Data visualization display method and system based on business model

Abstract

The embodiment of the present invention discloses a data visualization display method and system based on a business model, which belongs to the field of digital city technology. The embodiment of the present invention can effectively process three-dimensional surveying and mapping data containing multiple types of data such as terrain, land objects and textures, and mine visual element vectors through a first model visual mapping network to provide accurate input for subsequent rendering and improve visualization efficiency. The model visual rendering network performs rendering operations based on the mined vectors to achieve visual display of the target digital city area. Based on the first training error and the second training error, the two networks are jointly debugged to accurately optimize network performance, improve the accuracy and realism of the visualization effect, better present the spatial characteristics of the digital city, and meet various needs such as urban planning and management.

Description

Technical Field

[0001] The embodiments of the present invention belong to the field of digital city technology, and specifically relate to a data visualization display method and system based on a business model. Background Art

[0002] With the development of digital cities, the demand for 3D model visualization is growing. Traditional technologies face numerous challenges when processing 3D surveying and mapping data. For one thing, it's difficult to effectively integrate and utilize complex terrain, objects, and texture data, resulting in inefficient processing of the underlying data for visualization. Furthermore, the lack of effective mapping and rendering mechanisms during the data-to-visualization conversion process often results in inaccurate and inauthentic visualizations. Furthermore, the lack of a debugging method that can precisely optimize the relevant networks to enhance visualizations makes it impossible to meet the demand for precise visualization of digital cities in fields such as urban planning and management. Summary of the Invention

[0003] The embodiments of the present invention provide a data visualization display method and system based on a business model, which can solve or partially solve the technical problems involved in the above-mentioned background technology.

[0004] An embodiment of the present invention provides a data visualization display method based on a business model, which is applied to a data visualization display system. The method includes: obtaining three-dimensional surveying and mapping data to be processed, the three-dimensional surveying and mapping data to be processed is used to describe a multimodal spatial representation vector of a target digital city area, and the three-dimensional surveying and mapping data to be processed includes terrain data, object data, and texture data; performing visual element vector mining on the three-dimensional surveying and mapping data to be processed through a first model visual mapping network to obtain visual element vectors of the three-dimensional surveying and mapping data to be processed; performing visual rendering operations on the visual element vectors of the three-dimensional surveying and mapping data to be processed through a model visual rendering network to obtain a target visual rendering output result; the target visual rendering output result is used to realize model visualization display of the target digital city area; wherein the model visual rendering network and the first model visual mapping network are obtained by joint debugging based on a first training error and a second training error.

[0005] An embodiment of the present invention provides a data visualization display system, comprising at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor performs the above-mentioned method.

[0006] An embodiment of the present invention provides a readable storage medium, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the above method are implemented.

[0007] Embodiments of the present invention can effectively process three-dimensional mapping data containing multiple types of data, including terrain, objects, and textures. Through a first model visual mapping network, visual element vectors are mined, providing precise input for subsequent rendering and improving visualization efficiency. The model visual rendering network performs rendering operations based on the mined vectors, achieving visual display of the target digital city area. By jointly debugging the two networks based on the first training error and the second training error, network performance can be precisely optimized, improving the accuracy and fidelity of the visualization effect, better presenting the spatial characteristics of the digital city, and meeting various needs such as urban planning and management. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 This is a flowchart of a data visualization display method based on a business model provided by an embodiment of the present invention.

[0009] Figure 2 A structural diagram of a data visualization display system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0010] Figure 1 A data visualization display method based on a business model is shown, which is applied to a data visualization display system. The method includes the following steps 101 to 103.

[0011] Step 101: Acquire three-dimensional mapping data to be processed, wherein the three-dimensional mapping data to be processed is used to describe a multi-modal spatial representation vector of a target digital city area, and the three-dimensional mapping data to be processed includes terrain data, object data, and texture data.

[0012] Step 102: mining visual element vectors of the three-dimensional surveying and mapping data to be processed by a first model visual mapping network to obtain visual element vectors of the three-dimensional surveying and mapping data to be processed.

[0013] Step 103: Performing a visual rendering operation on the visual element vectors of the to-be-processed three-dimensional surveying and mapping data through a model visual rendering network to obtain a target visual rendering output result; the target visual rendering output result is used to realize model visualization display of the target digital city area.

[0014] The model visual rendering network and the first model visual mapping network are obtained by joint debugging based on the first training error and the second training error.

[0015] Based on the above steps 101 to 103, the debugging steps of the first model visual mapping network and the model visual rendering network include: obtaining X visual rendering output samples and Y three-dimensional mapping data samples included in the current debugging sample set, and performing visual element vector mining on each three-dimensional mapping data sample through the first model visual mapping network to obtain a first visual element vector of each three-dimensional mapping data sample, and performing visual element vector mining on each visual rendering output sample to obtain a first visual sample vector of each visual rendering output sample, wherein the three-dimensional mapping data sample is used to describe the multimodal spatial representation vector of the sample digital city area, and the X and Y are both positive integers; determining a first training error based on the first visual element vectors of the Y three-dimensional surveying and mapping data samples and the first visual sample vectors of the X visual rendering output samples; performing a visual rendering operation on the first visual element vectors of the Y three-dimensional surveying and mapping data samples using a model visual rendering network to obtain visual rendering training results corresponding to the Y three-dimensional surveying and mapping data samples, and determining a second training error based on the visual rendering training results corresponding to the Y three-dimensional surveying and mapping data samples; and jointly debugging the first model visual mapping network and the model visual rendering network based on the first training error and the second training error.

[0016] The data visualization and display system, as the execution entity of an embodiment of the present invention, first obtains unprocessed 3D mapping data in step 101. This unprocessed 3D mapping data is an important source of information for describing the multimodal spatial representation vector of the target digital city area, encompassing various aspects of data, including terrain data, object data, and texture data. Terrain data accurately represents the topographical conditions of the target digital city area. For example, the height of a hill in the city can be represented by a specific numerical value, including the slope and aspect of a hill ranging from 100 to 500 meters above sea level. Object data includes information on the 3D shape, size, and location of various man-made structures in the city, such as buildings. For example, a commercial center building may occupy tens of thousands of square meters and be dozens of stories tall. Road data includes information on its width and direction. For example, the width of a main road in a city may range from 50 to 100 meters, while the width of a branch road may range from 10 to 20 meters. Texture data gives the ground features a realistic appearance. For example, the material texture of a building wall may be marble or brick, and the road surface texture may be asphalt or cement.

[0017] Next, the process proceeds to step 102, where the data visualization display system utilizes the first model visual mapping network to mine visual element vectors for the three-dimensional mapping data to be processed. Visual element vector mining is a complex and critical process. Through the first model visual mapping network, visual element vectors can be extracted from the three-dimensional mapping data to be processed, which contains a variety of data such as terrain, objects, and textures. Taking buildings in a city as an example, the first model visual mapping network can mine element vectors related to visualization from comprehensive information such as the building's three-dimensional shape data and its surface texture data. These vectors may contain key features of the building in terms of visual presentation, such as the building's outline feature vector and color feature vector at a specific viewing angle. Through this mining process, the original complex three-dimensional mapping data can be converted into a visual element vector form that is more convenient for subsequent visualization processing, thereby laying the foundation for achieving model visualization display of the target digital city area.

[0018] In step 103, the data visualization display system uses the model visual rendering network to perform visual rendering operations on the visual element vectors of the previously obtained three-dimensional surveying and mapping data to be processed. The visual rendering operation is a key step in converting the mined visual element vectors into target visual rendering output results that can be directly presented on the display device. In this process, the model visual rendering network will perform operations such as light and shadow effect calculation, color adjustment, and spatial layout optimization based on various feature information in the visual element vectors. For example, for a park area in a city, realistic green grass and blue lakes are rendered based on the visual element vectors of the grass and lakes in its terrain data, and the shadow effects on the grass and lakes are calculated based on the direction of light, and the lake surface is rendered based on the reflected light from the surrounding buildings, making the lake surface look shimmering. The target visual rendering output result finally obtained can realize the model visualization display of the target digital city area, allowing users to intuitively see the three-dimensional model of the digital city area.

[0019] The debugging process of the first model visual mapping network and the model visual rendering network is an important step in ensuring the accuracy and effectiveness of the entire visualization system. During the debugging process, the data visualization display system first obtains the current debugging sample set. This debugging sample set contains X visual rendering output samples and Y three-dimensional mapping data samples, where X and Y are both positive integers. For example, X can take the value of 100 and Y can take the value of 50. For each three-dimensional mapping data sample, it is used to describe the multimodal spatial representation vector of the sample digital city area, just like the three-dimensional mapping data of the target digital city area mentioned above, containing multiple data such as terrain, objects, and textures.

[0020] The data visualization display system mines visual element vectors for each 3D surveying and mapping data sample using the first model's visual mapping network, obtaining a first visual element vector for each 3D surveying and mapping data sample. Simultaneously, visual element vector mining is also performed on each visual rendering output sample, obtaining a first visual sample vector for each visual rendering output sample. For example, for a visual rendering output sample representing a visualization of a historical block in a city, the mined first visual sample vector may include a color distribution feature vector for the block's buildings, an architectural style feature vector, and so on.

[0021] A first training error is determined based on the first visual element vectors of the Y 3D mapping data samples and the first visual example vectors of the X visual rendering output samples. Determining this first training error may involve measuring the difference between the two vectors, such as comparing the degree of difference between the two vectors in key visual element features. If there is a significant deviation between the first visual element vector of the 3D mapping data sample and the first visual example vector of the visual rendering output sample in a certain visual element feature, such as the representation of building height, this will be reflected in the first training error.

[0022] Then, the data visualization display system performs a visual rendering operation on the first visual element vectors of the Y three-dimensional mapping data samples through the model visual rendering network to obtain visual rendering training results corresponding to the Y three-dimensional mapping data samples. For example, for a three-dimensional mapping data sample that is an area containing a large bridge, the visual rendering training result obtained after the visual rendering operation is the visual presentation of the bridge area after rendering, including the color, material texture, and integration effect with the surrounding environment of the bridge. Based on the visual rendering training results corresponding to these Y three-dimensional mapping data samples, a second training error is determined. The determination of this second training error may be to compare the difference between the visual rendering training result and the expected result. For example, if the color of the bridge in the visual rendering training result deviates significantly from the actual color, the value of the second training error will be increased.

[0023] Finally, based on the first training error and the second training error, the first model visual mapping network and the model visual rendering network are jointly debugged. This joint debugging enables the two networks to coordinate and optimize each other. For example, if the first training error is mainly caused by inaccurate feature extraction of a certain type of land object during the visual element vector mining process, then the relevant parameters in the first model visual mapping network can be adjusted; if the second training error is caused by improper processing of light and shadow effects during the visual rendering operation, then the light and shadow calculation module in the model visual rendering network can be adjusted. Through this continuous joint debugging process, the performance of the entire visualization system is gradually improved, so that the final target visual rendering output result can more accurately and realistically realize the model visualization display of the target digital city area.

[0024] Throughout the implementation of this invention, the data visualization system continuously optimizes and improves the model visualization display of the target digital city area through the orderly execution of various steps, from data acquisition, visual element vector mining, visual rendering operations, to network debugging, providing users with a high-quality digital city 3D model visualization experience. This process involves the precise processing of multiple data, the construction and optimization of complex networks, and the close coordination of various links, thus achieving the transformation from raw 3D mapping data to high-quality visualization.

[0025] It can be understood that the data visualization display system is the core execution body of the embodiment of the present invention, in which the first model visual mapping network, the model visual rendering network and the joint debugging process of the two constitute the key part of the invention.

[0026] Throughout the invention, the first model visual mapping network undertakes the crucial task of mining visual element vectors from the 3D mapping data to be processed. This network is capable of performing in-depth analysis of the 3D mapping data to be processed, which includes terrain data, object data, and texture data. For example, a certain area within a city contains complex terrain, such as mountains with altitudes ranging from 50 to 300 meters, numerous different types of objects, including buildings ranging from 10 to 50 stories high, and roads of varying widths and orientations. The multimodal spatial representation vector of the target digital city area described by the 3D mapping data to be processed encompasses comprehensive information about these terrain, objects, and textures. For example, a multimodal spatial representation vector for a building might include numerical features such as its 3D coordinates (x, y, z), building height, and building area; for a road, it might include numerical features such as its start and end coordinates and road width; and for the terrain, it might include elevation values at different locations. These multimodal spatial representation vectors provide a naive and comprehensive description of the spatial information of the urban area.

[0027] The first model, the visual mapping network, processes this complex multimodal spatial representation vector through its own structure and algorithm to mine visual element vectors. A visual element vector is a vector representation that is important for visualization. For example, for a specific building, the visual element vector may include a feature vector of the building's outline when viewed from a certain perspective, whose numerical value can represent the shape parameters of the outline, such as the curvature of the corner; it may also include a feature vector of the color distribution on the building's surface, which numerically represents the proportion of different color areas; and a feature vector of the relative position relationship between the building and its surrounding environment, which numerically represents the distance ratio to adjacent roads and other buildings.

[0028] The model visual rendering network is responsible for performing visual rendering operations on the visual feature vectors mined by the first model visual mapping network. This network performs a series of rendering operations based on the various feature values in the visual feature vectors. For example, for the building outline feature vector in the visual feature vector, the model visual rendering network determines the building's shape in the visualization based on the outline's shape parameters. For the color distribution feature vector, it accurately assigns colors to the building surface based on the proportion of different color regions. For the relative position relationship feature vector, it rationally arranges the building's position within the overall urban area visualization scene. During the rendering process, the model visual rendering network also considers factors such as lighting effects and material texture to produce more realistic visualizations. For example, it calculates the position and size of shadows based on the sun's direction and the building's positional relationship, and determines the surface material texture based on the building's function (e.g., a residential building may have a softer texture, while a commercial building may have a more modern texture). This results in the target visual rendering output. The target visual rendering output result is the visual image or scene finally presented to the user. It can accurately and realistically display the three-dimensional model of the target digital city area, allowing users to intuitively observe the topography, landforms, and distribution of objects in the urban area.

[0029] The joint debugging process of these two networks is key to ensuring the accuracy and effectiveness of the entire visualization system. During the joint debugging process, a debugging sample set is first obtained, which contains X visual rendering output samples and Y 3D mapping data samples (for example, X=80, Y=60). For each 3D mapping data sample (these samples also contain the multimodal spatial representation vector information mentioned above, but only for the sample digital city area), the first model visual mapping network performs visual element vector mining to obtain the first visual element vector. Simultaneously, visual element vector mining is also performed on each visual rendering output sample to obtain the first visual sample vector.

[0030] For example, for a building in a 3D surveying data sample, the building height in its multimodal spatial representation vector is 100 meters, and the building area is 5,000 square meters. The contour feature vector value in the first visual element vector mined by the first model visual mapping network indicates that the curvature of a corner of the contour is 0.3, and the color distribution feature vector indicates that a certain color accounts for 30%. For a similar building in a visual rendering output sample, the contour feature vector value in the mined first visual sample vector indicates that the corner curvature is 0.35, and the color distribution feature vector indicates that the color accounts for 28%. A first training error is determined by comparing the first visual element vectors of these Y 3D surveying data samples with the first visual sample vectors of X visual rendering output samples. The determination of the first training error may be based on various difference metrics, such as calculating the sum of squared differences in contour feature vector values or the sum of absolute differences in color distribution feature vector values. The first training error value is then derived based on these calculation results.

[0031] Next, the model visual rendering network performs a visual rendering operation on the first visual element vectors of Y three-dimensional mapping data samples to obtain visual rendering training results. For example, for the building outline feature vector and color distribution feature vector in the visual element vector of one of the three-dimensional mapping data samples, the shape and color of the building rendered by the model visual rendering network may be different from the expected ones. The second training error is determined based on the difference between these visual rendering training results and the expected results. For example, if the rendered building height appears to be 5 meters shorter than the actual 100 meters, or the color deviates from the expected color to a certain extent, these will be reflected in the calculation of the second training error.

[0032] Finally, based on the first and second training errors, the first model visual mapping network and the model visual rendering network are jointly debugged. If the first training error is primarily due to inaccurate contour feature vector mining (e.g., corner curvature errors), the parameters related to contour feature mining in the first model visual mapping network are adjusted. If the second training error is due to improper color processing during the rendering process (e.g., deviations in color proportions), the modules related to color rendering in the model visual rendering network are adjusted. Through this joint debugging process, the performance of the two networks is continuously optimized, enabling the entire visualization system to more accurately obtain high-quality target visual rendering output results from multimodal spatial representation vectors through visual feature vector mining and visual rendering operations, achieving precise visualization of the target digital city area.

[0033] Based on the above content, it can be seen that the first training error and the second training error play a vital role. They are key indicators for measuring and optimizing the performance of the first model visual mapping network and the model visual rendering network.

[0034] The first training error is determined based on the first visual element vector of the 3D mapping data sample and the first visual sample vector of the visual rendering output sample. For each 3D mapping data sample, which contains a multimodal spatial representation vector describing the sample digital city area, the first model visual mapping network mines its visual element vectors to obtain the first visual element vector. For example, for a 3D mapping data sample containing a building, the first visual element vector may include the building's shape feature vector (e.g., a series of numerical values representing the complexity and symmetry of the building's outline) and a texture feature vector (e.g., a numerical value representing the texture's roughness or the color distribution ratio of the texture). For the visual rendering output sample, the first visual sample vector is obtained by mining the visual element vectors of the first model visual mapping network. Calculating the first training error may involve various methods to measure the difference between the two sets of vectors. For example, for the shape feature vector, the sum of the squared differences between the corresponding values can be calculated; for the texture feature vector, the sum of the absolute differences between the color distribution ratio values can be calculated. Combining the differences between these different components yields the first training error.

[0035] In traditional data visualization technology, there is often a lack of effective evaluation of the accuracy of the mining process from raw data to visual element vectors. The embodiment of the present invention introduces a first training error, which can accurately quantify the deviation between the first model visual mapping network and the ideal result when mining visual element vectors. This allows the system to optimize the first model visual mapping network in a targeted manner. For example, if the first training error shows a large deviation in shape feature vectors, it may mean that the first model visual mapping network has deficiencies in the mining algorithm for shape information such as building outlines, and it is necessary to adjust the relevant parameters in the network or improve the algorithm. This error quantification of the mining process is an innovative idea that helps to improve the accuracy of visual element vector mining, thereby laying the foundation for subsequent high-quality visual rendering.

[0036] The second training error is determined based on the visual rendering training results obtained by the model visual rendering network performing visual rendering operations on the first visual element vector of the 3D mapping data sample. When the model visual rendering network renders the first visual element vector, it generates a visual rendering training result, which is a visual representation of the sample digital city area. For example, for the first visual element vector of a 3D mapping data sample containing a road and surrounding buildings, the visual rendering training result will display the road's width, color, and material texture, as well as the building's height, color, and relative position relative to the road, as presented in the visual scene. The second training error is calculated by comparing the visual rendering training results with the expected results. The expected results can be based on known accurate information or a pre-defined ideal visualization effect. For example, if the visual rendering training results show that the building height is a certain value shorter than it should be, or if the road color deviates significantly from the expected color, these differences will be quantified in the second training error. This difference can be calculated in various ways, such as calculating the sum of the squares of the differences between each visual element (such as building height or color) and the expected value.

[0037] In previous data visualization technologies, the evaluation of rendering results is often rough and lacks precise quantitative error analysis. The second training error in the embodiment of the present invention can carefully measure the gap between the rendering effect of the model visual rendering network and the expectation. This allows the system to adjust the relevant parameters in the model visual rendering network, such as the light and shadow calculation module, the color allocation module, etc., in a targeted manner based on the feedback of the second training error. For example, if the second training error shows that there are major problems with the light and shadow effects of the building, the part of the model visual rendering network related to light and shadow calculation can be adjusted to improve the accuracy and realism of the rendering. This method of performing precise quantitative error analysis on the rendering results and using it to optimize the rendering network is the innovation of the embodiment of the present invention, which helps to improve the rendering quality of the entire visualization system and make the final target visual rendering output result more in line with actual needs.

[0038] Furthermore, the embodiments of the present invention achieve joint debugging of the first model visual mapping network and the model visual rendering network by simultaneously introducing the first training error and the second training error. This joint debugging is an innovative approach. Previous technologies often optimized either the mapping or the rendering process separately, making it difficult to achieve collaborative optimization of the entire visualization process. For example, in traditional technologies, one may only focus on optimizing the rendering results while ignoring the impact of the process from raw data to visual feature vector mining on the final visualization effect, or vice versa.

[0039] By adjusting the first model's visual mapping network based on the first training error and the model's visual rendering network based on the second training error, and by allowing the two networks to interact and optimize collaboratively during a joint debugging process, the accuracy and quality of the entire visualization system, from raw 3D mapping data to the final target visual rendering output, can be significantly improved. This joint debugging concept is a new exploration in the field of data visualization and provides a more effective solution for improving the visualization of models in digital city areas.

[0040] The introduction of two training errors helps to comprehensively improve the performance of the visualization system. From the perspective of the data processing process, the first training error ensures the accuracy of visual element vector mining and provides high-quality input for subsequent rendering; the second training error ensures the accuracy and realism of the rendering results. The combination of the two enables the entire visualization system to more accurately present the three-dimensional model of the target digital city area. For example, in application scenarios such as digital city planning and management, it can provide users with more accurate and intuitive urban spatial information, which is an effect that is difficult to achieve with traditional data visualization technology, reflecting the innovation of the embodiments of the present invention in improving visualization quality.

[0041] In an optional embodiment, determining the first training error based on the first visual element vectors of the Y three-dimensional mapping data samples and the first visual sample vectors of the X visual rendering output samples includes: determining a first visual vector chain and a second visual vector chain, where the first visual vector chain includes U second visual sample vectors, and the second visual vector chain includes V second visual element vectors, where U and V are both positive integers; and performing joint training based on the first visual element vectors of the Y three-dimensional mapping data samples, the first visual sample vectors of the X visual rendering output samples, the U second visual sample vectors included in the first visual vector chain, and the V second visual element vectors included in the second visual vector chain to determine the first training error.

[0042] In this embodiment, the data visualization display system determines the first training error by first determining a first visual vector chain and a second visual vector chain, wherein the first visual vector chain includes U second visual sample vectors and the second visual vector chain includes V second visual element vectors, where U and V are both positive integers.

[0043] The second visual sample vectors in the first visual vector chain are further derived or selected from the first visual sample vector of the visual rendering output sample. For example, the first visual sample vector of the visual rendering output sample is a vector representation of the visualization of a digital city area, which may include visual feature vectors of various aspects such as buildings and roads. The second visual sample vector selected from this first visual sample vector may focus more on specific visual elements, such as the appearance feature vector of a building. For example, the appearance feature vector of a building in the first visual sample vector may include a color distribution vector (numerically representing the proportion of different colors, such as red 30%, blue 20%) and a shape feature vector (numerically representing the curvature of the outline, such as a corner curvature of 0.3). The second visual sample vector selected may be further refined to a feature vector of a specific decorative component of the building's facade, such as a decorative texture with a roughness value of 0.5.

[0044] Similarly, the V second visual element vectors in the second visual vector chain are derived or selected from the first visual element vectors of the 3D mapping data sample. Taking a building in the 3D mapping data sample as an example, the first visual element vector may include the building's basic structural feature vector (such as height and number of floors, for example, 100 meters in height and 30 floors), and its geographic location feature vector (using coordinate values to represent its location in the city, such as x=100, y=200, etc.). The second visual element vector, on the other hand, may focus more on visual features related to the building's internal structure, such as the spatial proportion vector of the internal room layout (for example, the aspect ratio of a room is 2:1).

[0045] Then, a joint training is performed based on the first visual element vectors of the Y 3D mapping data samples, the first visual sample vectors of the X visual rendering output samples, the U second visual sample vectors included in the first visual vector chain, and the V second visual element vectors included in the second visual vector chain to determine the first training error. This joint training process involves in-depth exploration and quantification of the relationships between multiple sets of vectors.

[0046] For example, for a 3D mapping data sample of a digital city area containing multiple buildings and roads, the building height vector value in the first visual element vector is 100 meters, while the corresponding building height value in the first visual sample vector of the visual rendering output sample appears to be 95 meters. Simultaneously, a second visual sample vector in the first visual vector chain may represent the visual display effect value of the building's top decoration (e.g., the display completeness of the decoration is 80%), while a second visual element vector in the second visual vector chain may represent the original value of the building's top structure in the 3D mapping data (e.g., the complexity of the top structure is 0.6).

[0047] During joint training, the relationships between these different layers of vectors must be comprehensively considered to determine the first training error. Operations such as weighted summation can be performed on the numerical differences between different vectors. For example, differences in building height might be given a higher weight due to their significant impact on the overall visualization. Differences in the display quality of top decorations and the complexity of the top structure could be given different weights based on their relative importance to the overall visualization. This comprehensive weighted summation or other quantitative operations, which consider the relationships between multiple groups of vectors, ultimately determines the first training error.

[0048] This method for determining the first training error offers significant innovation and advantages over traditional methods. Traditional methods for determining the training error may simply compare partial vectors or employ single-level vector comparisons, making it difficult to fully and accurately reflect the complex relationship between 3D mapping data and visual rendering output. However, the method in this embodiment, by constructing multiple vector chains and conducting joint training, can more deeply explore the connections and differences between vectors at different levels, thereby more accurately determining the first training error.

[0049] As for the second training error, it has been mentioned above that it is determined based on the visual rendering training result obtained by performing a visual rendering operation on the first visual element vector of the three-dimensional mapping data sample by the model visual rendering network. It can be further seen here that the determination method involved in the embodiment of the present invention of the first training error is of great significance to the entire visualization system. Because the first training error reflects the accuracy of the visual element vector mining process from three-dimensional mapping data to visual rendering output, and this accuracy directly affects the subsequent visual rendering operations. If the first training error is large, it means that there are major problems in the visual element vector mining. Then, even if the model visual rendering network itself has good performance, the target visual rendering output result finally obtained may be far from expected.

[0050] For example, in a digital city area visualization, a large first training error can lead to inaccurate mining of building visual feature vectors. During visual rendering, the model's visual rendering network renders based on inaccurate visual feature vectors, resulting in significant deviations in the visualization of building heights, colors, and appearances. By precisely determining the first training error and making appropriate adjustments, the accuracy of visual feature vector mining can be improved, providing more accurate input to the model's visual rendering network and ultimately improving the quality of the target visual rendering output.

[0051] This method, which determines the first training error by constructing multiple sets of vector chains and conducting joint training, plays a key role in connecting the entire data visualization display system. It not only improves the accuracy of visual feature vector mining, but also, combined with the second training error, enables more effective joint debugging of the first model visual mapping network and the model visual rendering network. During the joint debugging process, the first model visual mapping network is adjusted based on the first training error, and the model visual rendering network is adjusted based on the second training error. This allows the two networks to be collaboratively optimized, continuously improving the accuracy and quality of the entire visualization system, from the original 3D mapping data to the final target visual rendering output.

[0052] This design, by constructing multiple sets of vector chains and jointly training them to determine the first training error, achieves a more accurate error assessment for visual element vector mining. Compared with traditional methods, it can more comprehensively consider the relationships between vectors at different levels, improving the accuracy of error determination. This helps improve the quality of visual element vector mining and provides more accurate input for the model visual rendering network. This, in conjunction with the second training error, enables effective joint debugging of the first model visual mapping network and the model visual rendering network within the entire visualization system, improving the accuracy and quality of the output results from 3D mapping data to the target visual rendering, and providing more reliable and high-quality technical support for the model visualization display of digital cities.

[0053] According to a preferred technical approach, determining the first visual vector chain and the second visual vector chain includes: performing visual element vector mining on the Y three-dimensional surveying and mapping data samples using a second model visual mapping network to obtain a second visual element vector for each three-dimensional surveying and mapping data sample, and performing visual element vector mining on the X visual rendering output samples to obtain a second visual sample vector for each visual rendering output sample; migrating the second visual element vectors of the Y three-dimensional surveying and mapping data samples into the second visual vector chain; and migrating the second visual element vectors of the X visual rendering output samples into the first visual vector chain.

[0054] Under this preferred technical idea, determining the first visible vector chain and the second visible vector chain is a key link in the entire technical solution, and the process has unique logic and function.

[0055] First, the second model visual mapping network mines visual element vectors for each of the Y 3D mapping data samples, obtaining the second visual element vector for each. Simultaneously, visual element vectors are mined for each of the X visual rendering output samples, obtaining the second visual sample vector for each. This process is similar to the operations of the first model visual mapping network, but has different purposes and subsequent effects.

[0056] Taking a specific digital city area as an example, for the building part in the three-dimensional mapping data sample, for example, some characteristic values of the building in the three-dimensional mapping data are: height of 80 meters, floor area of 1000 square meters, and texture complexity value of the building facade of 0.4 (0.4 in the embodiment of the present invention is a value set to represent the complexity of the texture). The second visual element vector obtained after visual element vector mining through the second model visual mapping network can convert these values into vector representations that are more directly related to visualization. For example, the height vector can be converted into a height ratio vector at a specific visualization perspective (for example, 0.6, indicating the height ratio relative to other elements in the entire visualization scene) according to a certain mapping relationship, the floor area may be converted into a space proportion vector (for example, 0.1, indicating the space proportion in the visualization scene), and the facade texture complexity value may be converted into a texture visual effect vector (for example, 0.3, indicating the visual significance of the texture in the visualization).

[0057] For example, in a visual rendering output example, a scene containing a building has been initially visualized. In this visual rendering output example, some of the building's characteristic values in the visualization are: from a specific viewing angle, the building's displayed height ratio appears to be 0.55 (different from the ideal value of 0.6 mined from the previous 3D mapping data sample), its spatial proportion in the visualization scene appears to be 0.09 (different from the ideal value of 0.1), and its texture visual saliency appears to be 0.28 (different from the ideal value of 0.3). The second visual sample vector mined by the second model visual mapping network will represent these actual visual representations.

[0058] Then, the second visual element vectors of the Y 3D mapping data samples are migrated to the second visual vector chain. This migration operation makes the second visual vector chain a collection of visualization-related vectors mined from the 3D mapping data samples. These vectors contain visualization-related information about terrain, landforms, and other aspects of the original 3D mapping data after processing by the second model visual mapping network. For example, in addition to the relevant vectors of the aforementioned buildings, for roads in the digital city area, the 3D mapping data may have values such as a road width of 20 meters and a length of 500 meters. After mining and conversion, the second visual element vectors may include a width ratio vector (e.g., 0.05) and a length ratio vector (e.g., 0.2) of the road in the visualization scene. These vectors are all migrated to the second visual vector chain.

[0059] Similarly, the second visual sample vectors of the X visual rendering output samples are transferred to the first visual vector chain. This first visual vector chain then becomes a collection of vectors mined from the visual rendering output samples. These vectors reflect information related to the actual visual representation in the visual output results. For example, for different building and road combinations in multiple visual rendering output samples, the vectors in the first visual vector chain will record their actual visual features in each visual output, such as the relative height ratio vectors between buildings and the relative position vectors of roads and buildings.

[0060] This method of determining the first and second visual vector chains is crucial to the overall technical solution. It provides the foundation for subsequent joint training to determine the first training error based on the first visual element vectors of Y 3D mapping data samples, the first visual sample vectors of X visual rendering output samples, the U second visual sample vectors included in the first visual vector chain, and the V second visual element vectors included in the second visual vector chain. The vector chain constructed in this manner can more comprehensively encompass all aspects of information, from the original 3D mapping data to the visual rendering output.

[0061] Traditional data visualization techniques often oversimplify the construction of vector chains or lack a method for mining and migrating vectors from different perspectives (for example, 3D mapping data samples and visual rendering output). Traditional methods may simply extract a subset of vectors from the raw data or visualization output for simple comparison, but they struggle to comprehensively consider the visualization-related feature vectors of the data at different stages, as this technical solution does.

[0062] This method of constructing vector chains plays a key role in the error determination and network optimization process of the entire visualization system. For example, when determining the first training error, the vectors in the first visual vector chain and the second visual vector chain can be collaboratively analyzed with the first visual element vectors of Y three-dimensional mapping data samples and the first visual sample vectors of X visual rendering output samples. Because these vector chains contain visualization-related vectors mined from different levels and angles, the differences and associations between different vectors can be more accurately discovered during the joint training process. For example, when comparing the visualization effects of buildings, not only can the differences in basic information such as building height and floor area in the first visual element vector and the first visual sample vector be compared, but the vectors in the first visual vector chain and the second visual vector chain can also be used to further compare more detailed visual feature differences such as the relative height ratio and space occupancy of the building in the entire visualization scene, thereby more accurately determining the first training error.

[0063] This ability to accurately determine the first training error further impacts the performance of the entire visualization system. This is because the accuracy of the first training error is directly related to the optimization of the first model's visual mapping network. If the first training error can more accurately reflect problems in the process of mining visual feature vectors from 3D mapping data to visual rendering output, then the first model's visual mapping network can be adjusted more specifically. For example, if joint analysis of vector chains reveals that the first training error is primarily caused by problems in the mining and conversion of building space share vectors, adjustments can be made to the parts of the first model's visual mapping network related to space share calculations.

[0064] At the same time, this method of determining the first visual vector chain and the second visual vector chain is also related to the optimization of the model visual rendering network. Although the second training error is mainly determined based on the visual rendering training result obtained by the model visual rendering network performing a visual rendering operation on the first visual element vector of the three-dimensional mapping data sample, the vectors in the first visual vector chain and the second visual vector chain can indirectly affect the visual rendering operation. This is because accurate vector chain construction helps to more accurately determine the first training error, and the optimization of the first training error will make the visual element vector provided to the model visual rendering network more accurate, thereby improving the input quality of the model visual rendering network, and ultimately helping to improve the rendering effect of the model visual rendering network, thereby improving the accuracy and quality of the entire visualization system from the original three-dimensional mapping data to the final target visual rendering output result.

[0065] Thus, the method of mining data samples and migrating vectors to construct the first visual vector chain and the second visual vector chain through the second model visual mapping network comprehensively covers various visualization-related information from three-dimensional mapping data to visual rendering output. This provides a richer information basis for accurately determining the first training error, and can more accurately reflect problems in the visual element vector mining process than traditional methods. The method involved in the embodiment of the present invention helps to optimize the first model visual mapping network in a more targeted manner, while indirectly improving the input quality of the model visual rendering network, and ultimately improving the accuracy and quality of the entire visualization system, providing a more reliable technical guarantee for the three-dimensional model visualization display of digital cities.

[0066] In an alternative embodiment, performing joint training based on the first visual element vectors of the Y three-dimensional mapping data samples, the first visual sample vectors of the X visual rendering output samples, the U second visual sample vectors included in the first visual vector chain, and the V second visual element vectors included in the second visual vector chain to determine the first training error includes:

[0067] (1) for a p-th three-dimensional surveying and mapping data sample among the Y three-dimensional surveying and mapping data samples, determining a first characteristic commonality value between a first visual element vector of the p-th three-dimensional surveying and mapping data sample and each second visual sample vector among the U second visual sample vectors, where p is a positive integer not greater than Y;

[0068] (2) for a qth visual rendering output sample among the X visual rendering output samples, determining a second characteristic commonality value between a first visual sample vector of the qth visual rendering output sample and each second visual element vector among the V second visual element vectors, where q is a positive integer not greater than X;

[0069] (3) Determining the first training error based on a first feature commonality value corresponding to each of the Y three-dimensional surveying and mapping data samples and a second feature commonality value corresponding to each of the X visual rendering output samples.

[0070] In this alternative embodiment, the process of determining the first training error is a key part of the entire technical solution, involving complex relationship analysis and quantification operations between multiple vectors. First, for the pth 3D mapping data sample among the Y 3D mapping data samples, it is necessary to determine the first feature commonality value between its first visual element vector and each second visual sample vector among the U second visual sample vectors, where p is a positive integer not greater than Y. Taking a specific digital city 3D mapping data sample as an example, for example, the pth 3D mapping data sample is about a certain area and contains a building. Some feature values of the building in the first visual element vector are: a height value of 100 meters, a relative position coordinate value of (50, 80) in the entire area (the coordinates in the embodiment of the present invention are a value set to represent a relative position), and a texture feature value of 0.5 for the building facade (representing a certain quantitative feature of the texture).

[0071] One of the U second visual sample vectors may be mined from another visual rendering output sample and contains some feature vector values related to the visualization of similar buildings. For example, the relative proportion value of the building height in the visualization in the second visual sample vector is 0.8 (relative to other elements in the entire visualization scene), its relative position relationship value in the visualization scene is (0.4, 0.6) (indicating the relative position ratio with other elements), and the visual salience value of the facade texture in the visualization is 0.4.

[0072] When determining the first feature commonality value, it's necessary to comprehensively consider the various numerical features within these vectors. For example, for building heights, the degree of commonality can be calculated according to certain rules. If the height values are converted to relative proportions within their respective reference frames, for example, a 100-meter building has a relative height proportion of 0.6 within its own region, and this is compared with the 0.8 relative proportion within another second visual sample vector, the height commonality value can be calculated based on the difference between the two values and a certain weight (this weight may be determined based on the importance of height in the overall visualization). A similar approach is used for relative position coordinates and texture features. The calculated commonality values are combined (e.g., using a weighted summation) to ultimately determine the first feature commonality value between the first visual element vector and the second visual sample vector. This operation is repeated for each of the U second visual sample vectors to determine the first feature commonality value between the pth 3D mapping data sample and each second visual sample vector.

[0073] Next, for the qth visual rendering output sample among the X visual rendering output samples, a second characteristic commonality value between its first visual sample vector and each second visual element vector in the V second visual element vectors is determined, where q is a positive integer not greater than X. For example, the qth visual rendering output sample is a visualization result of a specific urban area, where some characteristic values of a building in the first visual sample vector are: a height ratio value of the building from a specific viewing angle is 0.7, a relative distance ratio value of the building to the surrounding roads is 0.3, and a color ratio value of the building's exterior color in the visualization is 0.4 (indicating the proportion of a certain color in the overall visual sense).

[0074] For example, one of the V second visual element vectors, mined from a 3D mapping data sample, contains a vector after conversion of some original relevant numerical values of the building. For example, the actual height value of the building is converted to a relative height ratio of 0.65, the relative distance value to the surrounding objects is converted to a ratio of 0.25, and a quantitative value of the building facade color is 0.35.

[0075] When determining the second feature commonality value, all numerical features must also be comprehensively considered. For example, for building height ratios, the commonality value for height is calculated based on the difference between the two values and the importance of height in visualization. Similar methods are used for relative distance ratios and color ratios, combining the commonality values calculated for each aspect (e.g., weighted summation) to obtain the second feature commonality value between the first visual sample vector of the qth visual rendering output sample and this second visual element vector. This operation is repeated for each of the V second visual element vectors to obtain the second feature commonality value between the qth visual rendering output sample and each second visual element vector.

[0076] Finally, the first training error is determined based on the first feature commonality value corresponding to each of the Y three-dimensional mapping data samples and the second feature commonality value corresponding to each of the X visual rendering output samples. In this process, these feature commonality values can be combined in a variety of ways. For example, all of the first feature commonality values and the second feature commonality values can be considered as a whole data set, and certain statistical characteristics of this data set are calculated to represent the first training error. Alternatively, the first feature commonality values and the second feature commonality values can be processed separately, such as taking a weighted average of the first feature commonality values to obtain a composite value, performing a similar operation on the second feature commonality values, and then determining the first training error by combining these two composite values using a certain functional relationship (such as the absolute value of the difference or other more complex functions).

[0077] This method for determining the first training error has significant advantages over traditional methods. Traditional data visualization techniques often lack detailed analysis of the commonalities between vectors at different levels when determining training errors. They typically simply compare superficial vector values or employ relatively crude methods to assess errors. However, the method in this embodiment, by deeply analyzing the commonalities between different vectors, can more accurately reflect the accuracy of visual element vector mining in the process from 3D mapping data to visual rendering output.

[0078] For example, traditional methods might simply compare building heights, ignoring their relative proportions within their respective reference frames and their comprehensive relationships with other features (such as location and texture). However, this embodiment comprehensively considers these factors by calculating the first and second feature commonality values, enabling more accurate detection of differences and similarities between vectors and thus more precisely determining the first training error.

[0079] The ability to determine the first training error involved in the embodiments of the present invention directly affects the optimization of the first model visual mapping network. If the first training error can accurately reflect the problems in the visual element vector mining process, the relevant parameters in the first model visual mapping network can be adjusted in a targeted manner. For example, if it is found that the first training error is mainly caused by a low commonality value of features in calculating the relative positions of buildings, the part of the first model visual mapping network related to position calculation can be adjusted to improve the accuracy of visual element vector mining.

[0080] At the same time, this accurate determination of the first training error also indirectly affects the model visual rendering network. This is because an accurate first model visual mapping network can provide the model visual rendering network with more precise visual feature vectors, thereby improving the input quality of the model visual rendering network, ultimately helping to improve the accuracy and quality of the entire visualization system, from raw 3D mapping data to the final target visual rendering output results.

[0081] As can be seen, determining the first training error by calculating the commonality of features between different vectors comprehensively considers the multifaceted characteristic relationships between individual vectors in both the 3D mapping data samples and the visual rendering output samples. Compared to traditional methods, this method more accurately reflects the accuracy issues in the visual feature vector mining process, facilitates targeted optimization of the first model's visual mapping network, and indirectly improves the input quality of the model's visual rendering network, thereby enhancing the accuracy and quality of the entire visualization system, providing more reliable technical support for model visualization in digital cities.

[0082] In an exemplary technical solution, the determining of the first training error based on the first feature commonality value corresponding to each of the Y three-dimensional surveying and mapping data samples and the second feature commonality value corresponding to each of the X visual rendering output samples includes: determining a third feature commonality value between the second visual element vector of the p-th three-dimensional surveying and mapping data sample and each of the U second visual sample vectors; determining a third feature commonality value between the second visual sample vector of the q-th visual rendering output sample and each of the V second visual element vectors. a fourth characteristic commonality value between each second visual element vector; determining a first local training error between the visual element vector and the visual sample vector based on the first characteristic commonality value and the third characteristic commonality value corresponding to each three-dimensional mapping data sample in the Y three-dimensional mapping data samples; determining a second local training error between the visual sample vector and the visual element vector based on the second characteristic commonality value and the fourth characteristic commonality value corresponding to each visual rendering output sample in the X visual rendering output samples; and obtaining the first training error based on the first local training error and the second local training error.

[0083] In this exemplary technical solution, the process of determining the first training error is a multi-step and complex operation, which involves in-depth analysis and comprehensive processing of multiple feature commonality values. First, for the p-th three-dimensional mapping data sample in the Y three-dimensional mapping data samples, it is necessary to determine the third feature commonality value between its second visual element vector and each second visual sample vector in the U second visual sample vectors. Taking a three-dimensional mapping data sample of a digital city as an example, for example, the building part in the p-th three-dimensional mapping data sample, its second visual element vector contains some feature values. For example, the structural complexity value of the building is 0.6 (0.6 in the embodiment of the present invention is a value that quantifies the complexity of the structure), and the proportion value of its internal space layout is 0.3 (indicating the proportional relationship between different functional spaces).

[0084] For example, a second visual sample vector among the U second visual sample vectors includes feature values related to visualization presentation. For example, the structural visual effect value of the building in the visualization is 0.5 (indicating the visual complexity of the structure), and the space ratio of the interior space in the visualization is 0.25.

[0085] When determining the third characteristic commonality value, the individual numerical features within these vectors are comprehensively considered. For structural complexity and structural visual effect, the commonality value may be calculated based on the degree of difference between the two and their relative importance within the overall visualization. For example, if a structural complexity value of 0.6 significantly differs from a structural visual effect value of 0.5, and structural complexity is considered relatively important within the overall visualization, then this value will have a greater impact on the third characteristic commonality value. A similar approach is employed for the internal space layout proportions and the spatial proportion of the internal space within the visualization. The commonality values calculated from each aspect are combined (e.g., using a weighted summation) to ultimately determine the third characteristic commonality value between the second visual element vector and this second visual sample vector. This operation is repeated for each of the U second visual sample vectors, resulting in the third characteristic commonality value between the pth 3D mapping data sample and each second visual sample vector.

[0086] Next, for the qth visual rendering output sample among the X visual rendering output samples, the fourth characteristic commonality value between its second visual sample vector and each second visual element vector in the V second visual element vectors is determined. For example, the second visual sample vector of the building portion in the qth visual rendering output sample contains some visualization-related characteristic values. For example, the building's appearance detail richness in the visualization has a value of 0.7 (indicating the visual representation of appearance details), and its integration with the surrounding environment has a value of 0.4 (indicating the degree of visual integration with the surrounding environment).

[0087] For one of the V second visual element vectors, for example, it contains a vector after some relevant values of the original building are converted. For example, the relative value of the actual appearance complexity of the building after conversion is 0.65, and the actual correlation degree of the surrounding objects after conversion is 0.35.

[0088] When determining the fourth characteristic commonality value, all numerical features must be comprehensively considered. For appearance detail richness and appearance complexity, the commonality value is calculated based on the difference between their values and their importance weights in the visualization. A similar method is used for integration with the surrounding environment and the degree of actual connection with surrounding features. The commonality values calculated from each aspect are combined (e.g., by weighted summation) to obtain the fourth characteristic commonality value between the second visual sample vector and this second visual element vector. This operation is repeated for each of the V second visual element vectors to obtain the fourth characteristic commonality value between the qth visual rendering output sample and each second visual element vector.

[0089] Then, based on the first feature commonality value and the third feature commonality value corresponding to each of the Y 3D mapping data samples, a first local training error is determined between the visual element vector and the visual sample vector. For example, for a 3D mapping data sample, its first feature commonality value indicates the degree of commonality with the visual sample vector in terms of certain characteristics (such as the building's height and location), while the third feature commonality value indicates the degree of commonality with the visual sample vector in terms of other characteristics (such as the building's structural complexity and internal spatial layout).

[0090] When determining the first local training error, the first and third feature commonality values can be treated as a single dataset and the first local training error can be determined by analyzing the statistical characteristics of this dataset. For example, the degree of dispersion of this dataset can be calculated. A large degree of dispersion indicates that the visual element vectors and visual sample vectors have little commonality across multiple features, resulting in a large first local training error. Alternatively, a weighted approach can be employed, assigning different weights to the features associated with the first and third feature commonality values based on their importance in the overall visualization. The weighted composite value is then calculated to represent the first local training error.

[0091] Similarly, a second local training error between the visual sample vector and the visual element vector is determined based on the second feature commonality value and the fourth feature commonality value corresponding to each of the X visual rendering output samples. For example, for a visual rendering output sample, its second feature commonality value indicates the degree of commonality with the visual element vector in terms of certain visual features (such as the height ratio of the building in the visualization, the distance ratio to surrounding roads, etc.), while the fourth feature commonality value indicates the degree of commonality with the visual element vector in terms of other visual features (such as the richness of the building's appearance details, the degree of integration with the surrounding environment, etc.).

[0092] The second local training error is determined using a method similar to that used for the first local training error. The second local training error can be represented by calculating a statistical feature of the dataset containing the second and fourth feature commonality values, or by calculating a weighted composite value based on the importance of the features.

[0093] Finally, the first training error is obtained based on the first and second local training errors. This process may involve a combination of various methods. For example, the first training error can be obtained by directly adding the first and second local training errors. Alternatively, different weights can be assigned to the first and second local training errors based on their relative importance in the overall error assessment, and then the weighted sum is used to obtain the first training error.

[0094] The above method of determining the first training error has significant advantages. Traditional data visualization technology often lacks such a comprehensive and detailed analysis method when determining the training error. Usually, it simply compares the numerical values of some vectors or uses a single method to evaluate the error. However, this embodiment determines the commonality values of features at different levels, and then determines the local training error, and finally obtains the first training error. This can more accurately reflect the relationship between the visual element vector and the visual sample vector in the process of converting three-dimensional mapping data into visual rendering output.

[0095] The error determination method involved in the embodiments of the present invention directly affects the optimization of the first model visual mapping network. If the first training error can accurately reflect the differences and common problems between the visual element vector and the visual sample vector, the relevant parameters in the first model visual mapping network can be adjusted in a targeted manner. For example, if the first local training error is mainly caused by problems with features related to the structural complexity of the building, the part of the first model visual mapping network related to the calculation of structural complexity can be adjusted to improve the accuracy of visual element vector mining.

[0096] At the same time, accurate determination of the first training error also indirectly affects the model visual rendering network. This is because an accurate first model visual mapping network can provide the model visual rendering network with more precise visual element vectors, thereby improving the input quality of the model visual rendering network, ultimately helping to improve the accuracy and quality of the entire visualization system, from raw 3D mapping data to the final target visual rendering output results.

[0097] Therefore, calculating the first training error by determining multiple feature commonalities comprehensively and deeply considers the multi-faceted relationship between visual element vectors and visual example vectors. Compared to traditional methods, this approach more accurately reflects the relationship between the two, helping to specifically optimize the first model's visual mapping network and indirectly improve the input quality of the model's visual rendering network, thereby enhancing the accuracy and quality of the entire visualization system and providing more reliable technical support for model visualization in digital cities.

[0098] On the basis of the above-mentioned exemplary technical solution, the determining of the first local training error between the visual element vector and the visual sample vector based on the first feature commonality value and the third feature commonality value corresponding to each three-dimensional surveying and mapping data sample in the Y three-dimensional surveying and mapping data samples includes: obtaining the prior feature commonality value between the visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors; determining the first local training error based on the prior feature commonality value between the visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, the first feature commonality value between the first visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, and the third feature commonality value between the second visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors.

[0099] Furthermore, the determining of the first local training error based on the prior feature commonality value between the visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, the first feature commonality value between the first visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, and the third feature commonality value between the second visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, comprises: The method comprises the steps of: performing feature enhancement on a first common visual feature information and a third common visual feature value between the second visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, to obtain first common visual feature information; performing bit-wise weighting on the first common visual feature value between the first visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, to obtain second common visual feature information; and determining the first local training error based on the first common visual feature information and the second common visual feature information corresponding to each three-dimensional surveying and mapping data sample in the Y three-dimensional surveying and mapping data samples.

[0100] In the next step, determining the first local training error based on the first common visual feature information and the second common visual feature information corresponding to each three-dimensional mapping data sample in the Y three-dimensional mapping data samples includes: iteratively summing the first common visual feature information and the second common visual feature information corresponding to each three-dimensional mapping data sample in the Y three-dimensional mapping data samples to obtain the first local training error.

[0101] In this exemplary embodiment, the process of determining the first local training error first involves obtaining a priori feature commonality values between the visual element vector of the pth 3D mapping data sample and U second visual element vectors. For example, for a building in a specific 3D mapping data sample, the visual element vector contains feature information such as the building's height (e.g., 100 meters) and floor area (e.g., 2000 square meters). For each of the U second visual element vectors, similar feature values mined from other related visualization examples may be included, such as the building's relative height (e.g., 0.8, a relative ratio) and relative floor area (e.g., 0.2, also a relative ratio) in a particular visualization. Determining the priori feature commonality value comprehensively considers the relationships between the various numerical features in these vectors. This process calculates the degree of commonality based on specific rules and algorithms. For example, a comprehensive priori feature commonality value is determined based on the numerical relationships between height and relative height, floor area and relative floor area, and their importance in the overall visualization.

[0102] Next, a first local training error is determined based on the prior feature commonality value between the visual element vector of the p-th three-dimensional mapping data sample and the U second visual sample vectors, the first feature commonality value between the first visual element vector of the p-th three-dimensional mapping data sample and the U second visual sample vectors, and the third feature commonality value between the second visual element vector of the p-th three-dimensional mapping data sample and the U second visual sample vectors.

[0103] In this process, the prior feature commonality value between the visual element vector of the p-th 3D mapping data sample and the U second visual sample vectors is enhanced with the third feature commonality value between the second visual element vector of the p-th 3D mapping data sample and the U second visual sample vectors, thereby obtaining first common visual feature information. Feature enhancement is an operation that deeply explores and integrates vector relationships. For example, the prior feature commonality value may reflect the initial level of commonality of basic building features (such as height and area), while the third feature commonality value may focus on the commonality of the building's internal structure or other deeper features. Through the feature enhancement operation, these two common features are integrated and optimized to obtain first common visual feature information that more comprehensively reflects the relationship between the visual element vector and the visual sample vector.

[0104] At the same time, the first feature commonality values between the first visual element vector of the pth three-dimensional mapping data sample and the U second visual sample vectors are weighted bit by bit to obtain the second common visual feature information. Bitwise weighting is an operation that performs weighted processing based on the importance of each feature in the vector. For example, if the height of the building in the first visual element vector is more important in the overall visualization, then when calculating the first feature commonality value, the value of the highly relevant part will be given a higher weight. The second common visual feature information obtained by this bitwise weighting method can more accurately reflect the commonality relationship under different feature importances.

[0105] Then, a first local training error is determined based on the first common visual feature information and the second common visual feature information corresponding to each of the Y 3D mapping data samples. Specifically, the first common visual feature information and the second common visual feature information corresponding to each of the Y 3D mapping data samples are iteratively summed to obtain the first local training error. This iterative summation process gradually accumulates the relevant information of each 3D mapping data sample, thereby obtaining an error value that comprehensively reflects the relationship between the visual element vector and the visual sample vector for all 3D mapping data samples. This value is the first local training error.

[0106] The above-mentioned method of determining the first local training error has obvious advantages over the traditional method. Traditional data visualization technology often lacks such a method of comprehensively considering the commonality values of multiple features and performing in-depth integration when determining similar errors. Usually, it simply compares some vector features or adopts a relatively single calculation method. However, this embodiment obtains the prior feature commonality value first, then performs feature enhancement, bit-by-bit weighting and other operations, and finally obtains the first local training error through iterative summation, which can more accurately reflect the complex relationship between the visual element vector and the visual sample vector.

[0107] The method for determining the first local training error involved in the embodiment of the present invention directly affects the optimization of the first model visual mapping network. If the first local training error can accurately reflect the differences and common problems between the visual element vector and the visual sample vector at multiple feature levels, the relevant parameters in the first model visual mapping network can be adjusted in a targeted manner. For example, if it is found in the calculation of the first local training error that the height-related features of a certain building show a large deviation after feature enhancement and bit-by-bit weighting, the part related to height calculation or processing in the first model visual mapping network can be adjusted, thereby improving the accuracy of visual element vector mining.

[0108] At the same time, accurate determination of the first local training error also indirectly affects the model visual rendering network. This is because an accurate first model visual mapping network can provide the model visual rendering network with more precise visual feature vectors, thereby improving the input quality of the model visual rendering network. Ultimately, this helps improve the accuracy and quality of the entire visualization system, from raw 3D mapping data to the final target visual rendering output.

[0109] This design determines the first local training error by obtaining prior feature commonality values and performing operations such as feature enhancement, bit-by-bit weighting, and iterative summation. This approach comprehensively and deeply considers the various relationships between visual element vectors and visual example vectors, more accurately reflecting these relationships than traditional methods. This helps to specifically optimize the first model's visual mapping network, indirectly improving the input quality of the model's visual rendering network, thereby enhancing the accuracy and quality of the entire visualization system and providing more reliable technical support for model visualization in digital cities.

[0110] Based on the above exemplary technical solution, determining a second local training error between a visual sample vector and a visual element vector based on a second feature commonality value and a fourth feature commonality value corresponding to each of the X visual rendering output samples includes: obtaining a priori feature commonality value between the visual sample vector of the qth visual rendering output sample and the V second visual element vectors; and determining the second local training error based on the priori feature commonality value between the visual sample vector of the qth visual rendering output sample and the V visual sample vectors, a second feature commonality value between a first visual element vector of the qth visual rendering output sample and the V second visual element vectors, and a fourth feature commonality value between the second visual element vector of the qth visual rendering output sample and the V second visual element vectors.

[0111] Furthermore, the determining of the second local training error based on the prior feature commonality value between the visual sample vector of the qth visual rendering output sample and the V visual sample vectors, the second feature commonality value between the first visual element vector of the qth visual rendering output sample and the V second visual element vectors, and the fourth feature commonality value between the second visual element vector of the qth visual rendering output sample and the V second visual element vectors comprises: The method further comprises performing feature enhancement on the prior feature commonality value of the qth visual rendering output sample and the third feature commonality value between the second visual element vector of the qth visual rendering output sample and the V visual sample vectors to obtain third common visual feature information; determining fourth common visual feature information using the first feature commonality value between the first visual element vector of the qth visual rendering output sample and the V visual sample vectors; and determining the second local training error based on the third common visual feature information and the fourth common visual feature information corresponding to each visual rendering output sample in the X visual rendering output samples.

[0112] In the next step, determining the second local training error based on the third common visual feature information and the fourth common visual feature information corresponding to each of the X visual rendering output samples includes iteratively summing the third common visual feature information and the fourth common visual feature information corresponding to each of the X visual rendering output samples to obtain the second local training error.

[0113] In this embodiment based on the above-mentioned exemplary technical solution, the process of determining the second local training error is first to obtain the prior feature commonality value between the visual sample vector of the qth visual rendering output sample and the V second visual element vectors. For example, in the building part of a visual rendering output sample, the visual sample vector contains a specific visual feature value. For example, the color feature value of the building in the visual rendering output is 0.6 (0.6 in the embodiment of the present invention is a value that quantifies the color performance, for example, it can represent the proportion of a certain color in the overall vision or the degree of color vividness, etc.), and its position relationship value in the visualization scene is 0.3 (a quantitative value representing the relative position relationship with other elements). For one of the V second visual element vectors, it may contain relevant values mined from the three-dimensional surveying and mapping data sample, such as a certain quantitative value of the original color of the building is 0.55, and the ratio after conversion of the original position relationship value is 0.25. The acquisition of the prior feature commonality value is to comprehensively consider the relationship between the various numerical features in these vectors and calculate the degree of commonality between them based on specific algorithms and rules. This prior feature commonality value reflects the commonality relationship between the visual sample vector and the second visual element vector at the initial level.

[0114] Next, a second local training error is determined based on a prior feature commonality value between the visual sample vector of the qth visual rendering output sample and the V visual sample vectors, a second feature commonality value between the first visual element vector of the qth visual rendering output sample and the V second visual element vectors, and a fourth feature commonality value between the second visual element vector of the qth visual rendering output sample and the V second visual element vectors.

[0115] In this process, the prior feature commonality value between the visual sample vector of the qth visual rendering output sample and the V visual sample vectors is enhanced with the third feature commonality value between the second visual element vector of the qth visual rendering output sample and the V visual sample vectors to obtain third common visual feature information. For example, the prior feature commonality value reflects the initial commonality between the visual sample vector and the second visual element vector in basic visual features (such as color and positional relationship), while the third feature commonality value may focus on other deeper or different aspects of feature commonality (such as the relationship between the details of the building in the visualization and the corresponding features in the original data). Through the feature enhancement operation, the common features of these two aspects are integrated and optimized, thereby obtaining third common visual feature information that more comprehensively reflects the relationship between the visual sample vector and the visual element vector.

[0116] At the same time, the fourth common visual feature information is determined using the first feature commonality value between the first visual element vector of the qth visual rendering output sample and the V visual sample vectors. This process constructs the fourth common visual feature information based on the information contained in the first feature commonality value. The first feature commonality value reflects the degree of commonality between the first visual element vector of the visual rendering output sample and the V visual sample vectors in terms of certain specific features. By processing and transforming this commonality information, the fourth common visual feature information that accurately reflects the specific relationship is obtained.

[0117] Then, a second local training error is determined based on the third common visual feature information and the fourth common visual feature information corresponding to each of the X visual rendering output samples. Specifically, the third common visual feature information and the fourth common visual feature information corresponding to each of the X visual rendering output samples are iteratively summed to obtain the second local training error. This iterative summation process gradually accumulates the relevant information of each visual rendering output sample, ultimately obtaining a value that comprehensively reflects the error in the relationship between the visual sample vector and the visual element vector for all visual rendering output samples, i.e., the second local training error.

[0118] This method of determining the second local training error has significant advantages over traditional methods. Traditional data visualization technology often lacks this method of comprehensive processing of multiple feature commonality values in similar error determination. Traditional methods may simply compare some vector features or adopt a relatively single calculation method, which is difficult to fully and accurately reflect the complex relationship between the visual sample vector and the visual element vector. This embodiment obtains the prior feature commonality value, performs feature enhancement, determines the fourth common visual feature information, and finally obtains the second local training error through iterative summation, which can more accurately explore the intrinsic relationship between the visual sample vector and the visual element vector.

[0119] The second local training error determination method involved in the embodiment of the present invention directly affects the optimization of the first model visual mapping network and the model visual rendering network. For the first model visual mapping network, if the second local training error can accurately reflect the differences and common problems between the visual sample vector and the visual element vector at multiple feature levels, it can provide a basis for adjusting the first model visual mapping network. For example, if it is found that there is a deviation in the processing of features such as color or position relationship of a certain building when calculating the second local training error, the parts related to these features in the first model visual mapping network can be adjusted, thereby improving the accuracy of visual element vector mining.

[0120] For the model visual rendering network, accurate determination of the second local training error helps improve the input quality of the model visual rendering network. This is because accurate determination of the second local training error reflects the accuracy of the relationship between the visual sample vector and the visual feature vector, which indirectly affects the quality of the visual feature vector, and thus affects the input of the model visual rendering network. Ultimately, it helps improve the accuracy and quality of the entire visualization system, from the original 3D mapping data to the final target visual rendering output.

[0121] By obtaining prior feature commonality values, performing feature enhancement, identifying relevant common visual feature information, and iteratively summing the results, the second local training error is determined. This approach comprehensively considers the various relationships between visual example vectors and visual element vectors, more accurately reflecting these relationships than traditional methods. This helps optimize the first model's visual mapping network and model rendering network, improving the accuracy and quality of the entire visualization system and providing more reliable technical support for model visualization in digital cities.

[0122] Based on the above preferred technical ideas, in an independent embodiment, the first model visual mapping network and the model visual rendering network are jointly debugged based on the first training error and the second training error, including: optimizing the network model variables in the first model visual mapping network and the model visual rendering network through the first training error and the second training error to obtain the network model optimization variables of the first model visual mapping network and the network model optimization variables of the model visual rendering network; optimizing the network model variables in the first model visual mapping network based on the network model optimization variables of the first model visual mapping network, and optimizing the network model variables in the model visual rendering network based on the network model optimization variables of the model visual rendering network.

[0123] In addition, the method also includes: determining the network model optimization variables of the second model visual mapping network based on the network model optimization variables of the first model visual mapping network; and optimizing the network model variables in the second model visual mapping network based on the network model optimization variables of the second model visual mapping network.

[0124] Furthermore, determining the network model optimization variables of the second model visual mapping network based on the network model optimization variables of the first model visual mapping network includes: determining the optimized network model weight of the second model visual mapping network based on the current network model weight of the second model visual mapping network and the optimized network model weight of the first model visual mapping network.

[0125] In this independent embodiment, the joint debugging of the first model visual mapping network and the model visual rendering network based on the first training error and the second training error is a key part of the entire technical solution. For the first model visual mapping network and the model visual rendering network, network model variables are important factors that determine their functions and performance. These network model variables can be regarded as a set of elements with specific numerical characteristics, existing in the form of multidimensional numerical feature vectors. For example, the network model variables in the first model visual mapping network may contain vector components related to different types of visual element mining. For example, when processing building data in a digital city, the numerical feature vector of the network model variables in one dimension is related to the mining of the building's geometry, and its numerical value can be expressed as [0.3, 0.5, 0.2], and the other dimension is related to the mining of the building's texture, and its numerical value is [0.4, 0.1, 0.5]. The network model variables in the model visual rendering network also have such a multidimensional numerical feature vector form. For example, in the network model variables related to light and shadow effect rendering, one dimension may be related to light direction calculation with a value of [0.2, 0.6, 0.2], and another dimension may be related to shadow color rendering with a value of [0.5, 0.3, 0.2].

[0126] When optimizing the network model variables in these two networks using the first and second training errors, these multidimensional numerical feature vectors are analyzed and adjusted. For the first model visual mapping network, the first training error reflects deviations in the visual feature vector mining process. If significant deviations from the actual visual feature vectors of building geometry are detected (as reflected by the first training error), the numerical feature vectors of the network model variables related to building geometry mining ([0.3, 0.5, 0.2]) will be adjusted. For example, to [0.2, 0.6, 0.2], this adjustment process uses the information provided by the first training error to change the values of each dimension to reduce mining errors. Similarly, the numerical feature vectors of other network model variables related to the first training error are adjusted accordingly, resulting in the optimized network model variables for the first model visual mapping network. For the model visual rendering network, the second training error reflects deviations in the visual rendering operation. If, during the rendering of light and shadow effects, it is found that the shadow color rendering does not meet expectations (reflected by the second training error), then the numerical feature vector of the network model variables related to the shadow color rendering [0.5, 0.3, 0.2] will be adjusted, for example, to [0.4, 0.4, 0.2]. After such an adjustment, the network model optimization variables of the model visual rendering network are obtained.

[0127] Then, based on the network model optimization variables of the first model visual mapping network, the network model variables in the first model visual mapping network are optimized. This is the process of directly applying the optimization variables obtained above to the first model visual mapping network. For example, the adjusted network model variable numerical feature vector [0.2, 0.6, 0.2] related to building geometry mining replaces the original vector [0.3, 0.5, 0.2], thereby improving the accuracy of the first model visual mapping network in visual element vector mining. For the model visual rendering network, optimizing its own network model variables based on its network model optimization variables is a similar process, replacing the original vector with the adjusted vector related to shadow color rendering, thereby improving the accuracy of the visual rendering operation.

[0128] Furthermore, the network model optimization variables of the second model visual mapping network are determined based on the network model optimization variables of the first model visual mapping network. The key to this embodiment of the present invention is to determine the optimized network model weights of the second model visual mapping network based on the current network model weights of the second model visual mapping network (also in the form of a multidimensional numerical feature vector) and the optimized network model weights of the first model visual mapping network. For example, when processing a certain building feature (such as internal structure mining), the current network model weights of the second model visual mapping network have a numerical feature vector of [0.4, 0.3, 0.3]. After optimization, the numerical feature vector of the optimized network model weights related to similar building internal structure mining of the first model visual mapping network becomes [0.3, 0.4, 0.3]. Based on the relationship between the two (this relationship may be based on a complex algorithm, such as involving distance calculation, proportional relationship, or other logical relationship between vectors), the optimized network model weight numerical feature vector of the second model visual mapping network for building internal structure mining is calculated, such as [0.35, 0.35, 0.3]. Then, based on this optimized network model weight, the network model variables in the second model visual mapping network are optimized. For example, if the numerical feature vector of a network model variable related to building internal structure mining in the second model visual mapping network is [0.5, 0.2, 0.3], based on the new optimized network model weight [0.35, 0.35, 0.3], this vector can be adjusted to [0.4, 0.3, 0.3].

[0129] The above-mentioned joint debugging method has significant advantages. Traditional data visualization technology often lacks such a comprehensive, systematic optimization method based on network model variables (in the form of multidimensional numerical feature vectors) in network debugging. Traditional methods may simply adjust some parameters in the network without taking into account the multidimensionality of network model variables and the complex relationships between different networks. However, this embodiment uses the first training error and the second training error to jointly debug the first model visual mapping network and the model visual rendering network, and further links it to the optimization of the second model visual mapping network. It can comprehensively improve the accuracy of the entire visualization system from multiple dimensions and the relationship between multiple networks.

[0130] From the perspective of the first model visual mapping network and the model visual rendering network, optimizing the multidimensional numerical feature vectors of network model variables through precise training errors can improve the accuracy of visual feature vector mining and the quality of visual rendering operations. This makes the entire process from 3D mapping data to visual rendering output more precise, reducing the accumulation and propagation of errors across different dimensions and networks. For the second model visual mapping network, the associated optimization with the first model visual mapping network further ensures coordination and consistency between different links in the entire visualization process, improving the overall performance of the entire system from data mining to visualization.

[0131] In this way, the first model visual mapping network and the model visual rendering network are jointly debugged using the first and second training errors, including optimization of network model variables (in the form of multidimensional numerical feature vectors) and optimization of the associated second model visual mapping network. Compared to traditional methods, this approach more comprehensively and systematically considers the multidimensionality of network model variables and the relationships between networks, improving the accuracy of visual feature mining and visual rendering operations, as well as the coordination and consistency of the entire visualization system, and enhancing the overall performance from 3D mapping data to visual rendering output.

[0132] In another independent embodiment, determining the second training error based on the visual rendering training results corresponding to the Y three-dimensional mapping data samples includes: obtaining the visual rendering certification outputs corresponding to the Y three-dimensional mapping data samples; and determining the second training error based on the visual rendering certification outputs and the visual rendering training results corresponding to the Y three-dimensional mapping data samples.

[0133] Specifically, for Y three-dimensional mapping data samples, the visual rendering training results are the results obtained after the model visual rendering network performs a visual rendering operation on the first visual element vectors of the Y three-dimensional mapping data samples. These results contain information about how to convert the visual element vectors in the three-dimensional mapping data samples into visual presentations. For example, for a building portion in a certain three-dimensional mapping data sample, the visual rendering training results may have a numerical feature of the building's height in the visualization as 80 meters (indicating the height presented in the visualization), and a numerical feature of its color in the visualization as 0.6 (indicating the quantized value of a certain color, such as the vividness of the color or its relative position in the entire color space).

[0134] Next, obtain the visual rendering verification output corresponding to each of the Y 3D mapping data samples. A visual rendering verification output is a visualization result that serves as a standard or reference. For example, for the same building, the building height numerical feature in the visual rendering verification output should be 100 meters (different from the 80 meters in the visual rendering training result, reflecting possible errors), and the color numerical feature should be 0.7.

[0135] The second training error is determined based on the visual rendering verification output and the visual rendering training results corresponding to Y 3D mapping data samples. This process quantifies the error by comparing the differences between the two. For building height, there is a 20-meter difference between 80 meters in the visual rendering training results and 100 meters in the visual rendering verification output. This difference reflects the error in height rendering. For color numerical features, there is a 0.1 difference between 0.6 in the visual rendering training results and 0.7 in the visual rendering verification output. By combining the differences in these different aspects (such as height and color), the second training error can be determined. This combination may involve weighting different features. For example, because height is more important in the overall visualization, its difference may be given a larger weight in calculating the second training error, while color differences may be given a relatively smaller weight. In this way, a second training error value is ultimately determined that fully reflects the difference between the visual rendering training results and the visual rendering verification output.

[0136] In this way, the second training error is determined by comparing the visual rendering certification output with the visual rendering training results. This approach quantifies the error by comparing the difference between the reference standard and the actual rendering results, and comprehensively calculates the difference by taking into account the importance of different features. Compared with traditional methods, this method more accurately reflects the gap between the rendering effect of the model visual rendering network and the expected effect. It helps to optimize the model visual rendering network in a targeted manner, thereby improving the accuracy and quality of the entire visualization system and providing more reliable technical support for model visualization in digital cities.

[0137] In summary, the embodiments of the present invention can effectively process three-dimensional mapping data containing multiple types of data such as terrain, objects, and textures. The first model visual mapping network mines visual element vectors, providing accurate input for subsequent rendering and improving visualization efficiency. The model visual rendering network performs rendering operations based on the mined vectors to achieve visual display of the target digital city area. By jointly debugging the two networks based on the first training error and the second training error, network performance can be accurately optimized, the accuracy and fidelity of the visualization effect can be improved, the spatial characteristics of the digital city can be better presented, and various needs such as urban planning and management can be met.

[0138] Furthermore, Figure 2 FIG2 is a structural diagram of a data visualization display system 200 provided in an embodiment of the present invention. Figure 2 The data visualization display system 200 shown includes a processor 210, which can call and run a computer program from a memory to implement the method in the embodiment of the present invention. Figure 2 As shown, the data visualization display system 200 may further include a memory 230. The processor 210 may call and run a computer program from the memory 230 to implement the method in the embodiment of the present invention. The memory 230 may be a separate device independent of the processor 210 or may be integrated into the processor 210. Optionally, as Figure 2 As shown, data visualization display system 200 may further include a transceiver 220. Processor 210 may control transceiver 220 to interact with other devices. Specifically, it may send information or data to other devices or receive information or data sent by other devices. Optionally, data visualization display system 200 may implement the corresponding processes of the storage engine, or components within the storage engine (such as a processing module), or devices equipped with the storage engine, in the various methods of the embodiments of the present invention. For the sake of brevity, these processes are not further described here.

[0139] It should be understood that the processor in the embodiment of the present invention may be an integrated circuit chip with signal processing capabilities.

[0140] It is understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. It should be noted that the memory of the systems and methods described herein is intended to include but is not limited to suitable types of memory.

[0141] Based on the above, a readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, the steps of the above method are implemented.

[0142] The above describes an embodiment of the present invention in conjunction with the accompanying drawings, but the embodiment of the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of the embodiment of the present invention, ordinary technicians in this field can also make many forms without departing from the purpose of the embodiment of the present invention and the scope of protection of the embodiment of the present invention, all of which are protected by the embodiment of the present invention.

Claims

1. A data visualization display method based on a business model, characterized in that: The method is applied to a data visualization display system, and the method includes: Acquire three-dimensional surveying and mapping data to be processed, wherein the three-dimensional surveying and mapping data to be processed is used to describe a multi-modal spatial representation vector of a target digital city area, and the three-dimensional surveying and mapping data to be processed includes terrain data, object data, and texture data; Performing visual element vector mining on the three-dimensional surveying and mapping data to be processed by using a first model visual mapping network to obtain visual element vectors of the three-dimensional surveying and mapping data to be processed; Performing a visual rendering operation on the visual element vectors of the to-be-processed three-dimensional surveying and mapping data through a model visual rendering network to obtain a target visual rendering output result; the target visual rendering output result is used to realize a model visualization display of the target digital city area; The model visual rendering network and the first model visual mapping network are obtained by joint debugging based on the first training error and the second training error; The debugging steps of the first model visual mapping network and the model visual rendering network include: Obtaining X visual rendering output samples and Y three-dimensional surveying and mapping data samples included in the current debugging sample set, and performing visual element vector mining on each three-dimensional surveying and mapping data sample through a first model visual mapping network to obtain a first visual element vector of each three-dimensional surveying and mapping data sample, and performing visual element vector mining on each visual rendering output sample to obtain a first visual sample vector of each visual rendering output sample, wherein the three-dimensional surveying and mapping data sample is used to describe a multimodal spatial representation vector of a sample digital city area, and X and Y are both positive integers; Determining a first training error based on the first visual element vectors of the Y three-dimensional surveying and mapping data samples and the first visual sample vectors of the X visual rendering output samples; performing a visual rendering operation on the first visual element vectors of the Y three-dimensional surveying and mapping data samples through a model visual rendering network to obtain visual rendering training results corresponding to the Y three-dimensional surveying and mapping data samples, and determining a second training error based on the visual rendering training results corresponding to the Y three-dimensional surveying and mapping data samples; The first model visual mapping network and the model visual rendering network are jointly debugged according to the first training error and the second training error.

2. The method according to claim 1, wherein The determining a first training error based on the first visual element vectors of the Y three-dimensional surveying and mapping data samples and the first visual sample vectors of the X visual rendering output samples includes: Determine a first visual vector chain and a second visual vector chain, wherein the first visual vector chain includes U second visual sample vectors, and the second visual vector chain includes V second visual element vectors, where U and V are both positive integers; Joint training is performed based on the first visual element vectors of the Y three-dimensional mapping data samples, the first visual sample vectors of the X visual rendering output samples, the U second visual sample vectors included in the first visual vector chain, and the V second visual element vectors included in the second visual vector chain to determine the first training error.

3. The method according to claim 2, characterized in that The determining of the first visible vector chain and the second visible vector chain includes: Performing visual element vector mining on each of the Y three-dimensional surveying and mapping data samples using a second model visual mapping network to obtain a second visual element vector for each of the three-dimensional surveying and mapping data samples, and performing visual element vector mining on each of the X visual rendering output samples to obtain a second visual sample vector for each of the visual rendering output samples; Migrating the second visual element vectors of the Y three-dimensional surveying and mapping data samples into the second visual vector chain; Migrating the second visual element vectors of the X visual rendering output samples into the first visual vector chain.

4. The method according to claim 3, characterized in that The determining the first training error by performing joint training based on the first visual element vectors of the Y three-dimensional surveying and mapping data samples, the first visual sample vectors of the X visual rendering output samples, the U second visual sample vectors included in the first visual vector chain, and the V second visual element vectors included in the second visual vector chain includes: For a p-th three-dimensional surveying and mapping data sample among the Y three-dimensional surveying and mapping data samples, determining a first characteristic commonality value between a first visual element vector of the p-th three-dimensional surveying and mapping data sample and each second visual sample vector among the U second visual sample vectors, where p is a positive integer not greater than Y; For a qth visual rendering output sample among the X visual rendering output samples, determining a second characteristic commonality value between a first visual sample vector of the qth visual rendering output sample and each second visual element vector of the V second visual element vectors, where q is a positive integer not greater than X; The first training error is determined according to a first feature commonality value corresponding to each 3D surveying and mapping data sample in the Y 3D surveying and mapping data samples and a second feature commonality value corresponding to each 3D surveying and mapping output sample in the X visual rendering output samples.

5. The method according to claim 4, characterized in that The determining the first training error according to the first feature commonality value corresponding to each of the Y three-dimensional surveying and mapping data samples and the second feature commonality value corresponding to each of the X visual rendering output samples includes: Determining a third characteristic commonality value between the second visual element vector of the p-th three-dimensional surveying and mapping data sample and each second visual sample vector of the U second visual sample vectors; Determining a fourth characteristic commonality value between the second visual sample vector of the qth visual rendering output sample and each second visual element vector of the V second visual element vectors; Determining a first local training error between a visual element vector and a visual sample vector based on a first feature commonality value and a third feature commonality value corresponding to each of the Y three-dimensional surveying and mapping data samples; Determining a second local training error between the visual sample vector and the visual element vector based on the second feature commonality value and the fourth feature commonality value corresponding to each visual rendering output sample in the X visual rendering output samples; The first training error is obtained according to the first local training error and the second local training error.

6. The method according to claim 5, characterized in that Determining a first local training error between a visual element vector and a visual sample vector based on the first feature commonality value and the third feature commonality value corresponding to each of the Y three-dimensional surveying and mapping data samples includes: Obtaining a priori feature commonality values between the visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors; The first local training error is determined based on a prior feature commonality value between the visual element vector of the p-th three-dimensional mapping data sample and the U second visual sample vectors, a first feature commonality value between the first visual element vector of the p-th three-dimensional mapping data sample and the U second visual sample vectors, and a third feature commonality value between the second visual element vector of the p-th three-dimensional mapping data sample and the U second visual sample vectors.

7. The method according to claim 6, characterized in that Determining the first local training error based on a priori feature commonality values between the visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, a first feature commonality value between the first visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, and a third feature commonality value between the second visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, includes: Performing feature enhancement on a priori feature commonality value between the visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, and a third feature commonality value between the second visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors, to obtain first common visual feature information; performing bit-wise weighting on the first feature commonality values between the first visual element vector of the p-th three-dimensional surveying and mapping data sample and the U second visual sample vectors to obtain second common visual feature information; determining the first local training error based on the first common visual feature information and the second common visual feature information corresponding to each of the Y three-dimensional surveying and mapping data samples; The determining the first local training error based on the first common visual feature information and the second common visual feature information corresponding to each of the Y three-dimensional surveying and mapping data samples includes: iteratively summing the first common visual feature information and the second common visual feature information corresponding to each of the Y three-dimensional surveying and mapping data samples to obtain the first local training error.

8. The method according to claim 7, characterized in that The determining, based on the second feature commonality value and the fourth feature commonality value corresponding to each of the X visual rendering output samples, a second local training error between the visual sample vector and the visual element vector comprises: Obtaining a priori feature commonality values between the visual sample vector of the qth visual rendering output sample and the V second visual element vectors; The second local training error is determined based on a prior feature commonality value between the visual sample vector of the qth visual rendering output sample and the V visual sample vectors, a second feature commonality value between the first visual element vector of the qth visual rendering output sample and the V second visual element vectors, and a fourth feature commonality value between the second visual element vector of the qth visual rendering output sample and the V second visual element vectors.

9. The method according to claim 8, characterized in that The determining the second local training error based on a priori feature commonality value between the visual sample vector of the qth visual rendering output sample and the V visual sample vectors, a second feature commonality value between the first visual element vector of the qth visual rendering output sample and the V second visual element vectors, and a fourth feature commonality value between the second visual element vector of the qth visual rendering output sample and the V second visual element vectors, includes: performing feature enhancement on a priori feature commonality value between the visual sample vector of the qth visual rendering output sample and the V visual sample vectors, and a third feature commonality value between the second visual element vector of the qth visual rendering output sample and the V visual sample vectors, to obtain third common visual feature information; Determining fourth common visual feature information by using a first feature commonality value between the first visual element vector of the qth visual rendering output sample and the V visual sample vectors; determining the second local training error according to the third common visual feature information and the fourth common visual feature information corresponding to each of the X visual rendering output samples; Determining the second local training error based on the third common visual feature information and the fourth common visual feature information corresponding to each of the X visual rendering output samples includes iteratively summing the third common visual feature information and the fourth common visual feature information corresponding to each of the X visual rendering output samples to obtain the second local training error.

10. A data visualization display system, characterized in that: The method comprises at least one processor and a memory; the memory stores computer-executable instructions; the at least one processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the method according to any one of claims 1 to 9.

Images