frequency differences between the graphics card and the monitor.

[0087] Configure a buffer swapping barrier strategy so that buffer swapping operations are performed simultaneously on all compute nodes after all compute nodes have completed rendering of the current frame, and the rendered image is output. A swapping barrier is a synchronization primitive that blocks buffer swapping operations on all compute nodes until all nodes have finished rendering. Only after the master compute node receives rendering completion signals from all slave compute nodes will it send buffer swapping instructions to all nodes simultaneously. Upon receiving the instructions, all nodes immediately swap the contents of their background rendering buffers to their foreground display buffers and output the rendered image. For example, NVIDIA Mosaic multi-screen display technology can be used to recognize multiple physical outputs as a single logical canvas, and an nDisplay barrier strategy can be configured in Unreal Engine. Only when all nodes have completed rendering of the current frame will the master node allow all nodes to perform buffer swapping simultaneously, as expressed by the following formula:

[0088] in, Indicates the frame synchronization error time. This represents the single-frame rendering budget time. The formula expresses the frame synchronization error time between all computing nodes in the distributed rendering cluster. It must be strictly less than the total single-frame rendering budget time (approximately 16.67 milliseconds). This is a necessary condition to ensure that the system achieves smooth display at 60fps and that there is no screen tearing at the splicing points of multiple screens.

[0089] By using a hardware-level synchronization mechanism, the frame synchronization error between nodes is controlled within milliseconds, thus solving the problems of screen tearing and clipping at the point of multi-screen splicing.

[0090] In one implementation, after controlling the rendered images of all virtual cameras to be synchronously output to the corresponding flat display screen at the same time through the hardware synchronization mechanism, the following steps are further included: Step S610: Encode the rendered image of the three-dimensional geographic scene into a video stream, and push the video stream to the mobile control terminal; Step S620: Receive the operation command sent by the mobile control terminal and convert the operation command into a logical command that can be recognized by the 3D rendering engine; Step S630: Synchronously update the position, rotation angle, and scaling of all virtual cameras based on the logic instructions.

[0091] In this embodiment, the synchronously output 3D geographic scene rendering is encoded into a standard video stream, which can be a conventional video encoding format. The encoding process is accelerated by the graphics card hardware of the main computing node to reduce CPU utilization and improve encoding efficiency. The encoded video stream features low latency and high compression ratio, making it suitable for transmission over a network.

[0092] The video stream is pushed to the mobile control terminal based on a network transmission protocol, and a conventional network transmission protocol can be selected. This embodiment preferably uses the WebRTC (Web Real-Time Communication) protocol, which supports peer-to-peer real-time communication and achieves millisecond-level low-latency video transmission. The mobile control terminal can be a tablet computer, smartphone, laptop computer, virtual reality glasses, or other device with network connectivity and video decoding capabilities. The user accesses the service address of the main computing node through the browser on the mobile control terminal. The signaling server coordinates the peer-to-peer connection between the browser and the 3D rendering engine, and after completing the signaling handshake, establishes the video stream transmission channel and the data transmission channel. Figure 5 It demonstrates the complete interaction sequence from user-initiated operation to multi-screen synchronized output, including the entire timeline of command flow, data request, distributed rendering, and hardware synchronization.

[0093] The system receives touch or mouse operation commands from the mobile control terminal and converts these commands into logical instructions recognizable by the 3D rendering engine. The mobile control terminal receives user input via a touchscreen or mouse, such as swiping, clicking, zooming, and rotating. These commands are converted into JSON-formatted messages and sent back to the main computing node via the WebRTC data channel. (Touch coordinates on the mobile device are also mentioned.) The JS-SDK (JavaScript Software Development Kit) converts the data into logical instructions within the engine, and then calculates the logical instructions:

[0094] Among them, table Display logic instructions, This indicates a JS-SDK conversion operation. and These represent the width and height parameters of the corresponding screen.

[0095] After receiving the instruction, the main computing node parses it through the blueprint script system of the 3D rendering engine, mapping the external input instruction into the engine's internal logical instructions, such as camera movement, view rotation, scene scaling, and layer switching.

[0096] The position, rotation angle, and zoom level of all virtual cameras are updated synchronously based on logical instructions. All virtual cameras are configured with the observer's viewpoint as the origin, and updating the observer's viewpoint state synchronously updates the parameters of all virtual cameras. Based on the updated camera parameters, the asymmetric frustum projection matrix is ​​recalculated, triggering a new round of rendering. This multi-terminal interaction method retains high-performance rendering tasks on the server cluster, with mobile terminals serving only as display and control devices, reducing the requirements for terminal hardware while achieving low-latency interactive responses. For example, camera control classes and model display synchronization classes are implemented by encapsulating functional extension APIs. When a user drags the map on the terminal, the command is sent to the server via the JavaScript environment, triggering Blueprint events and synchronously updating the camera position, rotation angle, and zoom level on the large screen. This pixel streaming architecture is named Pixel Streaming.

[0097] Furthermore, it supports customized character-based scenes, distinguishing between audience scenes and actor immersion scenes. In audience scenes, the interface focuses on visualizing and statistically displaying macro data, while in actor immersion scenes, it supports first-person roaming, flight simulation, and manual annotation and interaction of key areas and buildings. In addition, it supports OSC (Open Sound Control) message communication and VRPN (Virtual Reality Peripheral Network) protocol integration. By defining message sending types through custom parameters, a global singleton receives events and triggers corresponding interface animations or level scene changes, enabling cross-device collaborative demonstrations.

[0098] In one implementation, the multi-sided screen is a four-sided screen, including a front screen, a left screen, a right screen and a bottom screen. The front screen, left screen and right screen are connected to each other in pairs, and the bottom screen is horizontally set at the bottom of the front screen, left screen and right screen, together enclosing a semi-enclosed display space. The four virtual cameras are facing directly forward, left, right, and down. The pre-built distributed rendering cluster includes one master computing node and at least three slave computing nodes. The master computing node is responsible for the rendering tasks of the virtual camera corresponding to the front screen and the issuance of global scheduling synchronization instructions. The slave computing nodes are respectively responsible for the rendering tasks of the virtual cameras corresponding to the left screen, right screen and ground screen.

[0099] In this embodiment, the multi-screen can be a four-screen structure, including a front screen, a left screen, a right screen, and a floor screen. The front screen, left screen, and right screen are arranged perpendicularly to each other in pairs, forming three perpendicular walls. The floor screen is horizontally set at the bottom of the front screen, left screen, and right screen, forming a horizontal floor. The four screens together enclose a semi-enclosed immersive display space, with the observer located at the center of this space.

[0100] The four virtual cameras face the front, left, right, and bottom directions, corresponding one-to-one with the four screens. The front camera faces the front screen and is responsible for rendering the main front view; the left camera faces the left screen and is responsible for rendering the left-side surround view; the right camera faces the right screen and is responsible for rendering the right-side surround view; and the ground camera faces the bottom screen and is responsible for rendering the ground-view overhead view.

[0101] like Figure 3 As shown, the pre-built distributed rendering cluster includes one master compute node and at least three slave compute nodes. The master compute node is responsible for the rendering tasks of the virtual camera corresponding to the front screen, and also undertakes the responsibilities of global scheduling and synchronous command issuance. The three slave compute nodes are responsible for the rendering tasks of the virtual cameras corresponding to the left screen, right screen, and ground screen, respectively. The rendering tasks of each screen and its corresponding node are shown in Table 1.

[0102] Table 1

[0103] Furthermore, multi-sided screens can also be five-sided screens including the ceiling or fully enclosed six-sided screens. Or, going further, for dome projection screens or curved LED arrays, the physical trapezoidal combination can be subdivided into multiple tiny triangular projection areas (Mesh Projection Policy). The viewport is mapped by creating a three-dimensional static mesh model of the physical screen in the engine. This physical mesh-based projection strategy is a geometric generalization of the trapezoidal configuration.

[0104] In summary, this embodiment utilizes distributed synchronization technology to achieve seamless and smooth four-screen splicing with a stable refresh rate exceeding 60fps, delivering visuals at a AAA game-level performance. It supports complex real-time lighting and physics effects, and accurately loads a realistic 3D city floor plan, enabling both city-wide macro-level displays and micro-level simulations of key areas. Furthermore, the system boasts strong openness, supporting the integration of various interactive sensing devices through API encapsulation, significantly enhancing the intuitiveness and scientific rigor of natural resource management.

[0105] Figure 2 This demonstrates the final implementation effect of this embodiment, showcasing excellent feasibility in actual engineering deployment and testing. This multi-screen display method achieved a 100% pass rate across 195 test cases, demonstrating particularly excellent performance in high-definition globe display and rapid viewpoint switching. To support the high load demands of natural resource simulation, the hardware configuration employs high-performance processors and large-capacity memory. Furthermore, at the network level, all node servers are required to be interconnected via at least gigabit fiber optic cables and assigned static IP addresses to ensure instantaneous response from the Switchboard when monitoring the status of each node.

[0106] Traditional 3D displays ignore the visual distortion caused by the physical angles of the screen. This embodiment proposes a physical trapezoidal combination configuration specifically adapted to multi-sided screens, and performs asymmetric offset correction on the projection matrix at the rendering pipeline level. The screen transition based on mathematical modeling ensures that the lines of 3D objects remain straight even at corners and cracks in the floor, eliminating visual blind spots. Thus, without the need for glasses, a strong sense of 3D depth is created by utilizing the perspective illusion of the human eye.

[0107] Furthermore, this embodiment resolves the inherent conflict between accuracy and realism in GIS systems and game engines. By employing a georeferencing plugin and 3D Tiles streaming technology, a realistic 3D model with accurate geographic coordinates is embedded into the UE4 environment, which boasts AAA-level rendering capabilities. This enables the simulation experiment to achieve millisecond-level synchronized visual smoothness and ensures that every operation and every measurement result is traceable and verifiable in the physical world.

[0108] Furthermore, this embodiment does not rely on a single software or hardware synchronization method. Instead, it constructs a three-layer defense system consisting of a hardware Genlock (Generator Lock), a graphics card swap barrier, and nDisplay application-layer synchronization. This combined approach solves the most challenging dynamic tearing problem in ultra-high resolution distributed rendering. Even in extreme test scenarios involving rapid flight over complex urban buildings, the four screens maintain a high degree of consistency. This industrial-grade stability is difficult for conventional display solutions to achieve.

[0109] In this embodiment, the multi-screen display also supports real-time lighting and shadow simulation. Based on the actual geographical location and time, the lighting system of Cesium (a 3D geospatial data visualization framework) is automatically updated to simulate the sun's trajectory. By overlaying post-processing volumes in UE4, color correction, screen space occlusion (SSAO), and fog effects can be applied to the real-world 3D data, ensuring that the final image not only possesses scientific accuracy but also artistic aesthetic value.

[0110] like Figure 6 As shown in the figure, this embodiment of the invention provides a three-dimensional geographic information system display system based on a multi-screen display. The system includes: a three-dimensional geographic scene construction module 10, an observer viewpoint determination module 20, a virtual camera deployment module 30, a physical trapezoidal combination calculation module 40, and a distributed rendering module 50.

[0111] Specifically, the 3D geographic scene construction module 10 is used to acquire 3D geographic information data and construct a 3D geographic scene in the 3D rendering engine; the observer viewpoint determination module 20 is used to determine the observer viewpoint in the physical space, wherein the physical space is provided with a multi-faceted screen, which is formed by at least three planar display screens that form an angle with each other; the virtual camera deployment module 30 is used to deploy the same number of virtual cameras as the multi-faceted screens in the 3D rendering engine, with the observer viewpoint as the origin, and each virtual camera faces the corresponding planar display screen; the physical trapezoidal combination calculation module 40 is used to perform physical trapezoidal combination configuration for each virtual camera based on the physical size and spatial position of each planar display screen relative to the observer viewpoint, and calculate the projection matrix of the corresponding asymmetric view frustum; the distributed rendering module 50 is used to allocate the rendering task of each virtual camera to the corresponding computing node of the pre-built distributed rendering cluster, and control the rendering screens of all virtual cameras to be synchronously output to the corresponding planar display screens at the same time through a hardware synchronization mechanism.

[0112] Based on the above embodiments, the present invention also provides a terminal device, the principle block diagram of which can be as follows: Figure 7As shown, the terminal device includes a processor, memory, network interface, display screen, and temperature sensor connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a method for displaying a multi-screen 3D geographic information system. The display screen can be an LCD screen or an e-ink screen. The temperature sensor is pre-installed inside the terminal device to detect the operating temperature of the internal components.

[0113] Those skilled in the art will understand that Figure 7 The schematic diagram shown is only a partial structural diagram related to the present invention and does not constitute a limitation on the terminal device to which the present invention is applied. The specific terminal device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0114] In one embodiment, a terminal device is provided, including a memory and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by one or more processors, the one or more programs including instructions for performing operations as described in the embodiments of the methods above.

[0115] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0116] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0117] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for displaying a three-dimensional geographic information system based on a multi-screen display, characterized in that, The method includes: Acquire 3D geographic information data and construct 3D geographic scenes in a 3D rendering engine; Determine the observer's viewpoint within a physical space, wherein a multi-sided screen is provided within the physical space, and the multi-sided screen is formed by at least three planar display screens that form an angle with each other; In the 3D rendering engine, virtual cameras are deployed with the observer's viewpoint as the origin, and the number of virtual cameras is the same as the number of multi-faceted screens. Each virtual camera faces the corresponding flat display screen. Based on the physical size and spatial position of each planar display screen relative to the observer's viewpoint, a physical trapezoidal combination configuration is performed for each virtual camera, and the projection matrix of the corresponding asymmetric view frustum is calculated. The rendering task of each virtual camera is assigned to the corresponding computing node of the pre-built distributed rendering cluster, and the rendering images of all virtual cameras are synchronously output to the corresponding flat display screen at the same time through a hardware synchronization mechanism. Based on the physical dimensions and spatial positions of each planar display screen relative to the observer's viewpoint, a physical trapezoidal combination configuration is performed for each virtual camera, and the projection matrix of the corresponding asymmetric view frustum is calculated, including: For any flat display screen, the projection boundary of the screen relative to the observer's viewpoint is determined based on the coordinates of the four corner points of the flat display screen in physical space. Configure the projection surface of the virtual camera corresponding to the flat display screen as a trapezoid that perfectly matches the projection boundary; The vertical distance from the observer's viewpoint to the plane of the flat display screen, the near clipping plane distance of the virtual camera, and the horizontal and vertical offsets of the observer's viewpoint relative to the center of the flat display screen are obtained as raw parameters. Based on the original parameters, the left and right boundary parameters and the top and bottom boundary parameters of the near-cut surface are calculated, and the projection matrix of the asymmetric view frustum is generated.

2. The method for displaying a three-dimensional geographic information system based on a multi-screen display according to claim 1, characterized in that, The acquisition of 3D geographic information data and the construction of a 3D geographic scene in a 3D rendering engine include: Obtain raw geospatial data in a geodetic coordinate system, wherein the raw geospatial data includes oblique photogrammetry model, terrain data, and image data; Convert the latitude and longitude coordinates of the original geospatial data into geocentric rectangular coordinates; The geocentric rectangular coordinates are converted into local tangent plane coordinates with the physical location of the observer's viewpoint as the origin, thus constructing the mapping of 3D geographic information data in the 3D rendering engine.

3. The method for displaying a three-dimensional geographic information system based on a multi-screen display according to claim 1, characterized in that, Determining the observer's viewpoint within the physical space includes: A predetermined fixed position in physical space is used as the observer's viewpoint; Alternatively, visual capture devices can be used to obtain the real-time position of the observer's head or eyeballs in the physical space as the observer's viewpoint.

4. The method for displaying a three-dimensional geographic information system based on a multi-screen display according to claim 1, characterized in that, The process of allocating rendering tasks for each virtual camera to the corresponding computing nodes of a pre-built distributed rendering cluster includes: For any virtual camera, based on the current position of the virtual camera and the coverage of the asymmetric view frustum, calculate the tile detail level and obtain the tile data of the corresponding level; The visibility of the tile data is determined, and tiles that are below the observer's horizon or are obscured by other objects are removed from the rendering task. The white models of buildings generated from vector data are overlaid on the three-dimensional geographic scene, and display materials are assigned according to the vector attributes; Based on the removed tile data and the white model of the building, the portion of the three-dimensional geographic scene corresponding to the virtual camera is allocated to the corresponding computing nodes of the distributed rendering cluster for rendering.

5. The method for displaying a three-dimensional geographic information system based on a multi-screen display according to claim 1, characterized in that, The distributed rendering cluster includes at least one master computing node. The graphics cards of all computing nodes in the distributed rendering cluster are daisy-chained together via synchronization cables. The hardware synchronization mechanism controls the rendering images from all virtual cameras to be synchronously output to their corresponding flat-panel display screens at the same time, including: The graphics card of the master computing node emits a synchronization pulse signal as the global clock source for the distributed rendering cluster. Align the refresh rate of the graphics cards on all computing nodes with the vertical synchronization signal of the corresponding flat-panel display screen; Configure a swap barrier strategy so that after all computing nodes have completed all rendering tasks for the current frame, they simultaneously perform a buffer swap operation and output the rendered screen.

6. The method for displaying a three-dimensional geographic information system based on a multi-screen display according to claim 1, characterized in that, After controlling all virtual camera renderings to be synchronously output to their corresponding flat-panel display screens at the same time via a hardware synchronization mechanism, the method further includes: The rendered image of the 3D geographic scene is encoded into a video stream, and the video stream is pushed to the mobile control terminal. Receive operation instructions sent by the mobile control terminal and convert the operation instructions into logical instructions that can be recognized by the 3D rendering engine; Based on the aforementioned logical instructions, the position, rotation angle, and scaling of all virtual cameras are updated synchronously.

7. The method for displaying a three-dimensional geographic information system based on a multi-screen display according to claim 1, characterized in that, The multi-sided screen is a four-sided screen, including a front screen, a left screen, a right screen and a floor screen. The front screen, left screen and right screen are connected to each other in pairs. The floor screen is horizontally set at the bottom of the front screen, left screen and right screen, together forming a semi-enclosed display space. The four virtual cameras are facing directly forward, left, right, and down. The pre-built distributed rendering cluster includes one master computing node and at least three slave computing nodes. The master computing node is responsible for the rendering tasks of the virtual camera corresponding to the front screen and the issuance of global scheduling synchronization instructions. The slave computing nodes are respectively responsible for the rendering tasks of the virtual cameras corresponding to the left screen, right screen and ground screen.

8. A three-dimensional geographic information system display system based on multi-screen displays, characterized in that, The steps for implementing the multi-screen-based 3D geographic information system display method as described in any one of claims 1-7, wherein the system comprises: The 3D geographic scene construction module is used to acquire 3D geographic information data and construct 3D geographic scenes in the 3D rendering engine. An observer viewpoint determination module is used to determine the observer viewpoint within a physical space, wherein a multi-sided screen is provided within the physical space, and the multi-sided screen is formed by at least three planar display screens that form an angle with each other. The virtual camera deployment module is used in the 3D rendering engine to deploy the same number of virtual cameras as the multi-faceted screen, with the observer's viewpoint as the origin, and each virtual camera facing the corresponding planar display screen; The physical trapezoidal combination calculation module is used to configure the physical trapezoidal combination of each virtual camera based on the physical size and spatial position of each planar display screen relative to the observer's viewpoint, and calculate the projection matrix of the corresponding asymmetric view frustum. The distributed rendering module is used to allocate the rendering tasks of each virtual camera to the corresponding computing nodes of the pre-built distributed rendering cluster, and control the rendering images of all virtual cameras to be synchronously output to the corresponding flat display screen at the same time through a hardware synchronization mechanism.

9. A terminal device, characterized in that, The terminal device includes a memory, a processor, and a multi-screen-based 3D geographic information system display program stored in the memory and executable on the processor. When the processor executes the multi-screen-based 3D geographic information system display program, it implements the steps of the multi-screen-based 3D geographic information system display method as described in any one of claims 1-7.

Images