Urban three-dimensional building model decryption method and device, electronic equipment and storage medium

Abstract

The present application belongs to the technical field of three-dimensional geographic information data decryption, and provides a city three-dimensional building model decryption method, device, electronic equipment and storage medium, wherein the method comprises: preprocessing first DEM data and first city three-dimensional building model data of a region to be decrypted to obtain second DEM data and second city three-dimensional building model data; constructing and training a GAN generative adversarial network through the second DEM data, performing encryption processing on the second DEM data through the GAN generative adversarial network to obtain third DEM data; performing detail enhancement on the third DEM data by using a content generation algorithm to obtain fourth DEM data; and performing decryption processing on the second city three-dimensional building model data through the fourth DEM data to obtain a decryption result. The present application has the beneficial effect of improving the data availability of a three-dimensional city model in a secret state.

Description

Technical Field

[0001] This invention relates to the field of three-dimensional geographic information data declassification technology, and in particular to a method, apparatus, electronic device and storage medium for declassifying urban three-dimensional building models. Background Technology

[0002] In recent years, with the deepening development of concepts such as smart cities and digital twins, detailed 3D urban building models have become a key digital infrastructure in fields such as urban planning, emergency response, traffic simulation, and security. These models, through high-precision data, reproduce the real physical world in virtual space, providing a solid foundation for various analyses and decision-making processes. However, while bringing enormous application value, 3D urban building models also face serious geographic information data security issues.

[0003] Current geographic information data declassification technologies exist, but most of these methods are applied to two-dimensional spatial data or have significant limitations in three-dimensional scenes. Common methods mainly include:

[0004] (1) Coordinate offset and perturbation method, which mainly achieves decryption by applying uniform translation or random noise to all coordinate points. However, the confidentiality of the overall translation is fragile. Once the attacker has a small number of real coordinate points, it can be easily cracked. At the same time, random perturbation will seriously damage the geometry and topology of the model, causing building distortion and road breakage, which greatly reduces the availability of data.

[0005] (2) Geometric transformation and simplification method. This method hides information by reducing the geometric details of the model (such as simplifying the fine model into a box) or removing some components. However, this is a decryption method at the cost of sacrificing the amount of information. The processed data will not meet the needs of refined applications such as solar radiation analysis and field of view analysis.

[0006] (3) Data aggregation and location fuzzification method. This method fuzzes the location by rounding the coordinate values ​​or reducing the storage precision. Its disadvantage is that the decryption effect is difficult to control precisely and may cause model topology errors. It cannot be used for GIS applications that rely on high-precision positioning (such as autonomous driving simulation) due to low location precision. Summary of the Invention

[0007] Aimed at at least in solving one of the technical problems existing in the prior art, the present invention provides a method, apparatus, electronic device and storage medium for declassifying urban three-dimensional building models, thereby improving the data availability of three-dimensional urban models when maintaining confidentiality.

[0008] One aspect of the present invention provides a method for decrypting urban three-dimensional building models, comprising:

[0009] Preprocess the first DEM data and the first city 3D building model data of the area to be de-densified to obtain the second DEM data and the second city 3D building model data.

[0010] A GAN (Generative Adversarial Network) is constructed and trained using the second DEM data. The second DEM data is then encrypted using the GAN to obtain the third DEM data.

[0011] The third DEM data is enhanced with a content generation algorithm to obtain the fourth DEM data;

[0012] The second city's three-dimensional building model data is de-encrypted using the fourth DEM data to obtain the de-encryption result.

[0013] According to the aforementioned method for de-declassifying urban 3D building models, the first DEM data and the first urban 3D building model data of the area to be de-declassified are preprocessed to obtain second DEM data and second urban 3D building model data, including:

[0014] The first DEM data and the first city 3D building model data are aligned using a GIS or programming library, wherein the alignment method includes coordinate system transformation and projection unification.

[0015] The aligned first DEM data and the first city 3D building model data are preprocessed to obtain the second DEM data and the second city 3D building model data. The preprocessing includes filling in missing values ​​and removing obvious erroneous or abnormal elevation points.

[0016] According to the urban 3D building model de-densification method, the accuracy of the first DEM data is less than 2 meters, and the first DEM data is high-precision DEM data generated from airborne LiDAR laser scanning point cloud data.

[0017] According to the aforementioned method for decrypting urban 3D building models, a GAN (Generative Adversarial Network) is constructed and trained using the second DEM data. The GAN is then used to encrypt the second DEM data to obtain third DEM data, including:

[0018] The second DEM data is sliced ​​to obtain DEM tiles, and the DEM tiles are normalized to obtain a training set;

[0019] Construct a Generative Adversarial Network (GAN), including a generator with a U-Net structure and a discriminator with a deep convolutional neural network;

[0020] The generator and discriminator of the GAN (Generative Adversarial Network) are trained for adversarial game using the loss function and training strategy of WGAN-GP on the training set.

[0021] The second DEM data is segmented using a sliding window. Each sliding window is processed by a trained generator of a GAN (Generative Adversarial Network) to obtain encrypted DEM tiles. The encrypted DEM tiles are then weighted and averaged to obtain the third DEM data.

[0022] According to the aforementioned method for declassifying urban 3D building models, the third DEM data is enhanced with a content generation algorithm to obtain fourth DEM data, including:

[0023] The Perlin noise or Simplex noise algorithm is used to generate a multi-level frequency harmonic fractal noise map;

[0024] The fractal noise map is weighted and fused with the third DEM data to obtain the fourth DEM data, wherein the weighted fusion is expressed as follows:

[0025] ;

[0026] in, This is the fourth DEM data. This is the third DEM data. This is a fractal noise diagram. Enhancement factor for details.

[0027] According to the aforementioned method for declassifying urban 3D building models, the second urban 3D building model data is declassified using the fourth DEM data to obtain the declassification result, including:

[0028] Calculate the elevation offset of each individual building in the second city's 3D building model data:

[0029] ;

[0030] in, This represents the building's elevation offset. The elevation values ​​of the buildings in the first DEM data. The elevation values ​​of the buildings in the fourth DEM data. The average position of the geometric center point or the vertex of the foundation of the building;

[0031] For the geometric vertices of all building models in the second city's 3D building model data, for each building model's geometric vertices... Update:

[0032] ;

[0033] in, For the updated vertices coordinate, The vertex before the update coordinate;

[0034] Repeat the process of adjusting the elevation offset and geometric vertices of the building model until all the vertices of all buildings in the second city's 3D building model data are updated to obtain the de-dense result.

[0035] According to the aforementioned method for declassifying urban 3D building models, calculating the elevation offset of each independent building in the second urban 3D building model data further includes:

[0036] Extract the base polygon for each individual building, and perform the following processing based on the area of ​​the base polygon:

[0037] If the area of ​​the base polygon is less than or equal to the preset area, the overall elevation offset is calculated directly.

[0038] If the area of ​​the base polygon is greater than the preset area, multiple points of the base polygon are randomly sampled. The elevation offset of each sample point is calculated, and the average or median value of the elevation offset of the sample points is used as the overall elevation offset.

[0039] Another aspect of the present invention provides a device for decrypting urban three-dimensional building models, comprising:

[0040] The first module is used to preprocess the first DEM data and the first city 3D building model data of the area to be de-densified to obtain the second DEM data and the second city 3D building model data.

[0041] The second module is used to construct and train a GAN generative adversarial network using the second DEM data, and to encrypt the second DEM data using the GAN generative adversarial network to obtain the third DEM data.

[0042] The third module is used to perform detail enhancement on the third DEM data using a content generation algorithm to obtain the fourth DEM data.

[0043] The fourth module is used to de-encrypt the second city's three-dimensional building model data using the fourth DEM data to obtain the de-encryption result.

[0044] Another aspect of the present invention provides an electronic device, including a processor and a memory;

[0045] The memory is used to store programs;

[0046] The processor executes the program to implement the method as described above.

[0047] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the methods described above.

[0048] The beneficial effects of this invention are as follows: The technical solution of this invention no longer directly transforms the location information of the building model itself. Instead, it uses a generative adversarial network model to learn the features of the underlying DEM carrying the building model, generating an encrypted DEM that is similar in macroscopic topography but completely different in elevation details. Then, the original 3D building model is translated as a whole onto this new virtual surface. That is, the elevation information of the urban 3D building model is encrypted by encrypting the background DEM, which changes the absolute elevation benchmark of the buildings. The height, size, geometric details, and relative spatial relationships between buildings of the building model itself can be well preserved. Thus, while achieving effective decryption, the availability of data is maximized, providing a solution to the current problem of secure sharing of 3D geographic information. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the urban 3D building model decryption process according to an embodiment of the present invention;

[0050] Figure 2 This is a schematic diagram of the DEM data encryption process according to an embodiment of the present invention;

[0051] Figure 3 This is a schematic diagram of the urban 3D building model decryption process according to an embodiment of the present invention;

[0052] Figure 4 This is a schematic diagram of a city 3D model decryption algorithm based on DEM terrain encryption according to an embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram of a de-decryption device for urban three-dimensional building models according to an embodiment of the present invention. Detailed Implementation

[0054] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings. Throughout the description, the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions. In the following description, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" can be used interchangeably. Terms such as "first," "second," etc., are used only to distinguish technical features and should not be construed as indicating or implying relative importance, or implicitly indicating the number of indicated technical features, or implicitly indicating the sequential relationship of the indicated technical features. In the following description, the consecutive reference numerals for method steps are for ease of review and understanding. Adjusting the implementation order of steps, in conjunction with the overall technical solution of the present invention and the logical relationship between the various steps, will not affect the technical effect achieved by the technical solution of the present invention. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0055] refer to Figure 1 ,in Figure 1 This is a schematic diagram of the decryption process for a 3D urban building model, which includes, but is not limited to, steps S100~S400:

[0056] S100: Preprocess the first DEM data and the first city 3D building model data of the area to be de-densified to obtain the second DEM data and the second city 3D building model data.

[0057] The first DEM data is the original, real DEM (Digital Elevation Model) data, and the second DEM data is the DEM data aligned with the first city 3D building model data (the original city 3D building model data).

[0058] In some embodiments, a GIS or programming library is used to align the first DEM data and the first city 3D building model data, wherein the alignment method includes coordinate system transformation and projection unification; the aligned first DEM data and the first city 3D building model data are preprocessed to obtain second DEM data and second city 3D building model data, wherein the preprocessing includes filling in missing values ​​and removing obvious erroneous elevation points or abnormal elevation points.

[0059] In some embodiments, the first DEM data, the DEM terrain data, is high-precision DEM data (accuracy better than 2m) or high-precision DEM data generated from aerial LiDAR laser scanning point cloud data, and the urban 3D building model data is in OBJ format.

[0060] S200: A GAN (Generative Adversarial Network) is constructed and trained using the second DEM data. The GAN is then used to encrypt the second DEM data to obtain the third DEM data.

[0061] In some embodiments, reference Figure 2 The diagram showing the DEM data encryption process includes, but is not limited to, steps S210 to S240:

[0062] S210, slice the second DEM data to obtain DEM tiles, normalize the DEM tiles to obtain the training set.

[0063] In some embodiments, a large-scale DEM map is cut into a series of DEM tiles at a preset size (e.g., 256×256) as training samples for a neural network.

[0064] In some embodiments, the elevation values ​​in each DEM tile are linearly normalized to the interval [-1, 1] to adapt to the output range of the neural network activation function and improve training stability.

[0065] S220 constructs a GAN (Generative Adversarial Network), which includes a generator with a U-Net structure and a discriminator with a deep convolutional neural network.

[0066] In some embodiments, the GAN generator uses the encoder-decoder structure of U-Net, with input being a random noise vector (standard GAN) or a real DEM tile (cGAN), and output being an encrypted DEM tile of the same size as the input. The discriminator uses a deep convolutional neural network (DCNN), with input being a DEM tile and output being a scalar value used to determine whether the tile is real or generated by the generator.

[0067] Specifically, a Wasserstein GAN with Gradient Penalty (WGAN-GP) model is constructed to improve the quality and diversity of generated terrain. The generator adopts the U-Net architecture. Its input is a 100-dimensional random noise vector z, which is upsampled into a 256×256 single-channel image, i.e., a encrypted DEM tile, through a series of transposed convolutional layers and ReLU activation functions. The last layer uses the Tanh activation function to ensure that the output value is in the range [-1,1]. The discriminator adopts a standard deep convolutional neural network (DCNN). Its input is a 256×256 DEM tile (real or generated), which is processed by a series of convolutional layers with LeakyReLU activation functions for feature extraction and downsampling, and finally outputs a scalar value (Critic Score) to evaluate the realism of the input tile.

[0068] S230 uses the WGAN-GP loss function and training strategy to train the generator and discriminator of the GAN generative adversarial network for adversarial game training on the training set.

[0069] In some embodiments, the loss function and training strategy of WGAN-GP are adopted, and the generator and discriminator engage in adversarial game until the model converges. The resulting generator has the ability to generate high-quality and diverse encrypted DEM tiles.

[0070] S240, the second DEM data is segmented using a sliding window, and each sliding window is processed by a trained generator of a GAN (Generative Adversarial Network) to obtain encrypted DEM tiles. The encrypted DEM tiles are then weighted and averaged to obtain the third DEM data.

[0071] The third DEM data is the encrypted DEM data.

[0072] In some embodiments, the complete real DEM map is divided into sliding windows of 256×256 pixels. Each window region is input into a trained generator to generate corresponding encrypted DEM tiles. Finally, all generated tiles are weighted and averaged in overlapping areas to create a complete encrypted DEM with the same extent as the original DEM, similar macroscopic geomorphic features, and no obvious seams.

[0073] S300, the third DEM data is enhanced with a content generation algorithm to obtain the fourth DEM data.

[0074] In some embodiments, to compensate for the potential local smoothing problem in terrain generated by GANs and to further increase its randomness, the PCG algorithm is used for detail enhancement, including:

[0075] Perlin noise or Simplex noise algorithms are used to generate multi-level frequency-harmonic fractal noise maps; the fractal noise maps are then weighted and fused with the third DEM data to obtain the fourth DEM data, where the weighted fusion is expressed as follows:

[0076] ;

[0077] in, This is the fourth DEM data. This is the third DEM data. This is a fractal noise diagram. The detail enhancement factor is an adjustable hyperparameter, typically ranging from 0.05 to 0.2. In this embodiment... This is to ensure that the original macroscopic topography is not destroyed while enriching the details.

[0078] In the multi-layered frequency-multiplying fractal noise diagram, the frequency of each layer of noise is twice that of the previous layer, and the amplitude is 0.5 times that of the previous layer. These multi-layered noise diagrams are superimposed according to amplitude weighting to form fractal noise with rich layering and self-similarity.

[0079] S400 uses the fourth DEM data to de-declassify the three-dimensional building model data of the second city, and obtains the de-declassification result.

[0080] In some embodiments, reference Figure 3 The diagram shown illustrates the process of declassifying a 3D urban building model, including but not limited to steps S410-S430:

[0081] S410, calculate the elevation offset of each individual building in the second city's 3D building model data;

[0082] Using bilinear interpolation, the elevation value of the center point is queried on both the original DEM (first DEM data) and the final encrypted DEM (fourth DEM data). The formula for calculating the elevation offset is:

[0083] ;

[0084] in, This represents the building's elevation offset. The elevation values ​​of the buildings in the first DEM data. The elevation values ​​of the buildings in the fourth DEM data. The average position of the geometric center point or the vertex of the foundation of the building;

[0085] S420 updates the geometric vertices of all building models in the second city's 3D building model data.

[0086] The geometric vertices of the architectural model The update method is as follows:

[0087] ;

[0088] in, For the updated vertices coordinate, The vertex before the update coordinate;

[0089] S430, repeatedly execute the elevation offset and geometric vertices of the building model until all the vertices of all buildings in the second city's 3D building model data are updated, and the de-dense result is obtained.

[0090] Embodiments of the present invention translate the entire building as a rigid body in the vertical direction, so that it is precisely located on the new encrypted surface.

[0091] Embodiments of the present invention further include extracting the base polygon of each independent building and performing the following processing based on the area of ​​the base polygon: if the area of ​​the base polygon is less than or equal to a preset area, the overall elevation offset is directly calculated; if the area of ​​the base polygon is greater than the preset area, multiple points of the base polygon are randomly sampled, and the elevation offset of each sample point is calculated. The average or median value of the elevation offset of the sample points is used as the overall elevation offset.

[0092] In some embodiments, the preset area is customized according to the size distribution of urban buildings.

[0093] This method avoids tilting of the building model due to single-point selection bias on sloping surfaces, thus improving the robustness of elevation offset calculation.

[0094] refer to Figure 4 ,in Figure 4 The embodiment is Figure 1 A decryption algorithm for urban 3D models based on DEM terrain encryption is described, and the steps are as described in the above embodiments.

[0095] Figure 5 This is a schematic diagram of a de-decryption device for urban three-dimensional building models according to an embodiment of the present invention. The device includes a first module 510, a second module 520, a third module 530, and a fourth module 540.

[0096] The system comprises the following modules: a first module for preprocessing the first DEM data and the first city 3D building model data of the area to be declassified, to obtain second DEM data and second city 3D building model data; a second module for constructing and training a GAN generative adversarial network using the second DEM data, and using the GAN to encrypt the second DEM data, to obtain third DEM data; a third module for performing detail enhancement on the third DEM data using a content generation algorithm, to obtain fourth DEM data; and a fourth module for declassifying the second city 3D building model data using the fourth DEM data, to obtain the declassified result.

[0097] Exemplarily, with the cooperation of the first, second, third, and fourth modules in the device, the embodiment device can implement any of the aforementioned methods for declassifying urban 3D building models, namely, preprocessing the first DEM data and the first urban 3D building model data of the area to be declassified to obtain second DEM data and second urban 3D building model data; constructing and training a GAN generative adversarial network using the second DEM data; encrypting the second DEM data using the GAN generative adversarial network to obtain third DEM data; performing detail enhancement on the third DEM data using a content generation algorithm to obtain fourth DEM data; and declassifying the second urban 3D building model data using the fourth DEM data to obtain the declassification result. The beneficial effects of this invention are as follows: By using a generative adversarial network model to learn the features of the underlying DEM carrying the building model, an encrypted DEM with similar macroscopic topography but completely different elevation details is generated. Then, the original 3D building model is translated onto this new virtual surface. That is, by encrypting the background DEM, the elevation information of the urban 3D building model is encrypted, changing the absolute elevation benchmark of the buildings. The height, size, geometric details, and relative spatial relationships between buildings of the building model itself can be well preserved. Thus, while achieving effective decryption, the availability of data is maximized, providing a solution to the current problem of secure sharing of 3D geographic information.

[0098] This invention also provides an electronic device, which includes a processor and a memory;

[0099] The memory stores the program;

[0100] The processor executes a program to perform the aforementioned method for decrypting urban 3D building models; the electronic device has the function of carrying and running the software system for decrypting urban 3D building models provided in the embodiments of the present invention, such as a personal computer, minicomputer, mainframe, workstation, network or distributed computing environment, standalone or integrated computer platform, or communicating with charged particle tools or other imaging devices, etc.

[0101] This invention also provides a computer-readable storage medium storing a program that is executed by a processor to implement the urban 3D building model decryption method described above.

[0102] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented in the embodiments of this invention. Alternative embodiments are contemplated, in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0103] This invention also discloses a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium and execute the computer instructions, causing the computer device to perform the aforementioned method for decrypting urban three-dimensional building models.

[0104] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the described functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, considering the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed in the embodiments of the invention, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0105] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0106] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can include, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0107] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0108] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0109] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0110] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0111] The above is a detailed description of the preferred embodiments of the present invention, but the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention, and these equivalent modifications or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for decrypting urban three-dimensional building models, characterized in that, include: Preprocess the first DEM data and the first city 3D building model data of the area to be de-densified to obtain the second DEM data and the second city 3D building model data. A GAN (Generative Adversarial Network) is constructed and trained using the second DEM data. The second DEM data is then encrypted using the GAN to obtain the third DEM data. The third DEM data is enhanced with a content generation algorithm to obtain the fourth DEM data; The second city's three-dimensional building model data is de-encrypted using the fourth DEM data to obtain the de-encryption result; The process involves constructing and training a GAN (Generative Adversarial Network) using the second DEM data, and then encrypting the second DEM data using the GAN to obtain the third DEM data, including: The second DEM data is sliced ​​to obtain DEM tiles, and the DEM tiles are normalized to obtain a training set; Construct a Generative Adversarial Network (GAN), including a generator with a U-Net structure and a discriminator with a deep convolutional neural network; The generator and discriminator of the GAN (Generative Adversarial Network) are trained for adversarial game using the loss function and training strategy of WGAN-GP on the training set. The second DEM data is segmented using a sliding window. Each sliding window is processed by a trained generator of a GAN (Generative Adversarial Network) to obtain encrypted DEM tiles. The encrypted DEM tiles are then weighted and averaged to obtain the third DEM data. The process of enhancing the details of the third DEM data using a content generation algorithm to obtain the fourth DEM data includes: The Perlin noise or Simplex noise algorithm is used to generate a multi-level frequency harmonic fractal noise map; The fractal noise map is weighted and fused with the third DEM data to obtain the fourth DEM data, wherein the weighted fusion is expressed as follows: in, This is the fourth DEM data. This is the third DEM data. This is a fractal noise diagram. For detail enhancement factor, The value is usually between 0.05 and 0.

2.

2. The method for decrypting urban three-dimensional building models according to claim 1, characterized in that, The first DEM data and the first city 3D building model data of the aligned de-densification area are preprocessed to obtain second DEM data and second city 3D building model data, including: The first DEM data and the first city 3D building model data are aligned using a GIS or programming library, wherein the alignment method includes coordinate system transformation and projection unification. The aligned first DEM data and the first city 3D building model data are preprocessed to obtain the second DEM data and the second city 3D building model data. The preprocessing includes filling in missing values ​​and removing obvious erroneous or abnormal elevation points.

3. The method for decrypting urban three-dimensional building models according to claim 1, characterized in that, The precision of the first DEM data is less than 2 meters. The first DEM data is high-precision DEM data generated from airborne LiDAR laser scanning point cloud data.

4. The method for decrypting urban three-dimensional building models according to claim 1, characterized in that, The process of declassifying the second city's 3D building model data using the fourth DEM data to obtain the declassification result includes: Calculate the elevation offset of each individual building in the second city's 3D building model data: in, This represents the building's elevation offset. The elevation values ​​of the buildings in the first DEM data. The elevation values ​​of the buildings in the fourth DEM data. The average position of the geometric center point or the vertex of the foundation of the building; For the geometric vertices of all building models in the second city's 3D building model data, for each building model's geometric vertices... Update: in, For the updated vertices coordinate, The vertex before the update coordinate; Repeat the process of adjusting the elevation offset and geometric vertices of the building model until all the vertices of all buildings in the second city's 3D building model data are updated to obtain the de-dense result.

5. The method for decrypting urban three-dimensional building models according to claim 4, characterized in that, The calculation of the elevation offset of each individual building in the second city 3D building model data also includes: Extract the base polygon for each individual building, and perform the following processing based on the area of ​​the base polygon: If the area of ​​the base polygon is less than or equal to the preset area, the overall elevation offset is calculated directly. If the area of ​​the base polygon is greater than the preset area, multiple points of the base polygon are randomly sampled. The elevation offset of each sample point is calculated, and the average or median value of the elevation offset of the sample points is used as the overall elevation offset.

6. A device for declassifying urban three-dimensional building models, characterized in that, include: The first module is used to preprocess the first DEM data and the first city 3D building model data of the area to be de-densified to obtain the second DEM data and the second city 3D building model data. The second module is used to construct and train a GAN generative adversarial network using the second DEM data, and to encrypt the second DEM data using the GAN generative adversarial network to obtain the third DEM data. The third module is used to perform detail enhancement on the third DEM data using a content generation algorithm to obtain the fourth DEM data. The fourth module is used to de-encrypt the second city's three-dimensional building model data using the fourth DEM data to obtain the de-encryption result; The process involves constructing and training a GAN (Generative Adversarial Network) using the second DEM data, and then encrypting the second DEM data using the GAN to obtain the third DEM data, including: The second DEM data is sliced ​​to obtain DEM tiles, and the DEM tiles are normalized to obtain a training set; Construct a Generative Adversarial Network (GAN), including a generator with a U-Net structure and a discriminator with a deep convolutional neural network; The generator and discriminator of the GAN (Generative Adversarial Network) are trained for adversarial game using the loss function and training strategy of WGAN-GP on the training set. The second DEM data is segmented using a sliding window. Each sliding window is processed by a trained generator of a GAN (Generative Adversarial Network) to obtain encrypted DEM tiles. The encrypted DEM tiles are then weighted and averaged to obtain the third DEM data. The process of enhancing the details of the third DEM data using a content generation algorithm to obtain the fourth DEM data includes: The Perlin noise or Simplex noise algorithm is used to generate a multi-level frequency harmonic fractal noise map; The fractal noise map is weighted and fused with the third DEM data to obtain the fourth DEM data, wherein the weighted fusion is expressed as follows: in, This is the fourth DEM data. This is the third DEM data. This is a fractal noise diagram. For detail enhancement factor, The value is usually between 0.05 and 0.

2.

7. An electronic device, characterized in that, Including the processor and memory; The memory is used to store programs; The processor executes the program to implement the method for decrypting urban three-dimensional building models as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The storage medium stores a program, which is executed by a processor to implement the method for decrypting urban three-dimensional building models as described in any one of claims 1-5.

Images