House plan reconstruction method, device and equipment

By extracting feature parameters from two-dimensional floor wireframe diagrams and combining the floor plan generation model, the problem of lack of realism and details in the existing technology is solved, and high-quality three-dimensional floor plan reconstruction is achieved, improving user experience and design efficiency .

CN119091045APending Publication Date: 2024-12-06KUATULI (GUANGZHOU) TECHNOLOGY CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411181398.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-27
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing floor plan display form cannot provide realism and details, and cannot meet the interactive needs of users.

Method used

By obtaining the two-dimensional floor wireframe diagram, extracting feature parameters, building a three-dimensional floor plan, and reconstructing the user-type diagram generation model, adding color, texture and light and shadow effects to generate high-quality three-dimensional color stereoscopic diagrams.

Benefits of technology

It improves the realism and interactivity of floor plans, and can quickly generate a variety of residential floor plans that meet user needs, providing a more intuitive and rich spatial display.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure BDA0005012270230000041
    Figure BDA0005012270230000041
  • Figure BDA0005012270230000042
    Figure BDA0005012270230000042
  • Figure BDA0005012270230000051
    Figure BDA0005012270230000051
Patent Text Reader

Abstract

The invention provides a house type diagram reconstruction method, device and equipment, and relates to the technical field of computer information processing, and the method comprises the steps: obtaining a two-dimensional house type wireframe diagram; performing feature parameter extraction on the two-dimensional house type wireframe diagram to obtain a feature quantity; obtaining a three-dimensional house type plane graph according to the characteristic quantity; inputting the three-dimensional house type planar graph into a house type graph generation model for house type graph reconstruction to obtain a house type graph; wherein the house type image generation model is obtained by training a preset model according to a historical house type image training set. According to the scheme of the invention, various residential house type images meeting user requirements can be quickly generated, and the sense of reality of the house type images is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of computer information processing, and in particular to a method, device and equipment for reconstructing a floor plan. Background Art

[0002] In recent years, with the hot real estate market, buyers' demands for house types have gradually diversified. However, the house types on the market are mostly displayed in the form of floor plans or bird's-eye models. They cannot experience the architectural space from a human perspective, lack realism and details, and do not have real-time interactivity, which cannot meet the needs of users. Summary of the invention

[0003] The present invention provides a method, device and equipment for reconstructing a floor plan, so as to solve the problem that the existing floor plans are insufficient in realism and details and cannot meet the needs of users.

[0004] In order to solve the above technical problems, the technical solution of the present invention is as follows:

[0005] A method for reconstructing a floor plan, comprising:

[0006] Get a 2D wireframe of the house type;

[0007] Extracting feature parameters from the two-dimensional house type wireframe diagram to obtain feature quantities;

[0008] According to the characteristic quantity, a three-dimensional floor plan is obtained;

[0009] The three-dimensional floor plan is input into a floor plan generation model to reconstruct the floor plan to obtain a floor plan; wherein the floor plan generation model is obtained by training a preset model based on a historical floor plan training set.

[0010] Optionally, feature parameters are extracted from the two-dimensional house type wireframe diagram to obtain feature quantities, including:

[0011] Preprocessing the two-dimensional house type wireframe diagram;

[0012] Identify the building structure in the preprocessed two-dimensional house type wireframe diagram to obtain building structure feature quantities;

[0013] The scale of the preprocessed two-dimensional house type wireframe diagram is identified to obtain a spatial feature quantity.

[0014] Optionally, obtaining a three-dimensional apartment floor plan according to the characteristic quantity includes:

[0015] Construct a three-dimensional space framework;

[0016] The building structure feature quantity is mapped to the three-dimensional space framework, and the building structure feature quantity is adjusted according to the space feature quantity to obtain a three-dimensional apartment plan.

[0017] Optionally, the three-dimensional floor plan is input into a floor plan generation model to reconstruct the floor plan to obtain the floor plan, including:

[0018] The floor plan generation model converts the three-dimensional floor plan into a three-dimensional color floor plan;

[0019] Filling is performed on the three-dimensional color plan to obtain a floor plan.

[0020] Optionally, the floor plan generation model converts the three-dimensional floor plan into a three-dimensional color floor plan, including:

[0021] The floor plan generation model renders colors in the three-dimensional floor plan according to the building structure feature.

[0022] Optionally, filling is performed on the three-dimensional color plan to obtain a floor plan, including:

[0023] Filling texture on the three-dimensional color plane image to generate a three-dimensional color stereogram;

[0024] Light and shadow effects are added to the three-dimensional color stereogram to obtain a floor plan.

[0025] Optionally, the floor plan generation model training process includes:

[0026] Obtaining a historical floor plan training set, wherein the historical floor plan training set includes real images and text descriptions;

[0027] Adding noise to the historical floor plan training set to obtain noisy data and inputting it into a preset model;

[0028] The preset loss function of the preset model is iterated through the noise-added data to obtain a floor plan generation model.

[0029] The present invention also provides a floor plan reconstruction device, comprising:

[0030] An acquisition module is used to obtain a two-dimensional house type wireframe diagram;

[0031] A processing module is used to extract feature parameters from the two-dimensional house type wireframe diagram to obtain feature quantities; obtain a three-dimensional house type floor plan based on the feature quantities; input the three-dimensional house type floor plan into a house type generation model to reconstruct the house type to obtain a house type; wherein the house type generation model is obtained by training a preset model based on a historical house type training set.

[0032] The present invention also provides a computing device, comprising: a processor and a memory storing a computer program, wherein when the computer program is executed by the processor, the method described above is executed.

[0033] The present invention also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enable the computer to execute the method described above.

[0034] The above solution of the present invention includes at least the following beneficial effects:

[0035] The above scheme of the present invention includes: obtaining a two-dimensional house type wireframe diagram; extracting feature parameters of the two-dimensional house type wireframe diagram to obtain feature quantities; obtaining a three-dimensional house type floor plan according to the feature quantities; inputting the three-dimensional house type floor plan into a house type generation model to reconstruct the house type to obtain a house type diagram; wherein the house type generation model is obtained by training a preset model according to a historical house type training set. The scheme of the present invention can quickly generate a variety of residential house types that meet user needs and improve the realism of the house type diagram. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 is a flow chart of a floor plan reconstruction method provided by an embodiment of the present invention;

[0037] Figure 2 It is a module schematic diagram of a floor plan reconstruction device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The exemplary embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present invention and to enable the scope of the present invention to be fully communicated to those skilled in the art.

[0039] like Figure 1 As shown, an embodiment of the present invention provides a method for reconstructing a floor plan, comprising:

[0040] Step 11, obtaining a two-dimensional house type wireframe diagram;

[0041] Step 12, extracting feature parameters from the two-dimensional house type wireframe diagram to obtain feature quantities;

[0042] Step 13, obtaining a three-dimensional apartment floor plan according to the feature quantity;

[0043] Step 14, input the three-dimensional floor plan into a floor plan generation model to reconstruct the floor plan to obtain a floor plan; wherein the floor plan generation model is obtained by training a preset model based on a historical floor plan training set.

[0044] In this embodiment, a corresponding three-dimensional interior plan design drawing is generated based on the two-dimensional floor plan drawn by the user. The three-dimensional interior plan design drawing is used as a wireframe and input into a floor plan generation model. Based on deep learning and graphics technology, a perfect fusion of realistic rendering and intelligent generation is achieved to generate high-quality, realistic image floor plans.

[0045] A 2D floor plan wireframe is usually a top-down view showing the relative positions and sizes of rooms, walls, doors, windows, and other fixtures and furniture. Converting a 2D floor plan wireframe into a 3D floor plan provides a richer and more intuitive spatial display, including third-dimensional information such as height, depth, and width. Then converting the 3D floor plan into a 3D floor plan provides a richer and more intuitive spatial display, allowing observers to understand the space from different angles and dimensions, achieving a highly realistic and detailed 3D spatial display, and providing users with a more intuitive and interactive experience.

[0046] In an optional embodiment of the present invention, step 12 may include:

[0047] Step 121, preprocessing the two-dimensional house type wireframe diagram;

[0048] Step 122, identifying the building structure in the pre-processed two-dimensional house type wireframe diagram to obtain building structure feature quantities;

[0049] Step 123, identifying the scale of the pre-processed two-dimensional house type wireframe diagram to obtain a spatial feature value.

[0050] In this embodiment, the two-dimensional house type wireframe image is firstly smoothed and contrast enhanced to reduce image noise, prevent noise points from being erroneously identified, and improve the accuracy of wireframe recognition.

[0051] Furthermore, the points with obvious brightness changes in the two-dimensional house wireframe are identified, which are the turning points of the corresponding object outline or shape. The center line of the wall is extracted using an edge detection algorithm, and the center line is optimized, such as merging center lines, deleting overlapping center lines, delaying center lines, and closing center lines, so as to obtain accurate wall recognition results.

[0052] Specifically, based on the gradient magnitude and direction of the image brightness, the image is convolved to calculate the approximate gradients in the horizontal and vertical directions:

[0053]

[0054] Among them, G x is the horizontal gradient, G y is the vertical gradient.

[0055] G x and G y Applied to each pixel of the image separately, calculate the gradient response of the pixel in the horizontal and vertical directions:

[0056] Calculate the gradient magnitude for each pixel:

[0057]

[0058] Calculate the gradient direction:

[0059]

[0060] I is the pixel value of the image. In the gradient direction, only the pixels with the local maximum value are retained to form a finer edge.

[0061] According to the edge strength required by different buildings, two high and low thresholds are set to determine strong edges and weak edges. The detected edge points are fitted into straight lines to represent the positions of the walls. The candidate area edges of doors and windows are extracted, and these edges are connected to form the final building structure feature quantity.

[0062] Furthermore, the scale information on the two-dimensional house type wireframe is identified and converted into digital form for subsequent size adjustment of the three-dimensional model.

[0063] Identify the types (such as doors, windows, sliding doors, etc.), sizes and locations of doors and windows on the two-dimensional house wireframe, and identify the room names or functional descriptions on the two-dimensional house wireframe.

[0064] Specifically, the text area in the two-dimensional house wireframe diagram is identified, which may include the positioning of text lines and text blocks, and the detected text area is further segmented into individual characters or words, and features (including shape descriptors, texture features, geometric features, etc.) are extracted from the segmented characters, and matched in a preset known character library to identify the characters; the recognition results are corrected, such as correcting spelling errors, formatting output, etc., to obtain spatial feature quantities, which are used to assign correct spatial names to each area.

[0065] Finally, the building structure characteristics and space characteristics are parsed into XML files.

[0066] In an optional embodiment of the present invention, step 13 may include:

[0067] Step 131, constructing a three-dimensional space framework;

[0068] Step 132, mapping the building structure feature quantity to the three-dimensional space framework, and adjusting the building structure feature quantity by the space feature quantity to obtain a three-dimensional apartment plan.

[0069] In this embodiment, a three-dimensional space framework is automatically constructed based on information such as walls, doors and windows, and scale information of the two-dimensional floor plan.

[0070] Specifically, according to the wall size and position in the 2D floor plan in the XML file, draw the wall in 3D space, set the appropriate height for the wall according to the actual required building space height; add door and window openings on the wall according to the position in the 2D floor plan wireframe, and adjust the shape and size of the openings according to the type of door and window. If the floor plan includes multiple floors, you also need to create floors for each floor and place walls at the corresponding height. Then add internal partition walls in 3D space to distinguish different rooms and areas.

[0071] Further, adjust the parameters in step 12 to ensure that the size, opening direction and position of the doors and windows are consistent with the 2D house line frame. Figure 1 The final result is a black and white 3D frame diagram, which clearly depicts the 3D structure and spatial layout of the room, but does not contain any color, texture or decorative details for the time being, which is concise and intuitive.

[0072] The user can select an optimal viewing angle to generate a black and white wireframe of the black and white 3D framework. This wireframe retains all the details of the room's 3D structure, but still maintains visual simplicity, making it easy to understand and modify. The resulting black and white wireframe can serve as the basis for design discussions or as input for subsequent rendering processes.

[0073] In an optional embodiment of the present invention, step 14 may include:

[0074] Step 141, the floor plan generation model converts the three-dimensional floor plan into a three-dimensional color floor plan;

[0075] Step 142, filling in the three-dimensional color plan to obtain a floor plan.

[0076] In this embodiment, based on the three-dimensional floor plan obtained in step 13, the input wireframe diagram is converted into a final rendered floor plan with vivid colors, textures, and light and shadow effects by combining the floor plan to generate a model.

[0077] Specifically, it automatically analyzes each part of the wireframe, such as walls, floors, ceilings, etc., and then fills these parts with corresponding colors and materials based on the preset design style or the user's specific requirements.

[0078] At the same time, the light and shadow effects of natural and artificial light sources in the space are simulated, making the rendered effect both realistic and artistic. In addition, the stable diffusion system can also handle more complex design elements, such as furniture, decorations and other interior design details. Finally, the 3D floor plan is converted into a 3D floor plan.

[0079] Automatically fill in colors, textures, and light and shadow effects based on the wireframe to generate the final rendered room image. This rendering not only truly restores the design intent, but is also colorful and lifelike, and can intuitively show the actual effect of the future space. The entire process, from 2D sketches to detailed 3D renderings, achieves fast and automated design visualization, greatly improves design efficiency and the quality of effect display, helps users and designers reach a design consensus in the shortest time, and accelerates the progress of design projects.

[0080] In an optional embodiment of the present invention, step 141 includes:

[0081] The floor plan generation model renders colors in the three-dimensional floor plan according to the building structure feature.

[0082] In this embodiment, the building structure characteristics can be used to obtain basic information such as room layout, door and window locations, wall thickness, and floor height, and the location and size of structures such as load-bearing walls, non-load-bearing walls, columns, and beams can be determined.

[0083] The floor plan generation model combines the building structure characteristics with the three-dimensional floor plan to generate an independent model for each building structure. Corresponding materials are set for different structures (such as walls, floors, ceilings, doors and windows, etc.) in the model. The materials can be automatically generated in the model according to the preset design style, or they can be generated according to the specific requirements of the text description provided by the user.

[0084] Next, select appropriate colors for different structures in the model. This includes the color of the walls, the color of the floor, the color of the doors and windows, etc. The colors can be automatically generated in the model according to the preset design style, or they can be rendered according to the specific requirements of the text description provided by the user.

[0085] In an optional embodiment of the present invention, step 142 may include:

[0086] Step 1421, filling texture on the three-dimensional color plane image to generate a three-dimensional color stereogram;

[0087] Step 1422, adding light and shadow effects to the three-dimensional color stereogram to obtain a floor plan.

[0088] In this embodiment, the system automatically identifies the material of the building structure, such as wood, metal, tile, etc., selects a suitable texture from a preset texture library, and applies the selected texture to the corresponding part of the building structure. Then the system automatically evaluates the effect after mapping, checks whether there are obvious texture duplications, unnatural seams, etc., and makes necessary optimization adjustments based on the evaluation results, such as adjusting the texture size, position, brightness, etc., to obtain better visual effects.

[0089] The texture library classifies and labels the textures corresponding to each material so that the matching texture can be quickly found based on the material recognition results.

[0090] The system uses a lighting model to simulate the effect of natural or artificial light on a room. This includes calculating parameters such as the position, intensity, and color of the light source, as well as the interaction between light and the surface of an object (such as reflection, refraction, and shadows). By adjusting lighting parameters and rendering settings, the system can generate realistic light and shadow effects, making the rendering more vivid and lifelike.

[0091] Specifically, first define the basic parameters of the light source, including:

[0092] Position: The coordinates of the light source in three-dimensional space.

[0093] Intensity: The brightness of a light source, usually expressed as RGB values ​​(red, green, blue) or in more complex color spaces such as HSV.

[0094] Color: The color of a light source, along with intensity, determines the overall appearance of the light source.

[0095] Type: The light source can be a point light source (such as a light bulb), a parallel light (such as sunlight), or a spotlight, etc.

[0096] Attenuation: How much a light source weakens with distance.

[0097] The material properties of the object surface are then determined based on the material of the building structure. These properties affect the way light interacts with the object:

[0098] Reflectivity: The proportion of light reflected from an object's surface.

[0099] Refractive index (for transparent or translucent objects): the degree to which light changes direction when passing through an object.

[0100] Diffuse reflection and specular reflection: Distinguish whether light is scattered evenly on the surface of an object (diffuse reflection) or reflected at a specific angle (specular reflection).

[0101] Roughness: Affects the blurriness of specular reflections.

[0102] Color: The color of an object's surface affects the color of reflected light.

[0103] Based on the lighting principle of physics, the lighting effect is decomposed into several components, including ambient light, diffuse light, and specular light. These components together determine the final color of the object surface. The lighting equation is used to calculate the interaction between light and the object surface:

[0104] Lighting = ambient light + diffuse light + specular reflection;

[0105] Among them, ambient light is a constant;

[0106] Diffuse reflected light = C light ×M diffuse ×max(0,n·l), where C light is the light source color, M diffuse is the diffuse color of the object, n is the normal direction of the object surface, l is the direction of the light source, and the dot product n·l represents the cosine value of the angle between the light source direction and the normal direction;

[0107] Specular Reflection = C light ×M specular ×(max(0,r·v) s ), where C light is the light source color, M specular is the highlight color of the object, r is the direction of the reflected light (which can be calculated by r = 2(n·l)nl), v is the viewing direction, and s is the highlight index (which controls the sharpness of the highlight);

[0108] Furthermore, shadow effects are added. Shadow Mapping is used to pre-calculate the depth map of the scene from the perspective of the light source, and then the depth value of the line of sight from the surface point of the object to the light source is compared during rendering, which increases the depth and realism of the scene.

[0109] It combines diffuse reflection, specular reflection, ambient light (simulating indirect light) and shadow, taking into account the lighting and shadow effects produced by the light source directly shining on the surface of the object. It can simulate the glossiness of the object's surface and generate realistic floor plans at a higher frame rate.

[0110] Finally, the lighting model is integrated with other parts of the scene (such as geometry, textures, animations, etc.), and debugged to ensure that the lighting effect is as expected. The generated images are adjusted and optimized, including adjusting the details, background, environment, etc. of the wireframe images, and adjusting the color, contrast, lighting, etc. of the realistic rendered images to achieve the best visual effect.

[0111] In an optional embodiment of the present invention, the floor plan generation model training process includes:

[0112] Step 21, obtaining a historical floor plan training set, wherein the historical floor plan training set includes real images and text descriptions;

[0113] Step 22, adding noise to the historical floor plan training set to obtain noise-added data and inputting it into a preset model;

[0114] Step 23, iterating the preset loss function of the preset model through the noise-added data to obtain a floor plan generation model.

[0115] In this embodiment, during the training of the floor plan generation model, a data set for training needs to be prepared first. The training data set consists of real images and corresponding text descriptions. The real images are real photos or high-quality 3D floor plans generated by professional rendering software. The text descriptions are text information provided by the user, such as:

[0116] Room layout: Describe the functional use and layout of the room, such as the location of the living room, number of bedrooms and kitchen facilities.

[0117] Home decor: Mention furniture placement, color style, and decorative elements, such as the sofa in the center of the living room, artwork on the wall, etc.

[0118] Structural details: These involve the structural details of the building, such as window shapes, staircase locations, and ceiling designs.

[0119] Materials and colors: Describe the types of building materials used, color combinations, floor materials, etc.

[0120] Example:

[0121] The living room is located in the center of the house, near the entrance. A large sofa is placed along the wall, echoing the wooden coffee table. The room adopts a modern style, with dark blue walls and light wooden floors. The living room faces a large French window, which can see the garden view through.

[0122] Furthermore, the preset model is trained using the above training data set, and the image generation system can be selected as the preset model for training. The image generation system can automatically fill in the color, texture and light and shadow effects according to the wireframe to generate the final rendered room image, which is an important application in current graphics processing and rendering technology. The system usually combines technologies from multiple fields such as computer vision, image processing, deep learning, and graphics rendering, and can efficiently and accurately complete the conversion from wireframe to final rendering.

[0123] Noise is gradually added to the training set in continuous time t, where t ranges from 0 to T (total number of time steps), converting the data into noisy data. The data is converted using the following formula:

[0124]

[0125] Among them, x t is the converted noise-added data, a t The variance schedule is set in advance to control the amount of noise increase at each step, where ∈ is the noise drawn from the standard normal distribution N(0,I).

[0126] This process converts the data into a Gaussian distribution, and the noise is gradually increasing, which means that as time tt increases, the data x t It will gradually lose its original structure and eventually turn into almost complete noise at t=T.

[0127] Furthermore, through the loss function: Iterate to denoise the data and restore the original data from the noisy data.

[0128] Here, ∈ is a noise vector randomly drawn from the standard normal distribution N(0,I), and x t It is the noisy data obtained from the original data x0 through some diffusion process (such as additive noise process), with a time step of t. Function ∈θ(x t ,t) is a parametric model whose goal is to predict given x t and t, used to generate x t The original noise ∈.

[0129] In the iterative process, the gradient descent method is used to calculate the loss function L simple Iterate and find the model parameter θ so that the loss function L simple At the same time, the difference between the production image and the real image is compared. When the difference is less than 1%, the training is completed and the floor plan generation model is obtained.

[0130] The floor plan reconstruction method described in the above embodiment of the present invention can quickly generate a variety of residential floor plans that meet user needs. Through training with a large number of realistic images and corresponding wireframe images, the model learns the mapping relationship between realistic images and wireframe images. It combines the intelligent generation of deep learning and the realistic rendering technology of computer graphics, and through algorithm training and optimization, it achieves a perfect combination of intelligent generation and realistic rendering, bringing new possibilities to the fields of design and visual presentation.

[0131] like Figure 2 As shown, in an embodiment of the present invention, a floor plan reconstruction device 20 is also provided, comprising:

[0132] An acquisition module 21 is used to acquire a two-dimensional house type wireframe diagram;

[0133] The processing module 22 is used to extract feature parameters from the two-dimensional house type wireframe diagram to obtain feature quantities; obtain a three-dimensional house type floor plan based on the feature quantities; input the three-dimensional house type floor plan into a house type generation model to reconstruct the house type to obtain a house type; wherein the house type generation model is obtained by training a preset model based on a historical house type training set.

[0134] Optionally, feature parameters are extracted from the two-dimensional house type wireframe diagram to obtain feature quantities, including:

[0135] Preprocessing the two-dimensional house type wireframe diagram;

[0136] Identify the building structure in the preprocessed two-dimensional house type wireframe diagram to obtain building structure feature quantities;

[0137] The scale of the preprocessed two-dimensional house type wireframe diagram is identified to obtain a spatial feature quantity.

[0138] Optionally, obtaining a three-dimensional apartment floor plan according to the characteristic quantity includes:

[0139] Construct a three-dimensional space framework;

[0140] The building structure feature quantity is mapped to the three-dimensional space framework, and the building structure feature quantity is adjusted according to the space feature quantity to obtain a three-dimensional apartment plan.

[0141] Optionally, the three-dimensional floor plan is input into a floor plan generation model to reconstruct the floor plan to obtain the floor plan, including:

[0142] The floor plan generation model converts the three-dimensional floor plan into a three-dimensional color floor plan;

[0143] Filling is performed on the three-dimensional color plan to obtain a floor plan.

[0144] Optionally, the floor plan generation model converts the three-dimensional floor plan into a three-dimensional color floor plan, including:

[0145] The floor plan generation model renders colors in the three-dimensional floor plan according to the building structure feature.

[0146] Optionally, filling is performed on the three-dimensional color plan to obtain a floor plan, including:

[0147] Filling texture on the three-dimensional color plane image to generate a three-dimensional color stereogram;

[0148] Light and shadow effects are added to the three-dimensional color stereogram to obtain a floor plan.

[0149] Optionally, the floor plan generation model training process includes:

[0150] Obtaining a historical floor plan training set, wherein the historical floor plan training set includes real images and text descriptions;

[0151] Adding noise to the historical floor plan training set to obtain noisy data and inputting it into a preset model;

[0152] The preset loss function of the preset model is iterated through the noise-added data to obtain a floor plan generation model.

[0153] It should be noted that the device is a device corresponding to the above method, and all implementation methods in the above method embodiments are applicable to the embodiments of the device and can achieve the same technical effects.

[0154] The embodiment of the present invention further provides a computing device, including: a processor, a memory storing a computer program, and when the computer program is executed by the processor, the method described in the above embodiment is executed. All implementations in the above method embodiment are applicable to this embodiment and can achieve the same technical effect.

[0155] In an embodiment of the present invention, a computer-readable storage medium is further provided, which stores instructions, and when the instructions are executed on a computer, the computer executes the method described in the above embodiment. All implementations in the above method embodiment are applicable to this embodiment, and can also achieve the same technical effect.

[0156] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0157] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0158] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0159] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0160] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0161] If the functions are implemented in the form of 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 the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, ROM, RAM, magnetic disks, or optical disks.

[0162] In addition, it should be noted that in the apparatus and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. Moreover, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order, and some steps can be performed in parallel or independently of each other. For those of ordinary skill in the art, it is understood that all or any steps or components of the method and apparatus of the present invention can be implemented in any computing device (including processors, storage media, etc.) or a network of computing devices in hardware, firmware, software or a combination thereof, which can be achieved by those of ordinary skill in the art using their basic programming skills after reading the description of the present invention.

[0163] Therefore, the purpose of the present invention can also be achieved by running a program or a group of programs on any computing device. The computing device can be a well-known general device. Therefore, the purpose of the present invention can also be achieved by simply providing a program product containing a program code that implements the method or device. That is to say, such a program product also constitutes the present invention, and the storage medium storing such a program product also constitutes the present invention. Obviously, the storage medium can be any well-known storage medium or any storage medium developed in the future. It should also be pointed out that in the device and method of the present invention, it is obvious that each component or each step can be decomposed and / or recombined. These decompositions and / or recombinations should be regarded as equivalent schemes of the present invention. In addition, the steps of performing the above-mentioned series of processing can naturally be performed in chronological order according to the order of description, but it is not necessary to perform them in chronological order. Some steps can be performed in parallel or independently of each other.

[0164] The above is a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for reconstructing a floor plan. It is characterized in that include: Get a 2D wireframe of the house type; Extracting feature parameters from the two-dimensional house type wireframe diagram to obtain feature quantities; According to the characteristic quantity, a three-dimensional floor plan is obtained; The three-dimensional floor plan is input into a floor plan generation model to reconstruct the floor plan to obtain a floor plan; wherein the floor plan generation model is obtained by training a preset model based on a historical floor plan training set.

2. The method for reconstructing a floor plan according to claim 1, It is characterized in that Extracting feature parameters from the two-dimensional house type wireframe diagram to obtain feature quantities includes: Preprocessing the two-dimensional house type wireframe diagram; Identify the building structure in the preprocessed two-dimensional house type wireframe diagram to obtain building structure feature quantities; The scale of the preprocessed two-dimensional house type wireframe diagram is identified to obtain a spatial feature quantity.

3. The method for reconstructing a floor plan according to claim 2, It is characterized in that According to the feature quantity, a three-dimensional floor plan is obtained, including: Construct a three-dimensional space framework; The building structure feature quantity is mapped to the three-dimensional space framework, and the building structure feature quantity is adjusted according to the space feature quantity to obtain a three-dimensional apartment plan.

4. The method for reconstructing a floor plan according to claim 1, It is characterized in that Inputting the three-dimensional floor plan into a floor plan generation model to reconstruct the floor plan to obtain the floor plan, including: The floor plan generation model converts the three-dimensional floor plan into a three-dimensional color floor plan; Filling is performed on the three-dimensional color plan to obtain a floor plan.

5. The method for reconstructing a floor plan according to claim 2, It is characterized in that The floor plan generation model converts the three-dimensional floor plan into a three-dimensional color floor plan, including: The floor plan generation model renders colors in the three-dimensional floor plan according to the building structure feature.

6. The method for reconstructing a floor plan according to claim 5, It is characterized in that Filling is performed on the three-dimensional color plan to obtain a floor plan, including: Filling texture on the three-dimensional color plane image to generate a three-dimensional color stereogram; Light and shadow effects are added to the three-dimensional color stereogram to obtain a floor plan.

7. The method for reconstructing a floor plan according to claim 1, It is characterized in that The floor plan generation model training process includes: Obtaining a historical floor plan training set, wherein the historical floor plan training set includes real images and text descriptions; Adding noise to the historical floor plan training set to obtain noisy data and inputting it into a preset model; The preset loss function of the preset model is iterated through the noise-added data to obtain a floor plan generation model.

8. A device for reconstructing a floor plan, It is characterized in that include: An acquisition module is used to obtain a two-dimensional house type wireframe diagram; A processing module, used for extracting feature parameters from the two-dimensional house type wireframe diagram to obtain feature quantities; A three-dimensional floor plan is obtained according to the feature quantity; the three-dimensional floor plan is input into a floor plan generation model to reconstruct the floor plan to obtain the floor plan; wherein the floor plan generation model is obtained by training a preset model based on a historical floor plan training set.

9. A computing device, It is characterized in that include: A processor and a memory storing a computer program, wherein when the computer program is executed by the processor, the method according to any one of claims 1 to 7 is performed.

10. A computer-readable storage medium, It is characterized in that Instructions are stored, and when the instructions are executed on a computer, the computer is caused to execute the method according to any one of claims 1 to 7.