A method for integrating modeling of shear wall residential enclosure and interior space by fusing semantic segmentation recognition and gan model prediction

Abstract

The application provides a shear wall residential enclosure and internal space integrated modeling method fusing semantic segmentation identification and GAN model prediction. The method obtains building wall outer contour lines and top traffic space contour lines and each facade elevation view. After identifying the elevation view of a certain facade, a window opening is constructed and matched to the outer contour line of the facade, so as to obtain an outer contour plane. The outer contour plane is extended to multiple layers to obtain a building enclosure model. The outer contour plane is preprocessed, a GAN model is applied to predict the preprocessed plane, the predicted plane is processed and coordinates are extracted, a single-story building internal model is constructed, the single-story building internal model is extended to multiple layers, and finally a three-dimensional model is obtained. Through the above method, a three-dimensional building model containing an indoor model can be quickly constructed, the modeling cost is greatly reduced, the modeling process is simplified, and the method has important significance for promoting urban informatization construction.

Description

Technical Field

[0001] This invention relates to the field of 3D model construction and digital twin technology, and in particular to an integrated modeling method for shear wall residential building envelope and interior space that combines semantic segmentation recognition and GAN model prediction. Background Technology

[0002] With the continuous advancement of urbanization and digitalization, Building Information Modeling (BIM) has been widely applied in various fields such as urban renewal, residential area renovation, environmental assessment, and smart operation and maintenance. By digitally modeling actual buildings, simulation analysis can be performed conveniently and quickly, providing important support for decision-making and management, construction and renewal, and smart city construction. In particular, it plays a significant role in City Information Modeling (CIM) and is of great importance to promoting urban informatization.

[0003] The construction of existing 3D building models mostly adopts traditional 3D modeling based on drawings, that is, processing the building information according to the design drawings. This method has good accuracy when building models, but it requires a lot of manpower and resources, and the management of drawings becomes difficult over time, making it difficult to build models of existing buildings.

[0004] Besides traditional 3D modeling based on drawings, there are also 3D model building methods based on images and point clouds. These methods often require high-precision equipment, are very costly, and require accurate operation by the operator. Error messages can significantly impact the model's quality. Furthermore, the constructed building models only include the building envelope and lack interior models.

[0005] Therefore, it is necessary to propose an integrated modeling method for shear wall residential building envelope and interior space that combines semantic segmentation recognition and GAN model prediction, so as to quickly establish a building envelope model and derive the building interior model from it, thereby achieving low-cost construction of the building 3D model. Summary of the Invention

[0006] The purpose of this invention is to solve the problems of high modeling cost and difficulty in deriving indoor models in existing technologies. It proposes an integrated modeling method for shear wall residential building envelope and interior space that combines semantic segmentation recognition and GAN model prediction.

[0007] This invention is achieved through the following technical solution: This invention proposes an integrated modeling method for shear wall residential building envelopes and internal spaces that combines semantic segmentation recognition with GAN model prediction. The method includes: Step 1: Obtain the horizontal outline of the building walls and the top traffic space outline of the building to be modeled; Step 2: Obtain the front view of each facade of the building to be modeled; Step 3: Identify the front view of a certain facade, obtain the corresponding information to construct window openings, match them to the outer contour line of the facade, and then obtain the outer contour plane containing the window opening information of each facade. Step 4: Based on the vertical direction information of the window openings obtained in Step 3, the outer contour plane containing the window opening information of each facade is extended to multiple layers to obtain the building's external envelope model. Step 5: Match the top traffic space outline to the outer outline plane obtained in Step 3. Based on the architectural characteristics of shear wall residential buildings, apply relevant knowledge in the field of architectural design to perform preset processing on the outer outline plane. Step 6: Predict the pre-processed plane to obtain the predicted plane; Step 7: Process the predicted plane and extract its coordinates to construct a single-story building interior model; Step 8: Using the vertical orientation information of the window openings obtained in Step 3, the single-story building interior model constructed in Step 7 is expanded to multiple stories, ultimately obtaining a three-dimensional model that includes the building's external envelope and interior.

[0008] Furthermore, the horizontal outline of the building's outer wall and the top traffic space outline are obtained by acquiring them from open-source websites or through drone modeling. The outer wall outline and the top traffic space outline are regularized and ensured to be on the same horizontal plane for subsequent processing. They are regularized by wrapping them with a regular rectangle and continuously approximating them inward.

[0009] Furthermore, the front views of each facade of the building to be modeled are obtained by correcting the tilted images of the building facades to obtain or crop the front views of the model obtained from the UAV modeling.

[0010] Furthermore, the identification of window openings in the front view of a certain facade is achieved through semantic segmentation; the identified image is then detected row by row and column by column in both horizontal and vertical directions, and the row and column with the best recognition effect are selected as representatives; in the horizontal direction, the color change points are the junctions between the window openings and the exterior walls; in the vertical direction, the distance between the centers of adjacent windows is the single-story height of the building, and the total number of window openings is the total number of floors of the building, with the window openings vertically centered on the wall.

[0011] Furthermore, in the process of matching window openings with the outer contour line, the representative window opening wall of a certain facade is first matched to the horizontal straight line length of the outer contour line of the facade, and the window openings are scaled up to achieve the actual length matching on this vertical facade; the scaled window openings are then horizontally and vertically aligned to the line segment of the outer contour line of the facade to complete the matching of window openings with the outer contour line of a certain facade, thereby obtaining the outer contour plane containing the window openings of each facade.

[0012] Furthermore, considering the architectural characteristics of shear wall residential buildings, the rules for the pre-planned layout include at least the following: (1) There will be a wall extending between horizontally adjacent exterior windows on the same vertical plane; (2) There will be a wall extending at the turning point of the outer contour line of the wall; (3) There will be a wall extending at the turning point of the top traffic space contour line; (4) The extended wall will not exceed the outer contour line of the wall, and will not connect with the exterior windows or cross the top traffic space contour line.

[0013] Furthermore, the prediction of the preset plane is achieved through a GAN model, which is trained on several datasets corresponding to the preset plane and the real plane. The optimal prediction effect is achieved by adjusting the model training parameters.

[0014] Furthermore, the prediction plane is processed and its coordinates are extracted using OpenCV. The prediction image is classified using HSV color segmentation. The geometric characteristics of each category are vectorized and imported into the modeling software for subsequent operations.

[0015] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the integrated modeling method for shear wall residential building envelope and internal space that integrates semantic segmentation recognition and GAN model prediction.

[0016] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the integrated modeling method for shear wall residential building envelope and internal space that integrates semantic segmentation recognition and GAN model prediction.

[0017] The beneficial effects of this invention are: (1) This invention proposes an integrated modeling method for shear wall residential building envelope and internal space that combines semantic segmentation recognition and GAN model prediction. The modeling method has a simple and easy-to-operate overall process and is highly applicable and predictive for shear wall residential buildings. Compared with the traditional three-dimensional building model construction process, it is greatly simplified.

[0018] (2) A method for rapidly constructing 3D building models including interior models has been achieved, and a method for constructing building envelope models for general buildings has also been proposed. The 3D building models constructed by the method described in this invention are richer in content and have better accuracy in interior models compared to building models obtained by modeling methods such as point cloud data and a large number of images.

[0019] (3) Traditional building 3D model construction requires a lot of manpower and material resources. This invention is based on open source data and real building scenes. Through deep learning, reasonable reasoning can be used to obtain a 3D model including the building interior, which greatly reduces the modeling cost and is of great significance to promoting urban informatization. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0021] Figure 1 This is a flowchart of a method for integrated modeling of shear wall residential building envelope and internal space that combines semantic segmentation recognition and GAN model prediction, as described in this invention.

[0022] Figure 2 This is an example diagram illustrating the identification of a front view window opening on a facade through semantic segmentation in an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram illustrating the matching of the window opening and the outer contour line in an embodiment of the present invention.

[0024] Figure 4 This is an example diagram of a building envelope structure model constructed in an embodiment of the present invention.

[0025] Figure 5 This is a sample plan view obtained by applying knowledge from the field of architectural design to the characteristics of shear wall residential buildings in this embodiment of the invention.

[0026] Figure 6 This is an example diagram of the predicted plane obtained by using GAN to predict a preset plane in an embodiment of the present invention.

[0027] Figure 7 In this embodiment of the invention, the internal model of a single-story building is extended to multiple stories, resulting in an example of a three-dimensional model that includes both the building's external envelope and its interior.

[0028] Figure 8 This is a real-world example image of a residential building with shear walls that is to be modeled. Detailed Implementation

[0029] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0030] Specifically, in combination Figures 1-8 This invention proposes an integrated modeling method for shear wall residential building envelopes and internal spaces that combines semantic segmentation recognition with GAN model prediction. The method includes: Step 1: Obtain the horizontal outline of the building walls and the top traffic space outline of the building to be modeled; Step 2: Obtain the front view of each facade of the building to be modeled; Step 3: Identify the front view of a certain facade, obtain the corresponding information to construct window openings, match them to the outer contour line of the facade, and then obtain the outer contour plane containing the window opening information of each facade. Step 4: Based on the vertical direction information of the window openings obtained in Step 3, the outer contour plane containing the window opening information of each facade is extended to multiple layers to obtain the building's external envelope model. Step 5: Match the top traffic space outline to the outer outline plane obtained in Step 3. Based on the architectural characteristics of shear wall residential buildings, apply relevant knowledge in the field of architectural design to perform preset processing on the outer outline plane. Step 6: Predict the pre-processed plane to obtain the predicted plane; Step 7: Process the predicted plane and extract its coordinates to construct a single-story building interior model; Step 8: Using the vertical orientation information of the window openings obtained in Step 3, the single-story building interior model constructed in Step 7 is expanded to multiple stories, ultimately obtaining a three-dimensional model that includes the building's external envelope and interior.

[0031] Furthermore, there are multiple ways to obtain the horizontal outline of the building's walls and the outline of its overhead traffic space. These can be obtained from open-source websites or through drone modeling. The outer and inner outlines need to be regularized and ensured to be on the same horizontal plane for subsequent processing. Regularization can be achieved by wrapping the outline with a regular rectangle and continuously approximating it inwards.

[0032] Furthermore, there are multiple ways to obtain the front view of each facade of the building to be modeled. It can be obtained by correcting tilted images of building facades obtained through various methods, or by cropping the front view of the model obtained by drone modeling.

[0033] Furthermore, the identification of window openings in a facade's front view can be achieved through semantic segmentation. The identified image is then inspected row by row and column by column in both horizontal and vertical directions, with the row and column showing the best recognition results selected as representatives. In the horizontal direction, the areas where color changes indicate the junction between the window opening and the exterior wall; in the vertical direction, the distance between the centers of adjacent windows represents the building's single-story height, and the total number of window openings represents the total number of stories. Window openings are generally vertically centered within the wall.

[0034] Furthermore, in the process of matching window openings with the outer contour line, it is first necessary to match the representative window opening wall of a certain facade to the horizontal straight line length of the outer contour line of that facade, and then scale the window openings accordingly to achieve a true length match on this vertical facade. The scaled window openings are then horizontally and vertically aligned to the outer contour line segment of the facade to complete the matching of window openings with the outer contour line of a certain facade, thereby obtaining the outer contour plane containing the window openings of each facade.

[0035] Furthermore, steps 1-4 can also be used to construct building envelope models for other types of buildings besides residential buildings. The method for constructing building envelope models for predicting the internal floor plan of residential buildings should not be construed as a limitation of the present invention, but should also fall within the scope of protection of the present invention.

[0036] Furthermore, considering the architectural characteristics of shear wall residential buildings, the rules for the pre-planned layout include at least the following: (1) There will be a wall extending between horizontally adjacent exterior windows on the same vertical plane; (2) There will be a wall extending at the turning point of the outer contour line of the wall; (3) There will be a wall extending at the turning point of the top traffic space contour line; (4) The extended wall will not exceed the outer contour line of the wall, and will not connect with the exterior windows or cross the top traffic space contour line.

[0037] Furthermore, the prediction of the preset plane can be achieved through a GAN model. The GAN model is trained by several sets of datasets corresponding to the preset plane and the real plane. The optimal prediction effect can be achieved by adjusting the model training parameters.

[0038] Furthermore, the prediction plane can be processed and its coordinates extracted using OpenCV. The prediction image can be classified using HSV color segmentation, and the geometric characteristics of each category can be vectorized and imported into the modeling software for subsequent operations.

[0039] This invention proposes an integrated modeling method for the envelope and interior space of shear wall residential buildings, combining semantic segmentation and recognition with GAN model prediction. The method acquires the building's external wall outline, top traffic space outline, and front views of each facade using open-source data and real-world building scenes. After recognizing the front view of a particular facade, window openings are constructed and matched to the external outline facade, resulting in an external outline plane containing window opening information for each facade. Based on the vertical information of the window openings, the external outline plane is extended to multiple floors, yielding a building envelope model. Considering the architectural characteristics of shear wall residential buildings, relevant knowledge from the field of architectural design is applied, combined with the top traffic space outline, to pre-process the external outline plane. A GAN model is then used to predict the pre-processed plane, which is then processed and its coordinates extracted to construct a single-story building interior model. Based on the vertical information of the window openings, the single-story building interior model is extended to multiple floors, ultimately obtaining a 3D model containing both the building envelope and interior. This method enables rapid construction of 3D building models including interior models, significantly reducing modeling costs and simplifying the modeling process, which is of great significance for promoting urban informatization.

[0040] Example 1 This invention proposes an integrated modeling method for shear wall residential building envelopes and internal spaces that combines semantic segmentation recognition with GAN model prediction. The method includes: (1) Obtain the outer contour lines of the building walls and the top traffic space contour lines and the front views of each facade from open source data and real building scenes. The outer contour lines and contour lines need to be regularized and ensured to be on the same horizontal plane for subsequent operations. (2) After identifying the front view of a certain facade, construct the window openings and match them to the outer contour line of the facade to obtain the outer contour plane containing the window opening information of each facade. Based on the vertical information of the window openings, extend the outer contour plane to multiple floors to obtain the building envelope structure model. The identification of the window openings in the front view of a certain facade can be achieved through semantic segmentation; (3) Based on the architectural characteristics of shear wall residential buildings, relevant knowledge in the field of architectural design is applied, and the outline of the top traffic space is combined with the pre-processing of the outer outline plane. The GAN model is used to predict the pre-processed plane. The GAN model is trained by several sets of datasets corresponding to the pre-processed plane and the real plane. The optimal prediction effect is achieved by adjusting the model training parameters. (4) The predicted plane is processed and the coordinates are extracted by OpenCV to construct a single-story building interior model. Based on the vertical information of the window openings, the single-story building interior model is extended to multiple floors, and finally a three-dimensional model containing the building's external envelope and interior is obtained.

[0041] Example 2 In this embodiment, a single shear wall residential building in an old community is used as an example to illustrate the integrated modeling method for the enclosure structure and internal space of a shear wall residential building that integrates semantic segmentation recognition and GAN model prediction.

[0042] (I) Modeling Scenarios and Data Acquisition A single shear wall residential building in an old residential area was selected as the modeling object. The building is a slab-type structure. A drone was used to take pictures to obtain the various facades of the building. The outlines of the outer walls and the top traffic space were obtained from an open source website. After regularization, they were aligned to the same horizontal plane for subsequent operations.

[0043] (ii) Facade identification and matching The front view of a residential facade is mapped onto the outer contour of the wall surface. Semantic segmentation is used for identification, and the identified window openings are matched to the horizontal straight line length of the outer contour facade. The window openings are then scaled to achieve a true length match on this vertical facade. Horizontally, the scaled window openings are vertically aligned to the line segment of the outer contour facade to complete the matching of window openings to the outer contour of a particular facade, thus obtaining an outer contour plane containing the window openings of all facades.

[0044] (III) Construction of Building Envelope Model During the recognition of a front view of a facade, detection is performed in the vertical direction. The center distance between adjacent windows is the single-story height information of the building, and the total number of windows in a vertical direction is the total number of floors of the building. Based on the vertical information of the window openings, a model of the building's external envelope structure is constructed.

[0045] (iv) Preset Plane Generation and Prediction The building's outer contour plane containing window opening location information is pre-processed. Considering the characteristics of shear wall residential buildings, the pre-defined plane rules include at least: (1) a wall extends between horizontally adjacent windows on the same vertical plane; (2) a wall extends at the turning point of the outer contour line; (3) a wall extends at the turning point of the top traffic space contour line; and (4) the extended wall does not exceed the outer contour line of the wall, does not connect with the windows, and does not cross the top traffic space contour line. The pre-defined plane is obtained based on these rules, and a predicted plane is obtained by predicting the pre-defined plane using a trained GAN model.

[0046] (V) Construction and Expansion of 3D Models The predicted plane is processed to obtain the corresponding coordinate information, which is then identified and constructed in the modeling software. The height of a single-layer model is the straight-line distance between two adjacent windows during the window opening identification process. The total number of layers in the model can also be obtained during the window opening identification process. The single-layer building interior model is then expanded to multiple layers to finally obtain a three-dimensional model that includes the building's external envelope and interior, thus completing the modeling process.

[0047] (vi) Explanation of Implementation Results The method described in this embodiment utilizes drone-captured images and deep learning for inference to rapidly model shear wall residential buildings. Compared to traditional modeling methods, the method of this invention is simpler, less costly, and capable of deriving interior details. Experimental results show that the modeling effect is good for a single shear wall residential building in an old community. The method of this invention can construct high-quality, high-precision building models with a simpler process and lower cost, thereby effectively promoting urban informatization.

[0048] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the integrated modeling method for shear wall residential building envelope and internal space that integrates semantic segmentation recognition and GAN model prediction.

[0049] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the integrated modeling method for shear wall residential building envelope and internal space that integrates semantic segmentation recognition and GAN model prediction.

[0050] The memory in this application embodiment can be volatile memory or non-volatile memory, or it can include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DR RAM). It should be noted that the memory used in the methods described in this invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0051] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).

[0052] In implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software. The steps of the method disclosed in the embodiments of this application can be directly implemented by a hardware processor, or by a combination of hardware and software modules in the processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, detailed descriptions are omitted here.

[0053] It should be noted that the processor in the embodiments of this application can be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method embodiments can be completed by the integrated logic circuitry in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied as execution by a hardware decoding processor, or as a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above methods.

[0054] The above provides a detailed description of the integrated modeling method for shear wall residential building envelope and internal space that combines semantic segmentation recognition and GAN model prediction. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A method for integrated modeling of shear wall residential building envelope and internal space by combining semantic segmentation recognition and GAN model prediction, characterized in that, The method includes: Step 1: Obtain the horizontal outline of the building walls and the top traffic space outline of the building to be modeled; Step 2: Obtain the front view of each facade of the building to be modeled; Step 3: Identify the front view of a certain facade, obtain the corresponding information to construct window openings, match them to the outer contour line of the facade, and then obtain the outer contour plane containing the window opening information of each facade. Step 4: Based on the vertical direction information of the window openings obtained in Step 3, the outer contour plane containing the window opening information of each facade is extended to multiple layers to obtain the building's external envelope model. Step 5: Match the top traffic space outline to the outer outline plane obtained in Step 3. Based on the architectural characteristics of shear wall residential buildings, apply relevant knowledge in the field of architectural design to perform preset processing on the outer outline plane. Step 6: Predict the pre-processed plane to obtain the predicted plane; Step 7: Process the predicted plane and extract its coordinates to construct a single-story building interior model; Step 8: Using the vertical orientation information of the window openings obtained in Step 3, the single-story building interior model constructed in Step 7 is expanded to multiple stories, ultimately obtaining a three-dimensional model that includes the building's external envelope and interior.

2. The method according to claim 1, characterized in that, The horizontal outlines of the building walls and the top traffic space are obtained by acquiring them from open-source websites or through drone modeling. The outlines of the walls and the top traffic space are regularized and ensured to be on the same horizontal plane for subsequent processing. They are regularized by wrapping them with a regular rectangle and continuously approximating them inward.

3. The method according to claim 1, characterized in that, The front views of each facade of the building to be modeled are obtained by correcting the tilted images of the building facades to obtain or crop the front views of the model obtained from the drone modeling.

4. The method according to claim 1, characterized in that, The identification of window openings in the front view of a facade is achieved through semantic segmentation. The identified image is then inspected row by row and column by column in both horizontal and vertical directions, and the row and column with the best recognition performance are selected as representatives. In the horizontal direction, the color change points indicate the junction between the window opening and the exterior wall. In the vertical direction, the distance between the centers of adjacent windows is the single-story height of the building, and the total number of window openings is the total number of stories of the building. The window openings are vertically centered on the wall.

5. The method according to claim 1, characterized in that, In the process of matching window openings with the outer contour line, the representative window opening wall of a certain facade is first matched to the horizontal straight line length of the outer contour line of the facade, and the window openings are scaled up to achieve the actual length matching on this vertical facade. The scaled window openings are then horizontally and vertically aligned to the line segment of the outer contour line of the facade to complete the matching of window openings with the outer contour line of a certain facade, thereby obtaining the outer contour plane containing the window openings of each facade.

6. The method according to claim 1, characterized in that, In view of the architectural characteristics of shear wall residential buildings, the rules of the pre-plan include at least (1) there will be a wall extending between the horizontally adjacent exterior windows on the same vertical plane; (2) there will be a wall extending at the turning point of the outer contour line of the wall; (3) there will be a wall extending at the turning point of the top traffic space contour line; (4) the extended wall will not exceed the outer contour line of the wall, and will not connect with the exterior windows, nor will it cross the top traffic space contour line.

7. The method according to claim 1, characterized in that, The prediction of the preset plane is achieved through the GAN model, which is trained by several sets of datasets corresponding to the preset plane and the real plane. The optimal prediction effect is achieved by adjusting the model training parameters.

8. The method according to claim 1, characterized in that, The prediction plane is processed and its coordinates are extracted using OpenCV. The prediction image is classified using HSV color segmentation. The geometric characteristics of each category are vectorized and imported into the modeling software for subsequent operations.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-8.

10. A computer-readable storage medium for storing computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the method according to any one of claims 1-8.

Images