An interior design method and system based on artificial intelligence

Abstract

The present application provides an interior design method and system based on artificial intelligence, which relates to the field of data analysis and processing. The method includes: obtaining indoor space sensor data and user historical behavior data; constructing an indoor space three-dimensional model according to the indoor space sensor data, and extracting indoor space features from the space three-dimensional model; determining the target indoor design elements preferred by the user according to the user historical behavior data; inputting the indoor space features and the target indoor design elements into a generative adversarial network to determine a corresponding plurality of preliminary indoor design schemes; inputting the plurality of preliminary indoor design schemes into a multi-objective optimization model to determine the target indoor design scheme, and the multi-objective optimization model adopts an improved particle swarm optimization algorithm. Thus, by combining user preferences and indoor space features, the generated design scheme can meet the user's high standard requirements in terms of demand personalization and practical functionality.

Description

Technical Field

[0001] The present application relates to the technical field of data analysis and processing, and in particular to an interior design method and system based on artificial intelligence. Background Art

[0002] In the past decade, the interior design industry has experienced multiple technological innovations from traditional manual drawing to computer-aided design (CAD) and now to intelligent design tools. In recent years, with the breakthroughs in artificial intelligence and machine learning technology, trying to apply these technologies to interior design has become a new trend in the industry. Intelligent interior design tools can automatically generate design plans, predict design trends, and even simulate user reactions to certain design elements.

[0003] However, despite the significant convenience brought by technological advances, existing smart interior design methods still have some technical defects that cannot be ignored. First, most existing smart design tools mainly rely on fixed design templates and style rules, which limits the personalization of design. For example, many tools only provide limited style options during the design process, resulting in the final design output often appearing the same and difficult to meet the unique needs of specific customers.

[0004] Secondly, these tools often ignore the complexity of the real world when dealing with design solutions. Interior design is not only a visual art, but also a comprehensive discipline that combines functionality, comfort and user experience. Existing tools often lack a deep understanding of the actual use functions of living spaces during the automated design generation process, such as insufficient assessment of the practicality of light, circulation or furniture layout.

[0005] In response to the above problems, the industry has not yet proposed a better technical solution. Summary of the invention

[0006] The present application provides an interior design method, system, storage medium, computer program product and electronic device based on artificial intelligence, which are used to at least solve the problem that the templated interior design solutions in the current related technology lack personalization and have poor practical functionality.

[0007] In a first aspect, an embodiment of the present application provides an interior design method based on artificial intelligence, comprising: obtaining indoor space sensor data and user historical behavior data; the user historical behavior data includes design case browsing records, design case feedback records and home accessories purchase records; constructing an indoor space three-dimensional model based on the indoor space sensor data, and extracting indoor space features from the three-dimensional model; the indoor space features include room structure features, furniture layout features and space functional area features; determining the target interior design elements preferred by the user based on the user historical behavior data; the interior design elements include interior design colors, interior design materials, interior design styles and home accessories types; inputting the indoor space features and the target interior design elements into a generative adversarial network to determine a corresponding plurality of preliminary interior design schemes; the preliminary interior design schemes include a color design module, a material design module, a furniture layout module, a light source design module and a space allocation module; inputting the plurality of preliminary interior design schemes into a multi-objective optimization model to determine a target interior design scheme; the multi-objective optimization model adopts an improved particle swarm optimization algorithm.

[0008] In a second aspect, an embodiment of the present application provides an interior design system based on artificial intelligence, comprising: a data acquisition unit, used to acquire indoor space sensor data and user historical behavior data; the user historical behavior data includes design case browsing records, design case feedback records and home accessories purchase records; a space feature extraction unit, used to construct an indoor space three-dimensional model based on the indoor space sensor data, and extract indoor space features from the space three-dimensional model; the indoor space features include room structure features, furniture layout features and space functional area features; a preference element determination unit, used to determine the target interior design elements preferred by the user based on the user historical behavior data; the interior design elements include interior design colors, interior design materials, interior design styles and home accessories types; a preliminary plan determination unit, used to input the indoor space features and the target interior design elements into a generative adversarial network to determine a corresponding plurality of preliminary interior design plans; the preliminary interior design plan includes a color design module, a material design module, a furniture layout module, a light source design module and a space allocation module; a target plan determination unit, used to input the plurality of preliminary interior design plans into a multi-objective optimization model to determine a target interior design plan; the multi-objective optimization model adopts an improved particle swarm optimization algorithm.

[0009] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the artificial intelligence-based interior design method of any embodiment of the present application.

[0010] In a fourth aspect, an embodiment of the present application provides a storage medium on which a computer program is stored, characterized in that when the program is executed by a processor, the steps of the artificial intelligence-based interior design method of any embodiment of the present application are implemented.

[0011] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program / instruction, which, when executed by a processor, implements the steps of the artificial intelligence-based interior design method of any embodiment of the present application.

[0012] The artificial intelligence-based interior design method and system provided by this application can produce at least the following technical effects:

[0013] (1) By obtaining user historical behavior data, including design case browsing records, design case feedback records, and home accessories purchase records, the user's preferences and needs can be accurately identified. Inputting this preference information into the generative adversarial network effectively avoids the lack of personalization caused by the fixed design templates and styles in existing tools, thereby generating a variety of design solutions to meet the user's unique personalized interior design needs.

[0014] (2) A three-dimensional model is constructed through indoor space sensor data, and room structure features, furniture layout features, and spatial functional area features are extracted to ensure that the design scheme is consistent with the actual space environment. The fitness function of the multi-objective optimization model is defined by comprehensively utilizing the illumination simulation evaluation function, air circulation simulation evaluation function, and space utilization simulation evaluation function. This allows the multi-objective optimization model to fully consider the complexity of the real world when performing feature space optimization calculations, especially the practical value of various design schemes, and ensure the practicality and functionality of the target design scheme output in the end.

[0015] Through this technical solution, in the process of design solution generation and optimization, by combining user preferences and indoor space characteristics, the generated design solution is more in line with user expectations. In addition, the multi-objective optimization model can comprehensively consider multiple optimization target directions, such as lighting, air circulation, and space utilization, so that the final output target design solution can meet the user's high standards in terms of demand personalization and practical functionality. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 A flowchart showing an example of an interior design method based on artificial intelligence according to an embodiment of the present application is shown;

[0018] Figure 2 A flowchart of an optimization process of an example of a multi-objective optimization model according to an embodiment of the present application is shown;

[0019] Figure 3 A flowchart showing an example of setting operations for each simulated illumination measurement point according to an embodiment of the present application is shown;

[0020] Figure 4 An operation flow chart showing an example of setting operations for each simulated air measurement point according to an embodiment of the present application is shown;

[0021] Figure 5 A structural block diagram of an example of an interior design system based on artificial intelligence according to an embodiment of the present application is shown;

[0022] Figure 6 It is a schematic structural diagram of an embodiment of an electronic device of the present application. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are 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 this application.

[0024] In the technical solution of this application, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved shall comply with the provisions of relevant laws and regulations and shall not violate public order and good morals.

[0025] Figure 1 A flowchart of an example of an interior design method based on artificial intelligence according to an embodiment of the present application is shown.

[0026] Regarding the executor of the method of the embodiment of the present application, it can be any controller or processor with computing or processing capabilities. By combining sensor data and user behavior data, personalized interior design scheme generation can be achieved, and multi-objective optimization can be performed using an improved particle swarm optimization algorithm. The comprehensive application of artificial intelligence algorithms and multi-objective optimization models can provide users with a more satisfactory interior design experience.

[0027] In some examples, it can be an interior design service platform, and can be integrated and configured in an electronic device or terminal through software, hardware, or a combination of software and hardware, and the type of terminal or electronic device can be diverse, such as a mobile phone, tablet computer, or desktop computer, etc.

[0028] like Figure 1 As shown, in step S110, indoor space sensor data and user historical behavior data are obtained, and the user historical behavior data includes design case browsing records, design case feedback records and home accessories purchase records.

[0029] In some embodiments, for indoor space sensor data, the geometric data of the indoor space is collected in real time through a variety of sensors installed indoors (such as laser radar, camera, infrared sensor), and the sensor data is transmitted to the interior design service platform through the Internet of Things (IoT) platform. For the details of the user's historical behavior data, all design cases that the user has browsed on the design platform can be recorded, including access time, browsing time, collection records, etc., thereby obtaining the corresponding design case browsing records; record the user's rating, comments and specific feedback on each design case, thereby obtaining the corresponding design case feedback record; obtain the user's purchase history on home accessories related websites through the authorized e-commerce platform API interface, including the type, quantity, price and other information of the purchased products, and obtain the corresponding home accessories purchase record through sorting. Through this step, rich and accurate basic data is provided, which is helpful for the accurate analysis of user preferences and indoor space characteristics in the subsequent steps, and provides a basis for the personalization and pertinence of interior design solutions.

[0030] In step S120, a three-dimensional model of the indoor space is constructed according to the indoor space sensing data, and indoor space features are extracted from the three-dimensional model. The indoor space features include room structure features, furniture layout features, and space functional area features.

[0031] In some embodiments, 3D modeling software (e.g., Revit) is used to convert the sensor data into a three-dimensional model, including a detailed three-dimensional geometric model of walls, floors, ceilings, and fixtures. Then, key features are extracted from the three-dimensional model, such as the size, shape, structural features (e.g., door and window locations), furniture layout features (e.g., furniture location and size), and functional area divisions (e.g., living room, bedroom, kitchen, etc.) of the room.

[0032] In some examples of the embodiments of the present application, the indoor space sensing data includes a laser point cloud data set and a group of indoor real-time photos. Specifically, a laser radar (LiDAR) device can be used to scan the indoor space to obtain high-precision three-dimensional point cloud data, and SLAM (simultaneous localization and mapping) technology can be used to achieve real-time positioning and data collection when the device moves indoors, thereby improving the efficiency and accuracy of data collection; in addition, a high-resolution camera is used to take indoor photos from multiple angles to ensure that each area and corner has sufficient image coverage. Then, based on a surface reconstruction algorithm (e.g., Poisson reconstruction or Delaunay triangulation), the laser point cloud data set is converted into a three-dimensional geometric model to obtain the three-dimensional geometric structure of the room. Afterwards, the indoor real-time photo group and the three-dimensional geometric model are aligned and matched based on a feature point matching algorithm to obtain a three-dimensional model of the indoor space. Exemplarily, by using a feature extraction algorithm of SIFT or ORB, key feature points are extracted from the real-time photos and the three-dimensional geometric model, and by using a feature point matching algorithm (e.g., FLANN fast nearest neighbor search package algorithm), the feature points in the photo are matched with the feature points in the geometric model, thereby improving the accuracy of image alignment.

[0033] Further, indoor space features are extracted from the spatial three-dimensional model by using a three-dimensional convolutional neural network 3D CNN. Here, the convolutional neural network can adopt a multi-task CNN model to simultaneously perform plane detection, edge detection and shape matching to improve the efficiency and accuracy of feature extraction. Specifically, the 3D CNN model is used to process the three-dimensional voxel grid data, extract the geometric features and furniture layout of the room, and preliminarily divide the functional areas of the room based on the extracted geometric features. Then, the spatial clustering algorithm (such as the K-means algorithm) is used to optimize the preliminarily divided functional areas to ensure that the area and shape of each functional area are reasonable to obtain the corresponding spatial functional area features. In addition, the plane detection function of the 3D CNN model is used to identify the walls, floors and ceilings in the three-dimensional model, thereby obtaining the room structure features. In addition, based on the training and learning of the data sample set, the 3D CNN model is used to process the three-dimensional voxel grid data, identify and classify the furniture, and thus extract the furniture layout features such as the position, size and type of the furniture.

[0034] In step S130, target interior design elements preferred by the user are determined according to the user's historical behavior data, where the interior design elements include interior design colors, interior design materials, interior design styles, and home accessories types.

[0035] In some implementations, association rule mining (e.g., Apriori algorithm) can be used to extract frequent item sets from the user's browsing history, feedback history, and purchase history to determine the user's preferred interior design elements, including color, material, style, and home accessories type, etc. Thus, personalized design elements ensure that the design scheme meets the user's aesthetic and functional requirements, providing a basis for highly customized design schemes and improving user satisfaction.

[0036] In step S140, the indoor space features and the target indoor design elements are input into a generative adversarial network to determine a corresponding plurality of preliminary indoor design solutions.

[0037] It should be noted that the generative adversarial network GAN includes a generator network (Generator) and a discriminator network (Discriminator). Through the generator network, a preliminary design plan is generated based on the input indoor space characteristics and user preference design elements. Through the discriminator network, it is responsible for evaluating whether the generated design plan is in line with reality and user preferences. More specifically, a large number of annotated interior design cases can be used for training. The generator continuously generates design plans and accepts feedback from the discriminator for improvement. Then, the discriminator improves its own judgment ability by distinguishing between real design cases and the plans generated by the generator. Here, the preliminary interior design plan includes a color design module, a material design module, a furniture layout module, a light source design module, and a space allocation module. Multiple different preliminary design plans that match user preferences are generated through multiple iterations of GAN.

[0038] In the exemplary preliminary interior design plan, the color scheme includes the colors of the walls, floors, ceilings and furniture, the material design plan includes the material types of various surfaces (such as walls, floors, furniture), the furniture layout plan includes the type, quantity, position and placement of furniture, the light source design plan includes the position, type and quantity of light sources, and the space allocation plan includes the area and layout of different functional areas.

[0039] In step S150, a plurality of preliminary interior design solutions are input into a multi-objective optimization model to determine a target interior design solution.

[0040] Here, the multi-objective optimization model can adopt various optimization algorithms, such as genetic algorithm, particle swarm optimization algorithm, etc. It should be noted that in the process of practicing this application, MOGA (multi-objective genetic algorithm) and PSO (particle swarm optimization) are compared, and it is found that in the intelligent interior design system, the PSO algorithm has more prominent advantages. In particular, the PSO algorithm can adapt well to complex and continuous design spaces, and is very suitable for multi-dimensional space optimization problems involved in intelligent interior design.

[0041] In the process of implementing this application, the PSO algorithm was also improved and optimized. Multi-objective optimization was performed by improving the particle swarm optimization algorithm, especially by integrating various practical functions into the fitness function of the particle swarm optimization algorithm. The fitness function was comprehensively determined using the lighting simulation evaluation function, air circulation simulation evaluation function and space utilization evaluation function, so as to achieve optimization of the functional practicality of the design scheme.

[0042] Figure 2 A flowchart of an optimization process for a multi-objective optimization model according to an embodiment of the present application is shown.

[0043] like Figure 2 As shown in the figure, the multi-objective optimization model adopts an improved particle swarm optimization algorithm, which includes fitness function definition, particle initialization, particle iterative optimization and optimal design solution output in the optimization process.

[0044] More specifically, the multi-objective optimization model is used to perform the optimization process in the following ways:

[0045] Each preliminary design scheme is initialized as a corresponding particle; each particle is configured with a corresponding position vector and velocity vector to represent the characteristic vector and change trend of the preliminary design scheme respectively:

[0046] x i =x min +r·(x max -x min ), Formula (1)

[0047] v i =v min +r·(v max -v min ), Formula (2)

[0048] In the formula, x i and v i They represent the initial position vector and initial velocity vector of particle i, respectively, min and x max are the minimum and maximum values ​​of the eigenvectors, v min and v max are the minimum and maximum values ​​of the velocity vector, and r represents a random number between 0 and 1.

[0049] It should be noted that the particle position vector x iRepresents a feature vector of a preliminary design. In combination with the example of this technical scenario, can the various dimensions in the preliminary design be feature-encoded, such as wall color (one of three colors, coded as 1, 2, 3), floor material (one of three materials, coded as 1, 2, 3), furniture position (the position of each piece of furniture is represented by a two-dimensional coordinate), light source position (the position of each light source is represented by a two-dimensional coordinate), and functional area allocation (expressed as a percentage). Accordingly, the exemplary preliminary design feature vector sequence x i Maybe x i =[1,2,(3,4),(5,6),30%], each feature in the sequence represents the wall color, floor material, the position coordinates of the first furniture, the position coordinates of the first light source, and the area ratio of the living area. In addition, the particle velocity vector v i represents the rate of change of the design solution features in the feature space. The velocity vector sequence of the position vector can be v i =[0.1,-0.2,(0.3,-0.4),(-0.5,0.6),1%], each feature in the sequence represents the rate of change of wall color (color coding change), floor material (material coding change), furniture position, light source position and living area. Thus, by initializing these vectors, each particle represents a specific design scheme and its changing trend in the feature space, and in the iterative process of the PSO algorithm, the speed and position of the particle will be continuously updated, and finally the optimal interior design scheme will be found.

[0050] By introducing adaptive inertia weight and learning factor, the algorithm has different search capabilities at different stages. The corresponding inertia weight adjustment formula is:

[0051]

[0052] In the formula, w represents the inertia weight, w max and w min Represent the maximum and minimum values ​​of the inertia weight respectively, t represents the iteration index, t max Indicates the maximum number of iterations.

[0053] Through the above operations, the inertia weight is dynamically adjusted from large to small, enhancing the global search capability in the early stage and the local search capability in the later stage.

[0054] Learning factor dynamic adjustment formula:

[0055]

[0056] In the formula, c1 represents the individual learning factor, which controls the degree to which the particle learns its own optimal position; c2 represents the group learning factor, which controls the degree to which the particle learns the global optimal position; 1_start and c 2_start is the starting value of the learning factor, c 1_end and c 2_end is the final value of the learning factor.

[0057] Here, the local search ability of particles is enhanced by the individual learning factor c1, so that particles can use their own historical experience to search, and a larger c1 value will make particles more inclined to approach their own optimal position. The global search ability of particles is enhanced by the group learning factor c2, so that particles can use the overall experience of the group to search, and a larger c2 value will make particles more inclined to approach the global optimal position. Therefore, balancing local search and global search in the calculation optimization process helps to improve the convergence speed and accuracy of the algorithm and reduce the risk of falling into a local optimal solution.

[0058] Then, the elite strategy is introduced in the model iteration process. In each iteration, a part of the best particles is retained to prevent the optimal solution from being replaced by a new suboptimal solution, thus ensuring the stability of the global optimal solution.

[0059] More specifically, during each iteration, population crossover and population mutation operations are performed:

[0060]

[0061] In the formula, x i and x j denote the selected particles i and j respectively, α is the crossover coefficient between 0 and 1, δ is the coefficient of variation, and N(0,1) is a standard normal distribution random number; represents the new particles generated by the population crossover operation, Represents new particles generated by population mutation operations.

[0062] Here, according to the fitness value, particles with high fitness are selected to enter the crossover operation, two particles are selected with a certain probability for the crossover operation, new particles are generated, and then the crossover particles are mutated with a certain probability. Therefore, in each iteration, population crossover and mutation operations are performed to enhance the diversity and global search capabilities of the algorithm.

[0063] The particle update formula is:

[0064]

[0065] In the formula, x i (t) and v i(t) represent the position vector and velocity vector of particle i at the tth iteration respectively; g best represents the global optimal position of all particles; represents the best historical position of particle i; r1 and r2 represent random numbers between 0 and 1 respectively.

[0066] After each iteration, the top k% particles with the highest fitness are selected as elite particles.

[0067] At the beginning of the next iteration, the elite particles are directly retained without updating their speed and position.

[0068] In this way, after each iteration, by retaining the elite particles with the highest fitness, the stability and continuity of the optimal solution can be ensured, preventing it from being replaced by a new suboptimal solution, improving the quality of the solution, and accelerating the convergence process.

[0069] If in d consecutive iterations, there exists |f(g best (t m ))-f(g best (t m -1))|<∈, then it is determined that the convergence judgment condition is met, and according to the corresponding t m The particle with the highest fitness in the iteration results determines the target interior design solution; ∈ represents the preset fitness change threshold, f(g best (t)) represents the global optimal position of all particles at the tth m The fitness function value of the iteration.

[0070] In this way, judging whether the algorithm has converged by the change in fitness of continuous iterations can avoid premature stopping and ensure that the solution found is optimal, thereby enhancing the robustness of the optimization process.

[0071] The fitness function F(x) of the improved particle swarm optimization algorithm is determined by the following method:

[0072] F(x)=β1f1(x)+β2f2(x)+β3f3(x) , formula (10)

[0073] Wherein, f1(x) represents the illumination simulation evaluation function, f2(x) represents the air circulation simulation evaluation function and f3(x) represents the space utilization simulation evaluation function; β1, β2, β3 are the weight coefficients of each evaluation function respectively.

[0074] Through the lighting simulation evaluation function, we ensure that the lighting intensity of the design scheme is evenly distributed at each measuring point, optimize the combination of natural light and artificial light, reduce dependence on artificial lighting, and reduce energy consumption. Through the air circulation simulation evaluation function, we evaluate the air velocity, air quality and temperature distribution in the design scheme to ensure good indoor air circulation, high air quality and suitable temperature, and provide a healthy and comfortable living environment. Through the space utilization simulation evaluation function, we evaluate the area ratio and layout flexibility of each functional area achieved by the design scheme to ensure reasonable space allocation, flexible layout, and improve space utilization efficiency and adaptability.

[0075] It should be understood that the illumination simulation evaluation function, the air circulation simulation evaluation function, and the space utilization simulation evaluation function can be implemented using various currently related evaluation functions or simulation evaluation software modules, and no limitation is imposed here. Furthermore, in the process of practicing this application, the above-mentioned various evaluation functions are also optimized and improved, and more details will be described below in conjunction with other examples.

[0076] Through this embodiment, an improved particle swarm optimization algorithm is adopted, and adaptive inertia weights and dynamically adjusted learning factors are introduced into the multi-objective optimization model, so that the algorithm has different search capabilities at different stages. Through population crossover and mutation operations, the global search capability and local search accuracy of the algorithm are improved. Elite particles are retained in each iteration process to ensure that high-quality design solutions can be retained and further optimized in subsequent iterations, effectively improving the accuracy and quality of the design solutions. In addition, by setting the fitness change threshold and convergence judgment conditions, the improved particle swarm optimization algorithm can converge quickly during the optimization process of the design solution, reducing the number of iterations and calculation time, thereby significantly improving the design efficiency while maintaining the optimization quality, and meeting the user's demand for quickly generating high-quality design solutions.

[0077] By setting the fitness function through comprehensive lighting simulation evaluation, air circulation simulation evaluation and space utilization simulation evaluation, we can comprehensively evaluate the practicality of each functional aspect of the design solution, ensure that the final design solution is optimal in terms of practical function, and thus guarantee the quality of the design solution.

[0078] In some examples of the embodiments of the present application, the expression of f1(x) is:

[0079]

[0080] Where, L s and They represent the total illumination intensity and natural illumination intensity of the sth simulated illumination measurement point in the three-dimensional model of the indoor space, ε represents the natural light weight coefficient, N represents the total number of simulated illumination measurement points, L target Indicates the target light intensity.

[0081] It should be noted that simulated lighting measurement points refer to certain specific positions preset in the three-dimensional model space for measuring light intensity. These measurement points can be set through simulation software (such as DIALux or Relux) to perform lighting simulation.

[0082] Here, the natural light weight ε is used to measure the impact of natural light on the total light intensity, and can be evaluated in two stages (e.g., daytime and nighttime) with different values. For example, during the daytime, the importance of natural light is high, and the natural light weight is set to 0.8, while at night, the importance of natural light is low, and the natural light weight is set to 0.2.

[0083] In this embodiment, by comprehensively considering the intensity of natural light and artificial light, the lighting simulation evaluation is made more comprehensive. The natural light introduction weight ε can adjust the lighting quality under different conditions during the day and night, ensure uniform light distribution, reduce areas with too strong or too weak light, and improve the comfort and efficiency of indoor lighting. Through reasonable lighting simulation, it is ensured that appropriate light intensity can be obtained in each functional area, ensuring the lighting practicality of the design scheme.

[0084] In one example of the embodiment of the present application, the simulated lighting measurement points can be set by matching the features of the indoor three-dimensional space model and can be adjusted manually, such as setting measurement points in the corners, central positions and near windows of the room, and ensuring that at least one measurement point is set for each functional area such as the living area, dining area, and work area. In another example of the embodiment of the present application, the measurement point layout can be adaptively and dynamically designed or adjusted according to the indoor three-dimensional space model.

[0085] Figure 3 A flowchart of an example of setting operations for each simulated illumination measurement point according to an embodiment of the present application is shown.

[0086] like Figure 3 As shown, in step S310, a preset number N is pre-arranged in the three-dimensional model of the indoor space. initial initial lighting measurement points to ensure that each functional area of ​​the 3D model of the indoor space contains at least one initial lighting measurement point.

[0087] Specifically, the measuring points are initially set according to the geometric structure and functional area division of the room, and at least one measuring point is set for each functional area. In addition, the positions of the initial measuring points can be distributed in a grid pattern according to the length and width of the room.

[0088] In step S320, the propagation path of light in the indoor space is simulated based on ray tracing technology and the building surrounding environment information, and the total light intensity of each initial light measurement point is sampled.

[0089] It should be noted that the intensity and distribution of natural light are affected by the surrounding environment of the building, such as the obstruction of surrounding buildings, terrain undulations and vegetation coverage. Specifically, the building's surrounding environment information (including surrounding buildings, terrain and vegetation information, etc.) is integrated with the indoor three-dimensional model to perform comprehensive lighting simulation. For example, ray tracing simulation software is used to simulate the propagation path of light in the room, and the light intensity of each measuring point is sampled and analyzed. Therefore, by comprehensively considering the obstruction factors of the surrounding environment, the distribution of natural light in the room can be more accurately evaluated.

[0090] In step S330, the total number of measurement points N of the simulated illumination measurement points is determined according to the total illumination intensity of each initial illumination point and the target illumination intensity:

[0091]

[0092] Where, L a It represents the total illumination intensity of the initial illumination point a; b represents the adjustment coefficient of the number of illumination measurement points.

[0093] For example, assuming that the room is 5 meters by 7 meters, 5 measuring points are initially set. The light intensity of each measuring point is calculated by ray tracing technology as follows:

[0094] -Measurement point 1: L1 = 250 lux

[0095] -Measurement point 2: L2 = 300 lux

[0096] -Measurement point 3: L3 = 200 lux

[0097] -Measurement point 4: L4 = 150 lux

[0098] -Measurement point 5: L5 = 350 lux

[0099] Assuming the target light intensity is 300 lux, the natural light weight coefficient is 0.5, and the adjustment coefficient is 1, the solution is obtained according to the above formula (12):

[0100] N new =5+1(0.1667+0+0.3333+0.5+0.1667)=5+1.1667=6.1667≈6

[0101] The final number of measurement points after adjustment was 6, and they were redistributed in areas with large changes in light intensity.

[0102] In this embodiment, by dynamically adjusting the positions of measuring points, increasing the density of measuring points in areas where light intensity changes greatly, reducing lighting blind spots, and ensuring that all key areas are covered by measuring points, the optimized lighting measuring point layout can more accurately reflect the distribution of light intensity and improve the overall lighting quality and comfort.

[0103] In step S340, N simulated illumination measurement points are rearranged in the indoor space three-dimensional model according to the light intensity change distribution information indicated by the total illumination intensity of each initial illumination point.

[0104] In this embodiment, the positions and number of measuring points are dynamically adjusted according to the light intensity distribution of the initial measuring points. For example, the density of measuring points in areas with large changes in light intensity is increased, and the density of measuring points in areas with small changes in light intensity is reduced.

[0105] The expression of f2(x) is:

[0106]

[0107] Where V h , Q h and T h represent the air velocity, air quality index and temperature of the hth simulated air measurement point in the three-dimensional model of the indoor space, τ and η represent the weight coefficients, V target represents the target air velocity, and M represents the total number of measuring points.

[0108] It should be noted that simulated air measurement points refer to certain specific locations preset in the three-dimensional model space for measuring relevant parameters such as air flow rate and air quality. These measurement points can be set through various CFD simulation software (for example, ANSYS Fluent, OpenFOAM) to record the air flow rate and quality at each measurement point.

[0109] In some examples of the embodiments of the present application, the number and location of air measurement points can be arranged using a dynamic measurement point adjustment algorithm. The dynamic measurement point adjustment algorithm dynamically adjusts the location and number of measurement points by analyzing the indoor air flow in real time to ensure that the measurement points cover all key areas and can accurately reflect the distribution of air velocity, air quality and temperature.

[0110] Figure 4 An operation flow chart of an example of setting operations for each simulated air measurement point according to an embodiment of the present application is shown.

[0111] like Figure 4 As shown, in step S410, a preset number M is pre-arranged in the three-dimensional model of the indoor space. initial initial air measurement points to ensure that each functional area of ​​the 3D model of the indoor space contains at least one initial air measurement point.

[0112] In some embodiments, measurement points are initially set according to the geometric structure of the room and the locations of vents, doors and windows, with at least one measurement point being set for each functional area. In addition, the initial measurement point locations can be distributed in a grid pattern according to the length and width of the room.

[0113] In step S420, the air flow path is simulated based on fluid dynamics technology and building surrounding environment information, and the air flow rate, air quality and temperature of each initial air measurement point are sampled.

[0114] It should be noted that indoor air flow is affected by the surrounding environment of the building, such as wind speed, wind direction, obstruction of surrounding buildings and terrain undulations. Computational fluid dynamics technology (CFD simulation software, such as ANSYS Fluent, OpenFOAM) can be combined with the external environment for simulation, and by setting boundary conditions and wind field data, the air flow situation can be evaluated and the distribution of air velocity, air quality and temperature can be analyzed. Therefore, by comprehensively considering the obstruction and diversion factors of the surrounding environment, the indoor air velocity and direction can be more accurately evaluated.

[0115] In step S430, the total number of simulated air measuring points M is determined according to the comprehensive air flow parameters of each initial air measuring point and the target comprehensive air flow parameters:

[0116] P e =V e +τQ e +ηT e , Formula (14)

[0117]

[0118] Where P e Represents the comprehensive air flow parameters of the initial air measurement point e; V e , Q e and T e They represent the air velocity, air quality index and temperature of the initial air measuring point e respectively; l represents the adjustment coefficient of the number of air measuring points.

[0119] For example, assuming that the room is 5 meters by 7 meters, 5 measuring points are initially set. The air flow parameters of each measuring point are calculated by CFD simulation as follows:

[0120] -Measurement point 1: V1=0.4m / s, Q1=40, T1=22 degrees Celsius

[0121] -Measurement point 2: V2 = 0.6 m / s, Q2 = 30, T2 = 21 degrees Celsius

[0122] -Measurement point 3: V3 = 0.5 m / s, Q3 = 50, T3 = 20 degrees Celsius

[0123] -Measurement point 4: V4 = 0.3 m / s, Q4 = 60, T4 = 23 degrees Celsius

[0124] -Measurement point 5: V5 = 0.7 m / s, Q5 = 20, T5 = 19 degrees Celsius

[0125] Assuming the target comprehensive air flow parameter P target =0.6m / s, air quality weight is 0.01, temperature weight is 0.05, and adjustment coefficient is 1.

[0126] According to formula (14) and (15), we can know that:

[0127] M=5+1(2.1667+2.25+2.3333+2.4167+2.0833)=5+11.25=16.25≈16

[0128] The final adjusted number of measuring points was 16, which were redistributed in areas with large changes in air velocity, air quality and temperature.

[0129] In this embodiment, by increasing the density of measurement points in areas where the air velocity changes greatly, the air flow can be accurately monitored and optimized, which is conducive to optimizing the ventilation design, ensuring smooth indoor air circulation, effectively removing harmful gases and particulate matter, and improving the health level of the living and working environment.

[0130] In step S440, M simulated air measuring points are rearranged in the indoor space three-dimensional model according to the air flow change information indicated by the comprehensive air flow parameters of each initial air measuring point.

[0131] In this embodiment, the position and number of the measuring points are dynamically adjusted according to the air flow conditions of the initial measuring points. For example, the density of measuring points in areas with large air velocity changes is increased, and the density of measuring points in areas with small air velocity changes is reduced.

[0132] The expression of f3(x) is:

[0133]

[0134] In the formula, A u represents the area of ​​the uth functional area, A total Represents the total indoor area, R u Indicates the area ratio of the target functional area, F u represents the flexibility index of spatial layout, ψ represents the adjustment weight coefficient, and P represents the total number of functional areas.

[0135] It should be noted that the spatial layout flexibility index F u It is used to measure whether the spatial layout can be flexibly adjusted to meet different usage requirements. It is usually measured by the modularity and versatility of the spatial design. For example, when the design involves an open spatial layout, which can adjust the functional areas by moving partitions, it is considered to have a high F. uTherefore, through the space utilization simulation evaluation function, the area ratio and layout flexibility of each functional area under the design scheme are evaluated to ensure reasonable space allocation, flexible layout, and improve space utilization efficiency and adaptability.

[0136] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of actions combined, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application. In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0137] Figure 5 A structural block diagram of an example of an artificial intelligence-based interior design system according to an embodiment of the present application is shown.

[0138] like Figure 5 As shown, the artificial intelligence-based interior design system 500 includes a data acquisition unit 510 , a space feature extraction unit 520 , a preference element determination unit 530 , a preliminary solution determination unit 540 and a target solution determination unit 550 .

[0139] The data acquisition unit 510 is used to acquire indoor space sensor data and user historical behavior data; the user historical behavior data includes design case browsing records, design case feedback records and home accessories purchase records.

[0140] The space feature extraction unit 520 is used to construct a three-dimensional model of the indoor space according to the indoor space sensing data, and extract indoor space features from the three-dimensional model; the indoor space features include room structure features, furniture layout features and space functional area features.

[0141] The preference element determination unit 530 is used to determine the target interior design elements preferred by the user according to the user historical behavior data; the interior design elements include interior design colors, interior design materials, interior design styles and home accessories types.

[0142] The preliminary scheme determination unit 540 is used to input the indoor space characteristics and the target interior design elements into the generative adversarial network to determine a corresponding plurality of preliminary interior design schemes; the preliminary interior design schemes include a color design module, a material design module, a furniture layout module, a light source design module and a space allocation module.

[0143] The target solution determination unit 550 is used to input the multiple preliminary interior design solutions into a multi-objective optimization model to determine a target interior design solution; the multi-objective optimization model adopts an improved particle swarm optimization algorithm.

[0144] In some embodiments, an embodiment of the present application provides a non-volatile computer-readable storage medium, which stores one or more programs including execution instructions, and the execution instructions can be read and executed by an electronic device (including but not limited to a computer, a server, or a network device, etc.) to execute the steps of any of the above-mentioned artificial intelligence-based interior design methods of the present application.

[0145] In some embodiments, the embodiments of the present application also provide a computer program product, which includes a computer program stored on a non-volatile computer-readable storage medium, and the computer program includes program instructions. When the program instructions are executed by a computer, the computer performs any step of the above-mentioned artificial intelligence-based interior design method.

[0146] In some embodiments, the embodiments of the present application also provide an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the steps of the artificial intelligence-based interior design method.

[0147] Figure 6 is a schematic diagram of the hardware structure of an electronic device for executing an interior design method based on artificial intelligence provided by another embodiment of the present application, such as Figure 6 As shown, the device includes:

[0148] One or more processors 610 and memory 620, Figure 6 A processor 610 is taken as an example.

[0149] The apparatus for executing the artificial intelligence-based interior design method may further include: an input device 630 and an output device 640 .

[0150] The processor 610, the memory 620, the input device 630 and the output device 640 may be connected via a bus or other means. Figure 6 The example of connecting through bus is taken in the following.

[0151] The memory 620, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as program instructions / modules corresponding to the artificial intelligence-based interior design method in the embodiment of the present application. The processor 610 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 620, that is, realizing the artificial intelligence-based interior design method in the above method embodiment.

[0152] The memory 620 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 620 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 620 may optionally include a memory remotely arranged relative to the processor 610, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0153] The input device 630 can receive input digital or character information and generate signals related to user settings and function control of the electronic device. The output device 640 can include a display device such as a display screen.

[0154] The one or more modules are stored in the memory 620, and when executed by the one or more processors 610, the artificial intelligence-based interior design method in any of the above method embodiments is executed.

[0155] The above-mentioned product can execute the method provided in the embodiment of the present application, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided in the embodiment of the present application.

[0156] The electronic devices of the embodiments of the present application exist in various forms, including but not limited to:

[0157] (1) Mobile communication equipment: This type of equipment is characterized by having mobile communication functions and its main purpose is to provide voice and data communications. This type of terminal includes: smart phones, multimedia phones, functional phones, and low-end phones.

[0158] (2) Ultra-mobile personal computer devices: These devices belong to the category of personal computers, have computing and processing functions, and generally also have mobile Internet access features. These terminals include: PDA, MID and UMPC devices, etc.

[0159] (3) Portable entertainment devices: These devices can display and play multimedia content. They include audio and video players, handheld game consoles, e-books, smart toys, and portable car navigation devices.

[0160] (4) Other onboard electronic devices with data interaction functions, such as on-board devices installed in vehicles.

[0161] The device embodiments described above are merely illustrative, wherein 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 modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0162] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a general hardware platform, and of course, by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An interior design method based on artificial intelligence, comprising: Obtain indoor space sensor data and user historical behavior data; The user historical behavior data includes design case browsing records, design case feedback records and home accessories purchase records; Constructing a three-dimensional model of the indoor space according to the indoor space sensing data, and extracting indoor space features from the three-dimensional model; the indoor space features include room structure features, furniture layout features and space functional area features; determining a target interior design element preferred by the user based on the user historical behavior data; Interior design elements include interior design colors, interior design materials, interior design styles, and home accessories types; Inputting the indoor space features and the target indoor design elements into a generative adversarial network to determine a corresponding plurality of preliminary indoor design schemes; the preliminary indoor design schemes comprising a color design module, a material design module, a furniture layout module, a light source design module and a space allocation module; Inputting the plurality of preliminary interior design schemes into a multi-objective optimization model to determine a target interior design scheme; the multi-objective optimization model adopts an improved particle swarm optimization algorithm; The fitness function F(x) of the improved particle swarm optimization algorithm is determined by the following method: F(x)=β1f1(x)+β2f2(x)+β3f3(x) Where, f1(x) represents the illumination simulation evaluation function, f2(x) represents the air circulation simulation evaluation function, and f3(x) represents the space utilization simulation evaluation function; β1, β2, β3 are the weight coefficients of each evaluation function respectively; The expression of f1(x) is: Where, L s and They represent the total illumination intensity and natural illumination intensity of the sth simulated illumination measurement point in the three-dimensional model of the indoor space, ε represents the natural light weight coefficient, N represents the total number of simulated illumination measurement points, L target Indicates the target light intensity; The expression of f2(x) is: Where V h , Q h and T h represent the air velocity, air quality index and temperature of the hth simulated air measurement point in the three-dimensional model of the indoor space, τ and η represent the weight coefficients, V target represents the target air velocity, and M represents the total number of measuring points; The expression of f3(x) is: In the formula, A u represents the area of ​​the uth functional area, A total Represents the total indoor area, R u Indicates the area ratio of the target functional area, F u represents the flexibility index of spatial layout, ψ represents the adjustment weight coefficient, and P represents the total number of functional areas.

2. The method according to claim 1, wherein: The multi-objective optimization model is used to perform optimization processing in the following manner: Each preliminary design scheme is initialized as a corresponding particle; each particle is configured with a corresponding position vector and velocity vector to represent the characteristic vector and change trend of the preliminary design scheme respectively: x i =x min +r·(x max -x min ) v i =v min +r·(v max -v min ) In the formula, x i and v i They represent the initial position vector and initial velocity vector of particle i, respectively, min and x max are the minimum and maximum values ​​of the eigenvectors, v min and v max are the minimum and maximum values ​​of the velocity vector, respectively, and r represents a random number between 0 and 1; By introducing adaptive inertia weight and learning factor, the algorithm has different search capabilities at different stages. The corresponding inertia weight adjustment formula is: In the formula, w represents the inertia weight, w max and w min Represent the maximum and minimum values ​​of the inertia weight respectively, t represents the iteration index, t max Indicates the maximum number of iterations; Learning factor dynamic adjustment formula: In the formula, c1 represents the individual learning factor, which controls the degree to which the particle learns its own optimal position; c2 represents the group learning factor, which controls the degree to which the particle learns the global optimal position; 1_start and c 2_start is the starting value of the learning factor, c 1_end and c 2_end is the final value of the learning factor; In each iteration, population crossover and population mutation operations are performed: In the formula, x i and x j denote the selected particles i and j respectively, α is the crossover coefficient between 0 and 1, δ is the coefficient of variation, and N(0,1) is a standard normal distribution random number; represents the new particles generated by the population crossover operation, Represents new particles generated by population mutation operation; The particle update formula is: x i (t+1)=x i (t)+v i (t+1) In the formula, x i (t) and v i (t) represent the position vector and velocity vector of particle i at the tth iteration respectively; g best represents the global optimal position of all particles; represents the best historical position of particle i; r1 and r2 represent random numbers between 0 and 1 respectively; After each iteration, the top k% particles with the highest fitness are selected as elite particles; At the beginning of the next iteration, the elite particles are directly retained without updating their speed and position; If in d consecutive iterations, there exists |f(g best (t m ))-f(g best (t m -1))|<∈, then it is determined that the convergence judgment condition is met, and according to the corresponding t m The particle with the highest fitness in the iteration results determines the target interior design solution; ∈ represents the preset fitness change threshold, f(g best (t)) represents the global optimal position of all particles at the tth m The fitness function value of the iteration.

3. The method according to claim 2, wherein: The setting operation for each of the simulated illumination measurement points includes: A preset number N of pre-arrangements are provided in the three-dimensional model of the indoor space. initial initial lighting measurement points to ensure that each functional area of ​​the three-dimensional model of the indoor space includes at least one initial lighting measurement point; Based on ray tracing technology and building surrounding environment information, simulate the propagation path of light in the indoor space, and sample the total light intensity of each initial light measurement point; According to the total illumination intensity and target illumination intensity of each of the initial illumination measurement points, the total number of measurement points N of the simulated illumination measurement points is determined: Where, L a represents the total light intensity at the initial light measurement point a; b represents the adjustment coefficient of the number of light measurement points; According to the light intensity change distribution information indicated by the total light intensity of each of the initial light measurement points, N simulated light measurement points are rearranged in the indoor space three-dimensional model.

4. The method according to claim 2, wherein: The setting operations for each of the simulated air measurement points include: A preset number M is pre-arranged in the three-dimensional model of the indoor space initial initial air measurement points to ensure that each functional area of ​​the indoor space three-dimensional model contains at least one initial air measurement point; Simulating the air flow path based on fluid dynamics technology and building surrounding environment information, sampling the air flow rate, air quality and temperature of each of the initial air measurement points; According to the comprehensive air flow parameters of each of the initial air measurement points and the target comprehensive air flow parameters, the total number of measurement points M of the simulated air measurement points is determined: P e =V e +τQ e +ηT e Where P e Represents the comprehensive air flow parameters of the initial air measurement point e; V e , Q e and T e They represent the air velocity, air quality index and temperature of the initial air measuring point e respectively; l represents the adjustment coefficient of the number of air measuring points; According to the air flow change information indicated by the comprehensive air flow parameters of each of the initial air measurement points, the M simulated air measurement points are rearranged in the indoor space three-dimensional model.

5. The method according to any one of claims 1 to 4, wherein: The indoor space sensing data includes a laser point cloud data set and a set of indoor real-life photos; The step of constructing a three-dimensional model of the indoor space according to the indoor space sensing data comprises: Based on a surface reconstruction algorithm, the laser point cloud data set is converted into a three-dimensional geometric model; The indoor real-life photo group and the three-dimensional geometric model are aligned and matched based on a feature point matching algorithm to obtain a three-dimensional model of the indoor space.

6. The method according to claim 5, wherein: The step of extracting indoor space features from the spatial three-dimensional model comprises: Indoor space features are extracted from the three-dimensional model of the space based on a three-dimensional convolutional neural network.

7. An interior design system based on artificial intelligence, comprising: A data acquisition unit, used to acquire indoor space sensor data and user historical behavior data; The user historical behavior data includes design case browsing records, design case feedback records and home accessories purchase records; A space feature extraction unit is used to construct a three-dimensional model of the indoor space according to the indoor space sensing data, and extract indoor space features from the three-dimensional model; the indoor space features include room structure features, furniture layout features and space functional area features; A preference element determination unit, used to determine a target interior design element preferred by the user according to the user historical behavior data; Interior design elements include interior design colors, interior design materials, interior design styles, and home accessories types; A preliminary scheme determination unit is used to input the indoor space characteristics and the target indoor design elements into a generative adversarial network to determine a corresponding plurality of preliminary indoor design schemes; the preliminary indoor design schemes include a color design module, a material design module, a furniture layout module, a light source design module and a space allocation module; A target solution determination unit, configured to input the plurality of preliminary interior design solutions into a multi-objective optimization model to determine a target interior design solution; The multi-objective optimization model adopts an improved particle swarm optimization algorithm; The fitness function F(x) of the improved particle swarm optimization algorithm is determined by the following method: F(x)=β1f1(x)+β2f2(x)+β3f3(x) Where, f1(x) represents the illumination simulation evaluation function, f2(x) represents the air circulation simulation evaluation function, and f3(x) represents the space utilization simulation evaluation function; β1, β2, β3 are the weight coefficients of each evaluation function respectively; The expression of f1(x) is: Where, L s and They represent the total illumination intensity and natural illumination intensity of the sth simulated illumination measurement point in the three-dimensional model of the indoor space, ε represents the natural light weight coefficient, N represents the total number of simulated illumination measurement points, L target Indicates the target light intensity; The expression of f2(x) is: Where V h , Q h and T h represent the air velocity, air quality index and temperature of the hth simulated air measurement point in the three-dimensional model of the indoor space, τ and η represent the weight coefficients, V target represents the target air velocity, and M represents the total number of measuring points; The expression of f3(x) is: In the formula, A u represents the area of ​​the uth functional area, A total Represents the total indoor area, R u Indicates the area ratio of the target functional area, F u represents the flexibility index of spatial layout, ψ represents the adjustment weight coefficient, and P represents the total number of functional areas.

Images