Three-dimensional house type diagram generation method

Abstract

The invention relates to the technical field of house type image generation, and discloses a three-dimensional house type image generation method, which comprises the steps of data preprocessing, training and obtaining of a dimension raising deep learning model, obtaining of a real-time three-dimensional model, obtaining of a rendered first three-dimensional house type image and obtaining of an adjusted real-time three-dimensional house type image. According to the method and the device, a data basis is provided for automatically raising the dimension from the two-dimensional image to the three-dimensional structure through data preprocessing, a technical means is provided for automatically identifying and generating the three-dimensional model by setting the space number and the element number, manual design and layout are omitted, the flexibility and the efficiency of application are improved, and the practicability is high. Automatic generation from a two-dimensional image to a three-dimensional model is realized through a dimension raising deep learning model, an accurate model basis is provided for image optimization, and automatic generation and optimization of a three-dimensional house type image based on the three-dimensional model are realized by obtaining a rendering parameter set and an image generation parameter set and combining an image generation system. Therefore, the sense of reality and the intelligent automatic generation capability of the image are improved.

Description

Technical Field

[0001] The present application relates to the technical field of floor plan generation, and specifically to a method for generating a three-dimensional floor plan. Background Art

[0002] Three-dimensional floor plans are a common technical means for modern home improvement companies to show their design effects to customers. They can intuitively and conveniently show customers the personalized home improvement plans that home improvement companies have made for customers. In practical applications, it is convenient for home improvement designers to explain home improvement plans to customers and for customers to understand the corresponding home improvement plans, and to achieve the maximum intuitive display of the actual effects after the home improvement is completed. Therefore, it is necessary to study three-dimensional floor plans and corresponding generation methods. However, although the existing rendering technology can generate high-quality images, it lacks the ability to generate intelligent scenes. This means that it cannot automatically identify and generate scene elements, but requires manual design and layout, which limits the flexibility and efficiency of its application. The existing scene images generated by artificial intelligence, although intelligent and able to automatically generate scene elements, are insufficient in terms of realism and details. The images generated by this technology may not be visually realistic enough, lacking sufficient details and texture, thus affecting the quality and fidelity of the final image.

[0003] Chinese patent number CN202111428640.9 discloses a method, device and system, electronic device and storage medium for determining the house type, the method comprising: obtaining a first image, the first image being a panoramic image and including all structural elements of the scene to be tested, the scene to be tested being the physical space of the house; determining a target acquisition point for each structural element, the target acquisition point being used to draw the scene to be tested; measuring the structural data of the target acquisition point; generating a second image of the scene to be tested according to the target acquisition point and the structural data corresponding to each target acquisition point, the second image being the floor plan of the house, including a three-dimensional floor plan and a plane floor plan; rendering the three-dimensional floor plan in combination with the first image to obtain a third image, the third image being a panoramic three-dimensional floor plan. This method proposes the application of AI technology to three-dimensional floor plans, but does not conduct further research, that is, it does not propose corresponding technical solutions for how to optimize how two-dimensional images are converted into three-dimensional images, how to optimize rendering effects, and how to adjust the final image display solution.

[0004] In summary, the existing technology urgently needs a technical solution for optimizing the generation of three-dimensional floor plans. Summary of the invention

[0005] The purpose of this application is to provide a three-dimensional floor plan generation method to solve the technical problems raised in the above background technology.

[0006] To achieve the above purpose, the present application discloses the following technical solution: a method for generating a three-dimensional floor plan, comprising:

[0007] S1: data preprocessing, the data preprocessing is used to obtain a training data set using existing images; the training data set includes a training two-dimensional floor plan with space numbers and training detail elements with element numbers; the space numbers correspond to the element numbers;

[0008] S2: using the training data set to train and obtain a dimension-upgraded deep learning model; the dimension-upgraded deep learning is used to convert the training two-dimensional house type image into a three-dimensional structure, and fill the three-dimensional structure with the training detail elements to obtain a three-dimensional model;

[0009] S3: Obtain a corresponding real-time data set, run the dimension-upgrading deep learning module, and obtain a real-time three-dimensional model; the real-time data set includes a real-time two-dimensional floor plan and real-time detail elements;

[0010] S4: importing the real-time 3D model into a preset image generation system, obtaining a rendering parameter set, and performing combined rendering to obtain a first 3D floor plan; the image generation system is used to perform the combined rendering, the rendering parameter set is used to provide rendering parameters, the rendering parameters include main color, material texture and light and shadow parameters, and the combined rendering is used to fill the 3D model with color, texture and light and shadow effects;

[0011] S5: Obtain a raw image parameter set, and perform combined adjustment on the first three-dimensional floor plan in the raw image system to obtain a real-time three-dimensional floor plan; the raw image parameter set stores parameters corresponding to the background environment, display color, contrast and light intensity for performing the combined adjustment, and the raw image system is also used to perform the combined adjustment, and the combined adjustment is used to adjust the background environment, display color, contrast and light intensity of the first three-dimensional floor plan.

[0012] Preferably, the existing images include any one or more of apartment photos, high-quality images generated by rendering software, and line models generated by 3D modeling software.

[0013] Preferably, the method of obtaining a training data set using existing images specifically includes:

[0014] Acquire image elements, wherein the image elements include image data, depth information, geometric features, and semantic information;

[0015] The image elements are annotated to obtain a two-dimensional floor plan and detail elements; the two-dimensional floor plan is used to represent the spatial structure of the floor plan and the corresponding points, lines and surfaces and is annotated with the corresponding spatial numbers, and the spatial numbers are used to match the corresponding spatial structure when converted into the three-dimensional structure; the detail elements are used to represent the real detail elements corresponding to the spatial structure and are annotated with the corresponding element numbers; the correspondence between the spatial numbers and the element numbers is used for matching when filling the three-dimensional structure;

[0016] The two-dimensional floor plan and the detail elements are output respectively to obtain the corresponding training two-dimensional floor plan and the training detail elements, and are packaged and defined as the training data set.

[0017] Preferably, the training using the training data set to train and obtain the dimension-raising deep learning model specifically includes training the dimension-raising deep learning model to perform the following steps:

[0018] A1: Acquire and parse the training data set to obtain a training two-dimensional floor plan with the space number and a training detail element with the element number;

[0019] A2: using the space numbers, correspondingly upgrading the points, lines and surfaces in the training two-dimensional floor plan into the three-dimensional structure, and retaining the space numbers;

[0020] A3: Based on the correspondence between the space number and the element number, and using the element number, the three-dimensional structure is filled accordingly, and the corresponding space number and the element number are retained to obtain the three-dimensional model.

[0021] Preferably, the obtaining of the corresponding real-time data set specifically includes:

[0022] Acquire real-time images and real-time image elements, annotate real-time image elements, and obtain corresponding real-time two-dimensional floor plan and real-time detail elements;

[0023] The real-time two-dimensional floor plan and real-time detail elements are output and packaged to be defined as the real-time data set.

[0024] Preferably, the step of running the dimension-upgrading deep learning module to obtain a real-time three-dimensional model specifically includes the following steps:

[0025] B1: Acquire and parse the real-time data set to obtain a real-time two-dimensional floor plan with the space number and a real-time detail element with the element number;

[0026] B2: using the space numbers, correspondingly upgrading the points, lines and surfaces in the real-time two-dimensional floor plan into a real-time three-dimensional structure, and retaining the space numbers;

[0027] B3: Based on the correspondence between the space number and the element number, and using the element number, the real-time three-dimensional structure is filled accordingly, and the corresponding space number and the element number are retained to obtain a real-time three-dimensional model.

[0028] Preferably, the step of importing the real-time three-dimensional model into a preset image generation system further comprises the following steps:

[0029] C1: performing color analysis on the real-time three-dimensional model to obtain a color analysis result; the color analysis is configured to assign colors to the real-time three-dimensional model based on the main color tone input by the designer;

[0030] C2: Performing texture analysis on the real-time three-dimensional model to obtain a texture analysis result; the texture analysis is configured to assign texture to the real-time three-dimensional model based on the material texture input by the designer;

[0031] C3: performing light and shadow analysis on the real-time three-dimensional model to obtain a light and shadow analysis result; the light and shadow analysis is configured to perform light and shadow allocation on the real-time three-dimensional model based on the light and shadow parameters input by the designer.

[0032] Preferably, the combining and rendering to obtain the first three-dimensional floor plan specifically includes:

[0033] Based on the color analysis result, the texture analysis result and the light and shadow analysis result, the real-time three-dimensional model is filled with corresponding colors, textures and light and shadow effects to obtain the first three-dimensional floor plan.

[0034] Preferably, the step of combining and adjusting the first three-dimensional floor plan in the image generation system to obtain a real-time three-dimensional floor plan specifically includes:

[0035] The background environment is acquired, the first three-dimensional floor plan is implanted into the background environment, and the display color, the contrast and the light intensity are adjusted accordingly after implantation.

[0036] Preferably, the step of combining and adjusting the first three-dimensional floor plan in the image generation system to obtain a real-time three-dimensional floor plan further includes:

[0037] Presetting a display color threshold, a contrast threshold and a light intensity threshold, and obtaining a corresponding raw image parameter redundancy set based on the raw image parameter set;

[0038] The first three-dimensional floor plan is combined and adjusted based on the redundant set of raw image parameters to obtain a real-time three-dimensional floor plan redundant atlas.

[0039] Beneficial effects: The three-dimensional floor plan generation method of the present application provides a data basis for automatically upgrading two-dimensional images to three-dimensional structures through data preprocessing, provides a technical means for automatically identifying and generating three-dimensional models by setting space numbers and element numbers, and eliminates manual design and layout, thereby improving the flexibility and efficiency of the application, realizes the automatic generation of two-dimensional images to three-dimensional models through the dimensionality upgrading deep learning model, and provides an accurate model basis for image optimization, and realizes the automatic generation and optimization of three-dimensional floor plans based on three-dimensional models by obtaining rendering parameter sets and image generation parameter sets, and combining with the image generation system, thereby improving the realism of the image and the ability of intelligent automatic generation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0041] Figure 1 A flowchart of a method for generating a three-dimensional floor plan provided in an embodiment of the present application. DETAILED DESCRIPTION

[0042] The technical solutions in the embodiments of the present application are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present application.

[0043] In this article, the term "comprising" is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprising..." do not exclude the presence of other identical elements in the process, method, article or device including the elements.

[0044] This embodiment discloses Figure 1 A method for generating a three-dimensional floor plan is shown, comprising:

[0045] S1: Data preprocessing, data preprocessing is used to obtain a training data set using existing images. The training data set includes a training two-dimensional floor plan with space numbers and training detail elements with element numbers. The space number corresponds to the element number. It should be noted that the data processing method used in this embodiment is any data set processing method in the prior art, and is intended to complete the numbering management of the same type of data.

[0046] S2: Use the training data set to train and obtain a dimensionality-enhanced deep learning model. Dimensionality-enhanced deep learning is used to convert the training two-dimensional house type image into a three-dimensional structure, and fill the three-dimensional structure with training detail elements to obtain a three-dimensional model.

[0047] S3: Get the corresponding real-time data set, run the dimension-upgrading deep learning module, and obtain the real-time 3D model. The real-time data set includes the real-time 2D floor plan and real-time detail elements.

[0048] S4: Import the real-time 3D model into a preset image generation system, obtain a rendering parameter set, and perform combined rendering to obtain a first 3D floor plan. The image generation system is used for combined rendering, and the rendering parameter set is used to provide rendering parameters, and the rendering parameters include main color, material texture, and light and shadow parameters. Combined rendering is used to fill the 3D model with color, texture, and light and shadow effects.

[0049] S5: Obtain a raw image parameter set, and perform combined adjustment on the first three-dimensional floor plan in the raw image system to obtain a real-time three-dimensional floor plan. The raw image parameter set stores parameters corresponding to the background environment, display color, contrast, and light intensity for combined adjustment. The raw image system is also used to perform combined adjustment, and the combined adjustment is used to adjust the background environment, display color, contrast, and light intensity of the first three-dimensional floor plan.

[0050] Through data preprocessing, a data basis is provided for automatically upgrading two-dimensional images to three-dimensional structures. By setting space numbers and element numbers, a technical means is provided for automatically identifying and generating three-dimensional models, eliminating manual design and layout, and improving the flexibility and efficiency of the application. Through the dimensionality upgrading deep learning model, the automatic generation of two-dimensional images to three-dimensional models is realized, and an accurate model basis is provided for image optimization. By obtaining rendering parameter sets and image generation parameter sets and combining with the image generation system, the automatic generation and optimization of three-dimensional floor plans based on three-dimensional models are realized, thereby improving the realism of the image and the ability of intelligent automatic generation.

[0051] Specifically, the existing images include any one or more of apartment photos, high-quality images generated by rendering software, and line models generated by 3D modeling software.

[0052] In this embodiment, the existing images are used to provide the necessary realism and structural data for training the dimensionality-enhanced deep learning model, ensuring that the generated content has a high degree of visual fidelity and detail.

[0053] Specifically, the training data set is obtained using existing images, including:

[0054] Image elements are obtained, where the image elements include image data, depth information, geometric features, and semantic information.

[0055] Label the image elements to obtain a two-dimensional floor plan and detail elements. The two-dimensional floor plan is used to represent the spatial structure of the floor plan and the corresponding points, lines and surfaces and is labeled with the corresponding spatial numbers. The spatial numbers are used to match the corresponding spatial structure when converted into a three-dimensional structure. The detail elements are used to represent the real detail elements corresponding to the spatial structure and are labeled with the corresponding element numbers. The correspondence between the spatial number and the element number is used for matching when filling the three-dimensional structure.

[0056] In a specific application of this embodiment, the image data of this embodiment includes important structural information of walls, doors and windows, etc., and the corresponding depth information includes the length, height and thickness of the wall and the height, width and thickness of the door and window and the corresponding spatial structure. The corresponding geometric features include the shape and material of the wall, door and window, etc., and the image elements are classified by the corresponding space number and element number, and the corresponding filling effect is automatically derived by combining the semantic information of each number. In a specific example, for example, Door 1_1 Indicates the door with space number 1 and element number 1. The semantic information of element number 1 is metal anti-theft door. 1_2 It represents a wall with a space number of 1 and an element number of 2, wherein the semantic information of the element number 2 is a cement wall. Then, according to the same space number, it can be obtained that the wall is the wall constituting the door. Furthermore, when filling, metal elements and cement elements are used respectively.

[0057] The two-dimensional floor plan and detail elements are output respectively to obtain the corresponding training two-dimensional floor plan and training detail elements, and are packaged and defined as a training data set.

[0058] By processing existing images, a standardized data set is provided for training the dimensionality-enhanced deep learning model, which simplifies the model's operations and helps shorten the running time, and provides data reference for real-time image processing.

[0059] Specifically, the training using the training data set to train and obtain the dimension-increased deep learning model specifically includes training the dimension-increased deep learning model to perform the following steps:

[0060] A1: Acquire and parse the training data set to obtain a training two-dimensional floor plan with space numbers and a training detail element with element numbers.

[0061] A2: Use space numbers to upgrade the points, lines, and surfaces in the training two-dimensional floor plan into a three-dimensional structure, and retain the space numbers.

[0062] A3: Based on the correspondence between the space number and the element number, and using the element number, the three-dimensional structure is filled accordingly, and the corresponding space number and element number are retained to obtain the three-dimensional model.

[0063] It should be noted that the dimensionality-enhanced deep learning model trained in this embodiment can be combined with any existing three-dimensional model generation software to achieve the technical effect of dimensionality enhancement. The dimensionality-enhanced deep learning model of this embodiment aims to improve the accuracy and intelligence during dimensionality enhancement, and further enhance the realism of the generated three-dimensional model.

[0064] Using the training data set to train the dimensionality-increasing deep learning model enables the model to automatically generate three-dimensional structures from two-dimensional images and fill in the necessary realism and details, so that it can automatically identify and generate scene elements, thereby improving the system's intelligent generation capabilities.

[0065] Specifically, obtain the corresponding real-time data set, including:

[0066] Acquire the real-time image and the real-time image elements, and annotate the real-time image elements. The annotating of the real-time image elements here adopts the same technical idea as the aforementioned annotating of the existing image, and obtain the corresponding real-time two-dimensional floor plan and real-time detail elements.

[0067] Output and package real-time 2D floor plans and real-time detail elements into real-time datasets.

[0068] Specifically, running the dimension-upgrading deep learning module to obtain a real-time three-dimensional model includes the following steps:

[0069] B1: Acquire and parse the real-time data set to obtain a real-time two-dimensional floor plan with space numbers and real-time detail elements with element numbers.

[0070] B2: Use space numbers to upgrade the points, lines and surfaces in the real-time two-dimensional floor plan to a real-time three-dimensional structure, and retain the space numbers.

[0071] B3: Based on the correspondence between the space number and the element number, and using the element number, the real-time three-dimensional structure is filled in accordingly, and the corresponding space number and element number are retained to obtain the real-time three-dimensional model.

[0072] By analyzing real-time data sets, it can automatically identify and generate three-dimensional space layouts, solving the shortcomings of traditional 3D rendering technology in intelligent recognition.

[0073] Specifically, the real-time 3D model is imported into a preset image generation system, and the following steps are also included:

[0074] C1: Perform color analysis on the real-time 3D model to obtain a color analysis result. The color analysis is configured to assign colors to the real-time 3D model based on the main color tone input by the designer.

[0075] It should be noted that the strategy adopted for color allocation in this embodiment is specifically:

[0076] Based on the color value of the main color, similarity calculation is performed to obtain colors close to the color value. Corresponding color allocation is performed based on the space number and element number, and real-time color allocation is modified based on the collected designer's adjustments.

[0077] C2: Perform texture analysis on the real-time 3D model to obtain texture analysis results. The texture analysis configuration is to assign texture to the real-time 3D model based on the material texture input by the designer.

[0078] It should be noted that the strategy adopted for texture allocation in this embodiment is specifically:

[0079] Based on the material textures exported by the designer, real-time texture assignment is performed based on the element number.

[0080] C3: Perform light and shadow analysis on the real-time 3D model to obtain light and shadow analysis results. The light and shadow analysis configuration is to distribute light and shadow to the real-time 3D model based on the light and shadow parameters input by the designer.

[0081] It should be noted that the strategy adopted for light and shadow allocation in this embodiment is specifically:

[0082] The designer inputs actual light and shadow parameters based on the actual lighting effect of the apartment, and distributes light and shadow based on the depth information and geometric features mentioned above in this embodiment, aiming to restore the actual lighting effect of the apartment.

[0083] Specifically, the first three-dimensional floor plan is obtained by combining rendering, including:

[0084] Based on the color analysis results, texture analysis results and light and shadow analysis results, the real-time three-dimensional model is filled with corresponding colors, textures and light and shadow effects to obtain a first three-dimensional floor plan.

[0085] It should be noted that the combined rendering process of this embodiment not only increases details and realism, but also allows customized design elements such as furniture and decorations to ensure consistency with the style and layout of the overall space.

[0086] Specifically, the first three-dimensional floor plan is combined and adjusted in the image generation system to obtain a real-time three-dimensional floor plan, which specifically includes:

[0087] A background environment is obtained, the first three-dimensional floor plan is implanted into the background environment, and corresponding display color, contrast and light intensity are adjusted after implantation.

[0088] By combining adjustments, designers and users can optimize renderings according to specific needs, further improving the accuracy and satisfaction of the design.

[0089] It should be noted that this embodiment allows designers to adjust the realism and intelligence of images as needed by obtaining rendering parameter sets and raw image parameter sets, providing highly flexible and personalized options. This means that designers can adjust design solutions based on real-time feedback from users to ensure that the final product meets the specific needs of customers.

[0090] Furthermore, the automatic combined rendering and combined adjustment of the 3D rendering system reduces the dependence on senior designers and 3D rendering experts, reducing labor costs. On average, each project can achieve effective total cost savings by reducing the number of manual modifications and iterations. In addition, the digitalization and automation of the design process reduces the use and waste of physical materials, which is in line with the goal of sustainable development. By reducing the need for printing and physical models, material waste can be further avoided every year.

[0091] Specifically, combining and adjusting the first three-dimensional floor plan in the image generation system to obtain a real-time three-dimensional floor plan also includes:

[0092] The display color threshold, contrast threshold and light intensity threshold are preset, and based on the raw image parameter set, a corresponding raw image parameter redundant set is obtained.

[0093] The first three-dimensional floor plan is combined and adjusted based on the redundant set of raw image parameters to obtain a real-time three-dimensional floor plan redundant atlas.

[0094] It should be noted that this embodiment obtains and outputs a series of atlases associated with the real-time three-dimensional floor plan by obtaining the real-time three-dimensional floor plan redundant atlas, which provides a data basis for designers to quickly switch when presenting to users, further enhancing the user experience of this embodiment.

[0095] In summary, the three-dimensional floor plan generation method of the present embodiment provides a data basis for automatically upgrading two-dimensional images to three-dimensional structures through data preprocessing, provides a technical means for automatically identifying and generating three-dimensional models by setting space numbers and element numbers, eliminates manual design and layout, and improves the flexibility and efficiency of the application. It realizes the automatic generation of two-dimensional images to three-dimensional models through the dimensionality upgrading deep learning model, and provides an accurate model basis for image optimization. It realizes the automatic generation and optimization of three-dimensional floor plans based on three-dimensional models by acquiring rendering parameter sets and image generation parameter sets, and combining with the image generation system, thereby improving the realism of the image and the ability of intelligent automatic generation.

[0096] In the embodiments provided in the present application, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: application specific integrated circuit (ASIC), digital signal processor (DSP), digital signal processing device (DSPD), programmable logic device (PLD), field programmable gate array (FPGA), processor, controller, microcontroller, microprocessor, other electronic units designed to implement the functions described herein or their combination. For software implementation, part or all of the flow of the embodiment can be completed by instructing the relevant hardware through a computer program. When implemented, the above program can be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein the communication medium includes any medium that is convenient for transmitting a computer program from one place to another. The storage medium can be any available medium that a computer can access. The computer-readable storage medium can include but is not limited to RAM, ROM, EEPROM, CD-ROM or other optical disk storage, disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer.

[0097] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for generating a three-dimensional floor plan. It is characterized in that include: S1: data preprocessing, the data preprocessing is used to obtain a training data set using existing images; the training data set includes a training two-dimensional floor plan with space numbers and training detail elements with element numbers; the space numbers correspond to the element numbers; S2: using the training data set to train and obtain a dimension-upgraded deep learning model; the dimension-upgraded deep learning is used to convert the training two-dimensional house type image into a three-dimensional structure, and fill the three-dimensional structure with the training detail elements to obtain a three-dimensional model; S3: Obtain a corresponding real-time data set, run the dimension-upgrading deep learning module, and obtain a real-time three-dimensional model; the real-time data set includes a real-time two-dimensional floor plan and real-time detail elements; S4: importing the real-time 3D model into a preset image generation system, obtaining a rendering parameter set, and performing combined rendering to obtain a first 3D floor plan; the image generation system is used to perform the combined rendering, the rendering parameter set is used to provide rendering parameters, the rendering parameters include main color, material texture and light and shadow parameters, and the combined rendering is used to fill the 3D model with color, texture and light and shadow effects; S5: Obtain a raw image parameter set, and perform combined adjustment on the first three-dimensional floor plan in the raw image system to obtain a real-time three-dimensional floor plan; the raw image parameter set stores parameters corresponding to the background environment, display color, contrast and light intensity for performing the combined adjustment, and the raw image system is also used to perform the combined adjustment, and the combined adjustment is used to adjust the background environment, display color, contrast and light intensity of the first three-dimensional floor plan.

2. The method for generating a three-dimensional floor plan according to claim 1, It is characterized in that The existing images include any one or more of apartment photos, high-quality images generated by rendering software, and line models generated by 3D modeling software.

3. The method for generating a three-dimensional floor plan according to claim 2, It is characterized in that The method of obtaining a training data set using existing images specifically includes: Acquire image elements, wherein the image elements include image data, depth information, geometric features, and semantic information; The image elements are annotated to obtain a two-dimensional floor plan and detail elements; the two-dimensional floor plan is used to represent the spatial structure of the floor plan and the corresponding points, lines and surfaces and is annotated with the corresponding spatial numbers, and the spatial numbers are used to match the corresponding spatial structure when converted into the three-dimensional structure; the detail elements are used to represent the real detail elements corresponding to the spatial structure and are annotated with the corresponding element numbers; the correspondence between the spatial numbers and the element numbers is used for matching when filling the three-dimensional structure; The two-dimensional floor plan and the detail elements are output respectively to obtain the corresponding training two-dimensional floor plan and the training detail elements, and are packaged and defined as the training data set.

4. The method for generating a three-dimensional floor plan according to claim 1, It is characterized in that The training of using the training data set to train and obtain the dimension-increased deep learning model specifically includes training the dimension-increased deep learning model to perform the following steps: A1: Acquire and parse the training data set to obtain a training two-dimensional floor plan with the space number and a training detail element with the element number; A2: using the space numbers, correspondingly upgrading the points, lines and surfaces in the training two-dimensional floor plan into the three-dimensional structure, and retaining the space numbers; A3: Based on the correspondence between the space number and the element number, and using the element number, the three-dimensional structure is filled accordingly, and the corresponding space number and the element number are retained to obtain the three-dimensional model.

5. The method for generating a three-dimensional floor plan according to claim 3, It is characterized in that The obtaining of the corresponding real-time data set specifically includes: Acquire real-time images and real-time image elements, annotate real-time image elements, and obtain corresponding real-time two-dimensional floor plan and real-time detail elements; The real-time two-dimensional floor plan and real-time detail elements are output and packaged to be defined as the real-time data set.

6. The method for generating a three-dimensional floor plan according to claim 4, It is characterized in that The step of running the dimension-upgrading deep learning module to obtain a real-time three-dimensional model specifically includes the following steps: B1: Acquire and parse the real-time data set to obtain a real-time two-dimensional floor plan with the space number and a real-time detail element with the element number; B2: using the space numbers, correspondingly upgrading the points, lines and surfaces in the real-time two-dimensional floor plan into a real-time three-dimensional structure, and retaining the space numbers; B3: Based on the correspondence between the space number and the element number, and using the element number, the real-time three-dimensional structure is filled accordingly, and the corresponding space number and the element number are retained to obtain a real-time three-dimensional model.

7. The method for generating a three-dimensional floor plan according to claim 1, It is characterized in that The method of importing the real-time three-dimensional model into a preset image generation system further includes the following steps: C1: performing color analysis on the real-time three-dimensional model to obtain a color analysis result; the color analysis is configured to assign colors to the real-time three-dimensional model based on the main color tone input by the designer; C2: Performing texture analysis on the real-time three-dimensional model to obtain a texture analysis result; the texture analysis is configured to assign texture to the real-time three-dimensional model based on the material texture input by the designer; C3: performing light and shadow analysis on the real-time three-dimensional model to obtain a light and shadow analysis result; the light and shadow analysis is configured to perform light and shadow allocation on the real-time three-dimensional model based on the light and shadow parameters input by the designer.

8. The method for generating a three-dimensional floor plan according to claim 7, It is characterized in that The above-mentioned combined rendering is performed to obtain a first three-dimensional floor plan, which specifically includes: Based on the color analysis result, the texture analysis result and the light and shadow analysis result, the real-time three-dimensional model is filled with corresponding colors, textures and light and shadow effects to obtain the first three-dimensional floor plan.

9. The method for generating a three-dimensional floor plan according to claim 1, It is characterized in that The step of combining and adjusting the first three-dimensional floor plan in the image generation system to obtain a real-time three-dimensional floor plan specifically includes: The background environment is acquired, the first three-dimensional floor plan is implanted into the background environment, and the display color, the contrast and the light intensity are adjusted accordingly after implantation.

10. The method for generating a three-dimensional floor plan according to claim 9, It is characterized in that The step of combining and adjusting the first three-dimensional floor plan in the image generation system to obtain a real-time three-dimensional floor plan further includes: Presetting a display color threshold, a contrast threshold and a light intensity threshold, and obtaining a corresponding raw image parameter redundancy set based on the raw image parameter set; The first three-dimensional floor plan is combined and adjusted based on the redundant set of raw image parameters to obtain a real-time three-dimensional floor plan redundant atlas.

Images