An ai indoor decoration design optimization system based on bim technology

Abstract

The application provides an AI indoor decoration design optimization system based on BIM technology, and relates to the technical field of data processing, comprising a plan design module, a BIM modeling module, a BIM analysis module, a BIM design optimization module and a user interaction module. The plan design module is used for obtaining a decoration design drawing of an indoor space of a user; the BIM modeling module is used for converting the two-dimensional decoration design drawing into a three-dimensional BIM model; the BIM analysis module is used for analyzing based on the BIM model and automatically generating multiple decoration design schemes; the BIM design optimization module is used for optimizing an intended decoration design scheme selected by the user from the multiple decoration design schemes; and the user interaction module is used for displaying the optimized decoration design scheme. The application can automatically generate multiple decoration design schemes for the customer to select, does not need to rely on a decoration designer to manually model, reduces the time cost and improves the indoor decoration design efficiency.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an AI interior decoration design optimization system based on BIM technology. Background Art

[0002] Interior decoration design is something that people often encounter in life. Decoration design is not just about decorating the space, but also focuses on the overall layout of the space, the selection of materials, the matching of colors and the use of light to meet the living and psychological needs of the residents or users.

[0003] BIM (Building Information Modeling) is a digital, three-dimensional modeling technology widely used in the architecture, engineering, and construction (AEC) sectors. By integrating multi-dimensional information, BIM digitizes every phase of a building project—from design and construction to operations and maintenance—to achieve greater efficiency, accuracy, and collaboration.

[0004] Although BIM-based interior design technology currently exists, it often relies on interior designers to plan and manually create 3D models of rooms, then add materials, lighting, furniture, and other elements to the 3D models to create a complete interior design. Intelligent, complete interior design solutions are still lacking, and manual modeling and design are time-consuming and inefficient. Furthermore, manual design methods make it difficult for designers to quickly generate multiple styles or layouts for clients to choose from. Typically, only a few alternatives are available. This lack of diversity and choice limits client choice and the ability to meet individual needs. Summary of the Invention

[0005] In order to solve the technical problems in the existing technology that rely on decoration designers to carry out design planning and manually perform three-dimensional modeling of rooms, which has high time cost and low efficiency of interior decoration design, and in the manual design mode, it is difficult for designers to generate a variety of different styles or layouts for customers to choose from in a short time, and usually only a few alternatives can be provided. The lack of diversity and selection space in the solution generation method limits customers' choices and the satisfaction of personalized needs, the present invention provides an AI interior decoration design optimization system based on BIM technology.

[0006] The technical solutions provided by the embodiments of the present invention are as follows:

[0007] A first aspect of an embodiment of the present invention provides an AI interior decoration design optimization system based on BIM technology, including:

[0008] Graphic design module, used to obtain the decoration design drawings of the user's interior space;

[0009] A BIM modeling module, used to convert the two-dimensional decoration design drawing into a three-dimensional BIM model;

[0010] A BIM analysis module, used to analyze based on the BIM model and automatically generate multiple decoration design solutions;

[0011] BIM design optimization module is used to optimize the intended decoration design scheme selected by the user from multiple decoration design schemes;

[0012] The user interaction module is used to provide an interactive interface between the system and the user and display the optimized decoration design plan.

[0013] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0014] In the present invention, the two-dimensional decoration design drawing can be automatically converted into a three-dimensional BIM model, and based on the analysis of the BIM model, a variety of decoration design schemes can be automatically generated for customers to choose from, so as to meet the personalized needs of customers. The intended decoration design scheme selected by the user from the multiple decoration design schemes can also be optimized without relying on manual modeling by decoration designers, thus reducing time costs and improving the efficiency of interior decoration design. The environmental friendliness and resource sustainability of materials can also be considered in the optimization process, thus helping to realize green buildings and providing users with a healthier and more environmentally friendly indoor environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 A schematic diagram of the structure of an AI interior decoration design optimization system based on BIM technology provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0017] The technical solution of the present invention is described below in conjunction with the accompanying drawings.

[0018] In the embodiments of the present invention, sometimes a subscript such as W1 may be mistakenly written as a non-subscript form such as W1. When the difference is not emphasized, the meanings to be expressed are the same.

[0019] In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0020] Reference Manual Figure 1 , which shows a flow chart of an AI interior decoration design optimization system based on BIM technology provided by an embodiment of the present invention.

[0021] An embodiment of the present invention provides an AI interior decoration design optimization system based on BIM technology, including: a graphic design module, a BIM modeling module, a BIM analysis module, a BIM design optimization module and a user interaction module.

[0022] The graphic design module is used to obtain the decoration design drawings of the user's interior space.

[0023] Specifically, designers or users can upload interior space decoration design drawings.

[0024] BIM modeling module is used to convert two-dimensional decoration design drawings into three-dimensional BIM models.

[0025] In a possible implementation, the BIM modeling module is specifically configured to execute S201 to S204:

[0026] S201: Identify walls, doors, and windows in two-dimensional decoration design drawings using digital image technology based on convolutional neural networks.

[0027] Convolutional Neural Networks (CNNs) are deep learning models specifically designed for processing image and video data. CNNs extract local features from images through convolution operations, offering significant advantages in computational efficiency and recognition accuracy. They are a core technology in fields such as computer vision, image recognition, and object detection.

[0028] Optionally, S201 specifically includes:

[0029] In the input layer of the convolutional neural network, input the decoration design drawing.

[0030] In the convolutional layer of the convolutional neural network, the graphic features in the decoration design picture are extracted.

[0031] In the pooling layer of the convolutional neural network, the extracted line features are subjected to dimensionality reduction.

[0032] In the fully connected layer of the convolutional neural network, the features in the convolutional layer and the pooling layer are summarized.

[0033] In the classification layer of the convolutional neural network, walls, doors, and windows in the decoration design drawings are identified.

[0034] In this paper, CNN-based image recognition technology provides a precise and efficient solution for the automated processing of 2D interior design drawings. By extracting features using convolutional layers, reducing dimensionality using pooling layers, summarizing features using fully connected layers, and identifying features using classification layers, this solution can accurately and efficiently identify walls, doors, and windows in drawings, providing precise data support for subsequent 3D modeling and optimization.

[0035] S202: Determine room size information based on wall, door, and window recognition results.

[0036] Optionally, S202 specifically includes: determining the closed area of ​​the decoration design drawing based on the recognition results of the walls and doors in combination with a connectivity detection algorithm, and determining the size information of each room.

[0037] Connected component labeling is an image processing technique that can label connected regions in an image as independent regions. Using a connectivity detection algorithm, we can identify closed regions in an image, i.e., the outlines of rooms.

[0038] Specifically, the recognition results of walls, doors, and other elements can be binarized to generate a binary image containing the wall and door outlines. For thin wall lines, dilation can be performed to make the wall areas more distinct, thereby improving the accuracy of subsequent connectivity detection. Using the binary image of walls and doors, a connectivity detection algorithm (such as the Flood Fill algorithm or a DFS / BFS traversal) is used to identify all closed spatial regions. This algorithm iterates through the image pixels one by one, marking each connected region and excluding unclosed spaces. When detecting closed areas, door locations are treated as interruptions in wall connections to ensure that their location does not affect the closure detection, thereby accurately identifying the room boundary. The closed area obtained by the connectivity detection algorithm is the room outline. Based on the boundary points of the closed area, the room's length, width, and area are calculated. The pixel values ​​in the image can be converted to actual dimensions using a BIM scale to obtain the accurate room area.

[0039] In this paper, by identifying walls and doors and combining them with a connectivity detection algorithm to determine enclosed areas, we can quickly and accurately obtain room boundaries and dimensions. This method not only enables efficient and standardized data processing but also enhances the intelligence of the design process, laying a solid foundation for subsequent BIM modeling, interior design, and construction management.

[0040] S203: Generate a three-dimensional BIM model based on the identified room size information, walls, doors, and windows.

[0041] Specifically, the basic three-dimensional structure of each room is generated based on the room dimensions, including its length, width, and height. The ground plane can generally be considered the XY plane, with the room height extending along the Z axis. The spatial dimensions of each floor are generated first, and then layered one by one to construct the complete building structure. Wall outlines are generated along the room boundaries based on their planar position and thickness. Using standard building floor heights or specific room height information, the walls are extended from the ground to the ceiling, forming a three-dimensional wall structure. In the BIM model, each wall is assigned attributes such as material, color, and thickness for use in subsequent design and construction phases. For example, different types can be designated as load-bearing, soundproof, or decorative. Next, based on the door locations identified in the recognition results, the door openings are precisely positioned at the corresponding locations in the walls. This ensures that the door's location and dimensions conform to the floor plan annotations. The door's 3D geometry, including the door frame and leaf, is generated based on its width, height, and thickness. In the BIM software, the door's opening and closing direction and angle are set to conform to the requirements of the architectural design drawings. For example, it can be designated as an inward-opening, outward-opening, or sliding door. Assign door attributes such as material, color, and model for reference during later construction and maintenance. Based on the identified window location and size, add window openings to the corresponding walls, ensuring that the window size and location match the design drawings. Generate a 3D window model based on the window width, height, and depth, including the frame and glazing. The frame thickness and material can be customized to meet actual needs. Specify the window orientation and opening method (such as casement, sliding, or fixed) in the BIM model to meet functional and ventilation requirements. Define window properties such as material, transparency, and thermal insulation to provide data support for subsequent daylighting and energy analysis. Based on the room dimensions, add floor components and assign the material, texture, and thickness. For example, choose from wood, tile, or carpet. Generate a ceiling at the upper level of the room to enclose the upper boundary and assign properties such as the material and light fixture opening location. Assign realistic building materials and textures to each component of the BIM model to ensure realism in the 3D model. Choose from a variety of materials, including concrete, brick, plaster, glass, and wood. Add furniture and fixtures (such as kitchen and bathroom equipment, and lighting) to the BIM model based on the functional requirements of the design plan to make it more realistic. You can further refine the room's interior decorative elements, such as curtains and lighting fixtures, to enhance the model's integrity and visual quality.

[0042] After generating a three-dimensional BIM model, the detection function of the BIM software is used to check whether there are problems such as space overlap and structural conflicts in the model to ensure the accuracy and constructability of the model.

[0043] In this paper, the identification information of walls, doors, and windows can be used to automatically generate a 3D BIM model that meets design requirements. This model not only contains the building's 3D spatial structure but also includes rich information on materials, dimensions, and attributes, providing data support for subsequent design optimization, energy consumption analysis, and construction.

[0044] S204: Determine room function information.

[0045] Optionally, S204 specifically includes:

[0046] When there are text annotations in the decoration design drawings, OCR technology is used to determine the functional information of each room.

[0047] Optical Character Recognition (OCR) is a technology used to identify text in images and convert it into editable text. Through image processing and pattern recognition, OCR extracts character information from scanned documents, photos, or other images, thereby digitizing the text content.

[0048] In the absence of text annotations on the decoration design drawings, feature matching technology is used to determine the functional information of each room.

[0049] Specifically, images of common room function symbols (such as sinks, beds, sofas, toilets, and bathtubs) are collected and organized to create a feature template library, with each symbol representing a room function. The interior design drawings are then de-noised and binarized to enhance the contrast between the room symbols and the background, facilitating feature extraction. Methods such as Canny edge detection are used to extract the edges of objects within the room, highlighting the symbol outlines and reducing background interference. Key points are detected within the room symbol region using algorithms such as SIFT (Scale-Invariant Feature Transform), SURF, or ORB, thereby extracting the symbol's features. Descriptors (such as feature vectors) are generated for each key point, converting the symbol's local features into numerical representations to facilitate subsequent matching with the template library. The extracted room symbol features are then matched against the feature templates in the template library. Algorithms such as FLANN (Fast Approximate Nearest Neighbor) or BFMatcher (Brute Force Matching) are used to compare the characteristic symbols in the drawings against the symbols in the template library one by one. Each match result is scored, with higher-scoring templates being more consistent with the symbol in the image. A scoring threshold is set to filter out qualified matches and avoid misidentification.

[0050] This invention utilizes OCR and feature matching technology to automatically identify room function information, enabling fast, efficient, and accurate function annotation, enhancing the intelligence of the design and modeling process. This automated recognition technology not only improves work efficiency and data accuracy, but also makes the entire renovation design process more flexible, intelligent, and user-friendly, providing a solid data foundation for BIM model construction and subsequent applications.

[0051] The BIM analysis module is used to analyze based on the BIM model and automatically generate a variety of decoration design plans.

[0052] In a possible implementation, the BIM analysis module is specifically configured to execute S301 to S303:

[0053] S301: Analyze the room dimensions and room functional attributes of the BIM model.

[0054] S302: Obtain the decoration style preferred by the user.

[0055] S303: Based on the decoration style, room size and room functional attributes, a variety of decoration design plans are automatically generated based on the artificial neural network.

[0056] Artificial neural networks (ANNs) are computational models that mimic the neural networks of the human brain and are widely used in fields such as image recognition, speech recognition, and natural language processing. By mimicking the learning and processing methods of the human brain, neural networks gradually learn the characteristics and patterns of data, thereby completing tasks such as classification and prediction.

[0057] Optionally, S303 specifically includes:

[0058] The decoration feature vector is determined according to the decoration style, room size and room functional attributes.

[0059] In the input layer of the artificial neural network, the decoration feature vector is input.

[0060] Optionally, the decoration characteristic vector includes: room length, width, height, room function, material preference, style preference, and color preference.

[0061] In the hidden layer of the artificial neural network, the input state vector is weighted and summed through each hidden layer neuron to obtain the hidden state vector:

[0062]

[0063] Among them, y j represents the hidden state of the jth hidden layer neuron output, σ1 represents the hidden layer activation function, W jrepresents the weight vector of the jth hidden layer neuron, T represents the transpose operation, X represents the input state vector, b j represents the bias term of the jth hidden layer neuron, ω ij represents the weight of the i-th state parameter in the j-th hidden layer neuron, x i represents the value of the i-th state parameter, and n represents the total number of state parameters.

[0064] In the matching layer of the artificial neural network, the recommendation index of each decoration design scheme is calculated based on the hidden state vector:

[0065]

[0066] Among them, ρ i represents the recommendation index of the i-th decoration design scheme, σ2 represents the output layer activation function, ω ij represents the connection weight between the jth hidden layer neuron and the output node represented by the i-th decoration design scheme, b i It represents the bias term of the output node represented by the i-th decoration design scheme, and m represents the total number of hidden layer neurons.

[0067] In the output layer of the artificial neural network, a preset number of decoration design plans with high recommendation index rankings are output.

[0068] It's important to note that by using artificial neural networks' feature vector inputs and latent state vector calculations, the system can efficiently and automatically generate and recommend a variety of personalized renovation design solutions. This approach not only improves design efficiency and intelligence, but also meets the personalized needs of users, improving user satisfaction and providing strong data support and a convenient user experience throughout the entire design-to-construction process.

[0069] This invention utilizes an automated design approach based on the BIM analysis module, generating design solutions through data analysis and artificial intelligence, making interior design more efficient, accurate, and intelligent. This approach not only provides users with diverse design options but also meets their needs for personalization, functionality, and sustainability, providing solid data support and technical assurance for design, construction, and post-construction management.

[0070] The BIM design optimization module is used to optimize the intended decoration design scheme selected by the user from a variety of decoration design schemes.

[0071] In a possible implementation, the BIM design optimization module is specifically configured to execute S401 and S402:

[0072] S401: Obtaining a desired decoration design scheme selected by a user from among multiple decoration design schemes.

[0073] S402: With the goal of reducing construction energy consumption, improving light transmittance, thermal comfort, and compatibility with decoration styles, an optimization objective function is constructed:

[0074] f(θ)=-λ1E+λ2T+λ3PMV+λ4S

[0075] Where f represents the optimization objective function, θ represents the parameter combination of the decoration design scheme, E represents the construction energy consumption, T represents the transmittance, PMV represents the predicted mean voting index, which is used to evaluate thermal comfort, S represents the decoration style fit, λ1 represents the weight coefficient of construction energy consumption, λ2 represents the weight coefficient of transmittance, λ3 represents the weight coefficient of thermal comfort, and λ4 represents the weight coefficient of decoration style fit.

[0076] Among them, those skilled in the art can set the weight coefficient λ1 of construction energy consumption, the weight coefficient λ2 of transmittance, the weight coefficient λ3 of thermal comfort, and the weight coefficient λ4 of decoration style compatibility according to actual conditions, and the present invention does not limit them.

[0077] Among them, the decoration design parameters include: windows, doors, furniture styles and placement, wall materials, color matching, lighting design, etc.

[0078] It's important to note that by combining the weighted coefficients of four indicators—construction energy consumption, light transmittance, thermal comfort, and style compatibility—a balance can be achieved between different design requirements. Optimizing the objective function improves the practicality and comfort of the space while still satisfying users' aesthetic needs.

[0079]

[0080] Where E represents the construction energy consumption, m i represents the construction area of ​​the i-th decoration construction link, c i It represents the energy consumption per unit area of ​​the i-th decoration construction link, and N represents the total number of decoration construction links.

[0081]

[0082] Where, T represents the transmittance, a j represents the area of ​​the jth light-transmitting area, t j represents the transmittance of the jth light-transmitting area, and M represents the total number of light-transmitting areas.

[0083] PMV=(0.303×e -0.036×M +0.028)×L

[0084] L=(MW)-0.00305×[5733-6.99×(MW)-P a ]

[0085] -0.42×(MW-58.15)-0.000017×(5867-P a )

[0086] -0.0014×M×(34-T a )-3.96×10 -8 ×f d ×[(T cl +273) 4 -(T r +273) 4 ]

[0087] -f cl ×h c ×(T cl -T a )

[0088] Where PMV is the predicted mean vote index, which is used to evaluate thermal comfort, e is the natural constant, M is the metabolic rate, W is the external work, and P is the a represents the water vapor partial pressure, T a Indicates the air temperature, T r represents the radiation temperature, f cl Indicates the clothing surface coefficient, h c represents the convective heat transfer coefficient, T cl Indicates the surface temperature of the garment.

[0089]

[0090] Among them, S represents the degree of fit of the decoration style, ω k represents the weight coefficient of the kth design feature, F k represents the kth design eigenvalue of the decoration design scheme, P k The kth design feature value represents the user’s preferred style, max represents the larger value, and K represents the total number of design features.

[0091] S403: Optimizing the details of the intended decoration design scheme through a genetic algorithm according to the optimization objective function.

[0092] Among them, Genetic Algorithm (GA) is an optimization algorithm that simulates the natural evolution process. It optimizes a set of candidate solutions through operations such as "selection", "crossover" and "mutation" to approach or reach the global optimal solution to the problem.

[0093] Specifically, the optimization objective function can be used as the fitness function of the genetic algorithm.

[0094] Initialize the population Q1, which includes multiple individuals X. Each individual X represents a feasible decoration design parameter combination θ.

[0095] Calculate the fitness value of the fitness function of each individual in the initial population Q1.

[0096] An elite selection strategy is adopted to remove the 20% individuals with the lowest fitness values ​​to form a new population Q2.

[0097] Perform a crossover operation on population Q2, randomly select two individuals from population H2 as parents, generate a random number, and compare the random number with the crossover probability p e Compare the size, if the random number is less than the crossover probability p e , then the parent is cross-operated to generate new individuals to form a new population Q3. The new individuals are generated as follows:

[0098] Y1=rand×X1+(1-rand)×X2

[0099] Y2=rand×X2+(1-rand)×X1

[0100] Among them, Y1 and Y2 represent new individuals, X1 represents the first parent, X2 represents the second parent, and rand represents a random number between 0 and 1.

[0101] Perform mutation operation on population Q3, randomly select an individual from population Q3 as the parent, generate a random number, and compare the random number with the mutation probability p m Compare the size, if the random number is less than the mutation probability p m , then the parent is mutated to generate new individuals to form a new population Q4. The generation of new individuals is as follows:

[0102]

[0103] Among them, Y3 represents the new individual, X3 represents the parent, and X max represents the individual with the largest fitness value, X min represents the individual with the smallest fitness value, and rand represents a random number between 0 and 1.

[0104] Repeat the above steps and iterate until the preset number of iterations is reached. The solution with the largest fitness value is output as the optimal decoration design parameter combination, and the details of the intended decoration design scheme are optimized.

[0105] In this paper, by combining the optimization objective function in the BIM design optimization module with a genetic algorithm, the system achieves comprehensive optimization in multiple aspects, including construction energy consumption, light transmittance, thermal comfort, and style compatibility. This approach not only makes design more scientific, accurate, and efficient, but also meets personalized needs, reduces construction costs, improves user satisfaction and comfort, and achieves higher-quality, more environmentally friendly interior design solutions.

[0106] The user interaction module is used to provide an interactive interface between the system and the user and display the optimized decoration design plan.

[0107] In a possible implementation, the user interaction module is specifically configured to display an optimized decoration design plan and the corresponding estimated cost and estimated decoration duration.

[0108] In this system, the user interaction module displays optimized renovation design plans, estimated costs, and estimated duration, helping users make more informed choices and plans, enhancing their engagement and satisfaction. This information also enhances the system's professionalism, transparency, and practicality, helping users find a balance between budget, time, and results, ultimately achieving a higher-quality renovation experience.

[0109] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least:

[0110] In the present invention, the two-dimensional decoration design drawing can be automatically converted into a three-dimensional BIM model, and based on the analysis of the BIM model, a variety of decoration design schemes can be automatically generated for customers to choose from, so as to meet the personalized needs of customers. The intended decoration design scheme selected by the user from the multiple decoration design schemes can also be optimized without relying on manual modeling by decoration designers, thus reducing time costs and improving the efficiency of interior decoration design. The environmental friendliness and resource sustainability of materials can also be considered in the optimization process, thus helping to realize green buildings and providing users with a healthier and more environmentally friendly indoor environment.

[0111] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. An AI interior decoration design optimization system based on BIM technology, characterized by: include: Graphic design module, used to obtain the decoration design drawings of the user's interior space; A BIM modeling module, used to convert the two-dimensional decoration design drawing into a three-dimensional BIM model; A BIM analysis module, used to analyze based on the BIM model and automatically generate multiple decoration design solutions; BIM design optimization module is used to optimize the intended decoration design scheme selected by the user from multiple decoration design schemes; User interaction module, used to provide an interactive interface between the system and the user and display the optimized decoration design plan; The BIM analysis module is specifically used for: S301: Analyze the room size and room functional attributes of the BIM model; S302: Obtaining the user's preferred decoration style; S303: Automatically generate multiple decoration design solutions based on the artificial neural network according to the decoration style, the room size, and the room functional attributes; The S303 specifically includes: Determining a decoration characteristic vector according to the decoration style, the room size, and the room functional attributes; In the input layer of the artificial neural network, the decoration characteristic vector is input; In the hidden layer of the artificial neural network, the input state vector is weighted and summed through each hidden layer neuron to obtain the hidden state vector; In the matching layer of the artificial neural network, the recommendation index of each decoration design scheme is calculated according to the hidden state vector; In the output layer of the artificial neural network, a preset number of decoration design plans with high recommendation index rankings are output.

2. The AI ​​interior decoration design optimization system based on BIM technology according to claim 1 is characterized in that: The BIM modeling module is specifically used for: S201: Using digital image technology based on convolutional neural networks, identifying walls, doors, and windows in the two-dimensional decoration design drawing; S202: Determine room size information based on wall, door, and window recognition results; S203: Generate a three-dimensional BIM model based on the identified room size information, walls, doors, and windows; S204: Determine room function information.

3. The AI ​​interior decoration design optimization system based on BIM technology according to claim 2 is characterized in that: The S201 specifically includes: Input the decoration design drawing into the input layer of the convolutional neural network; In the convolutional layer of the convolutional neural network, the graphic features of the decoration design are extracted; In the pooling layer of the convolutional neural network, the extracted line features are subjected to dimensionality reduction processing; In the fully connected layer of the convolutional neural network, the features in the convolutional layer and the pooling layer are summarized; In the classification layer of the convolutional neural network, walls, doors, and windows in the decoration design drawings are identified.

4. The AI ​​interior decoration design optimization system based on BIM technology according to claim 2 is characterized in that: The S202 is specifically as follows: Based on the recognition results of walls and doors, combined with the connectivity detection algorithm, the closed areas of the decoration design drawings are determined, and the size information of each room is determined.

5. The AI ​​interior decoration design optimization system based on BIM technology according to claim 2 is characterized in that: The S204 specifically includes: In the case of text annotations on the decoration design drawings, OCR technology is used to determine the functional information of each room; In the absence of text annotations on the decoration design drawings, feature matching technology is used to determine the functional information of each room.

6. The AI ​​interior decoration design optimization system based on BIM technology according to claim 1 is characterized in that: The decoration characteristic vector includes: room length, width, height, room function, material preference, style preference, and color preference.

7. The AI ​​interior decoration design optimization system based on BIM technology according to claim 1 is characterized in that: The BIM design optimization module is specifically used to: S401: Obtaining a desired decoration design scheme selected by a user from among multiple decoration design schemes; S402: Construct an optimization objective function with the goal of reducing construction energy consumption, improving light transmittance, thermal comfort, and compatibility with decoration styles; S403: Optimizing the details of the intended decoration design scheme through a genetic algorithm according to the optimization objective function.

8. The AI ​​interior decoration design optimization system based on BIM technology according to claim 1 is characterized in that: The user interaction module is specifically used for: Display the optimized decoration design plan and the corresponding estimated cost and estimated decoration time.

Images