Intelligent space distribution method based on layout constraint

Abstract

The invention relates to the technical field of intelligent space distribution, and discloses an intelligent space distribution method based on layout constraint, which comprises the following steps: 1) planning and designing process; 2) an image processing and computer vision identification process; 3) analyzing the process in combination with a mathematical model; 4) finding out a distribution rule and a hot spot region of the spatial data through clustering analysis; and 5) an intelligent space regression global distribution model. The method for finding out the distribution rule and the hot spot area of the spatial data through clustering analysis comprises the four steps of data preparation, clustering algorithm selection, clustering calculation, hot spot area recognition and distribution rule analysis, clustering algorithm selection and K-means clustering algorithm adoption, the method is suitable for the conditions that data distribution is relatively uniform and cluster shapes are relatively regular, the number K of clusters needs to be specified in advance, and the number K of the clusters needs to be determined in advance. And then clustering calculation is carried out, parameters are firstly determined, corresponding parameters are determined according to the selected algorithm, for example, a K value in a K-means algorithm can be determined through methods such as experience, multiple tests or cross validation.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent space distribution, and specifically provides an intelligent space distribution method based on layout constraints. Background Art

[0002] There are often various constraint conditions in spatial layout, such as the limitations of physical space, the constraints of functional requirements, the requirements of safety specifications, etc. The constraint conditions vary and are intertwined in different application scenarios, making the layout problem very complex. Traditional layout methods are difficult to consider and meet all these constraint conditions simultaneously, and an intelligent space distribution method is needed to model and process these constraints to find the optimal or sub-optimal layout scheme that meets all constraints.

[0003] In the existing spatial layout, generally, the design party or IE engineer determines the requirements, the construction is carried out according to the established requirements, and on-site adjustments are made according to the actual situation during use. Secondary construction will cause a lot of waste of time and materials.

[0004] Because the existing design solutions need to cover various situations, generally, a downward compatibility mode is adopted, and the solution with the largest space and conditions is used as the standard solution for implementation, resulting in a lot of waste. In the case of high environmental requirements, there will be problems of different branch loads. Currently, all these are manually controlled by engineers according to the on-site usage conditions, frequently causing situations such as circuit overload tripping and gas circuit non-compliance, affecting quality control and delivery time. If the distribution law and hot spot area of spatial data are found through cluster analysis, this kind of problem can be well solved. Summary of the Invention

[0005] (1) Technical Problems to be Solved

[0006] Aiming at the deficiencies of the existing technology, the present invention provides an intelligent space distribution method based on layout constraints, which has the advantages of not requiring secondary construction, and can be quickly adjusted according to the actual situation on site. By finding the distribution law and hot spot area of spatial data through cluster analysis to solve existing problems, etc., and solves the problem that in the existing spatial layout, generally, the design party or IE engineer determines the requirements, the construction is carried out according to the established requirements, and on-site adjustments are made according to the actual situation during use, and secondary construction will cause a lot of waste of time and materials.

[0007] (2) Technical Solutions

[0008] To achieve the above object, the present invention provides the following technical solution: an intelligent space distribution method based on layout constraints, including the following steps:

[0009] 1) Planning and design process;

[0010] 2) Image processing and computer vision recognition process;

[0011] 3) Process of combining mathematical model analysis;

[0012] 4) Find the distribution law and hot spots of spatial data through cluster analysis;

[0013] 5) Intelligent spatial regression global distribution model;

[0014] Preferably, the planning and design process includes the following steps:

[0015] S1. Strengthen communication and collaboration. Before determining the spatial layout requirements, the design party or IE engineer communicates fully with relevant personnel such as the construction team and the user side to jointly discuss possible problems and requirements. For example, in a construction project, the design team communicates with the construction side in advance about the construction process and difficulties, and communicates with future users about actual usage requirements to avoid adjustments due to unclear requirements in the later stage.

[0016] S2. Conduct in-depth on-site research. Designers should conduct a detailed survey of the construction site to fully understand the actual situation of the space, including building structure, circuit routing, gas pipeline distribution, etc. For example, before office decoration, designers should measure the space size on the spot, understand the location of the original circuit interfaces and the routing of gas pipelines, etc., to provide a basis for accurate design.

[0017] S3. Establish a digital model. Use digital technology to establish a three-dimensional model of the space to visually present the spatial layout and the routing of various equipment and pipelines, and discover possible conflicts and problems in advance. For example, in hospital construction, through the BIM model, simulate the layout of medical equipment, gas pipelines, circuits, etc. in advance to avoid collisions and rework during construction.

[0018] S4. Develop a flexible design plan. Consider a certain degree of flexibility and scalability during design to adapt to possible future changes and requirements, rather than excessive downward compatibility. For example, movable partitions are used in office spaces to facilitate re-dividing the space according to personnel changes or business adjustments in the later stage.

[0019] Preferably, in the image processing and computer vision recognition process, for image acquisition and recognition, use devices such as cameras to collect spatial images, and use image processing technology to analyze and recognize the images to extract information such as objects, boundaries, and regions in the space. For example, determine the structure of urban space by recognizing the contours of buildings, the routing of roads, etc. In addition, for space segmentation and understanding, with the help of computer vision algorithms, segment and understand the space in the image, and divide it into different functional areas, such as distinguishing residential areas, commercial areas, industrial areas, etc. In an indoor environment, the position of furniture and the spatial layout can also be recognized through image processing to achieve intelligent indoor space management.

[0020] Preferably, in the process of combining mathematical model analysis, spatial statistical analysis methods are adopted, such as spatial autocorrelation analysis and hotspot analysis, to study the distribution characteristics and correlations of spatial data. By calculating statistical indicators such as Moran's Index, it is determined whether there is a tendency of aggregation or dispersion in the spatial data, and hotspot areas and cold spot areas in the space are identified;

[0021] The formula for calculating Moran's Index is as follows:

[0022]

[0023] where n is the number of samples, x i and x j are the observed values of the i-th and j-th sample points, is the average observed value of all sample points, and ω ij is the spatial connection weight between sample points i and j.

[0024] Preferably, finding the distribution law and hotspot areas of spatial data through cluster analysis is divided into four steps: data preparation, selection of clustering algorithm, clustering calculation, and identification and analysis of distribution law of hotspot areas.

[0025] Preferably, the data preparation includes the following steps:

[0026] S1. Data collection: Collect various data related to spatial positions, such as geographical coordinates, population density, commercial activity frequency, traffic flow, etc. These data can come from sensors, GPS devices, questionnaires, statistical reports, remote sensing images, etc.;

[0027] S2. Data cleaning: Check the integrity and accuracy of the data, remove duplicate, incorrect or incomplete data records, fill in missing values, and handle outliers that deviate significantly from other data according to the actual situation, such as deletion or reasonable correction;

[0028] S3. Data transformation: According to the characteristics of the data and the requirements of the clustering algorithm, transform the data, which may include converting categorical data into numerical data, and standardizing or normalizing numerical data to ensure comparability between different features.

[0029] Preferably, for the selected clustering algorithm, the K-means clustering algorithm is adopted, which is applicable to the situation where the data distribution is relatively uniform and the cluster shape is relatively regular. It is necessary to specify the number of clusters K in advance. Then, for the clustering calculation, first determine the parameters. According to the selected algorithm, determine the corresponding parameters. For example, for the K value in the K-means algorithm, appropriate parameter values can be determined through methods such as experience, multiple experiments, or cross-validation. Then, perform clustering. Input the prepared data into the selected clustering algorithm for clustering calculation. The algorithm will divide the data points into different clusters according to metrics such as the distance or similarity between data points. Finally, evaluate the clustering results. Use evaluation metrics such as the silhouette coefficient, Davies-Bouldin index, and within-cluster sum of squared deviations to evaluate the quality of the clustering results. The closer the silhouette coefficient is, the better the clustering effect. The smaller the Davies-Bouldin index, the better the clustering effect.

[0030] Preferably, the calculation formula of the K-means clustering algorithm is as follows:

[0031]

[0032] where x j is the j-th feature of the sample x, and μi j is the j-th feature of the cluster center μi.

[0033] Preferably, for identifying the hot spot area, for density-based clustering algorithms, clusters with higher density can usually be regarded as hot spot areas. For algorithms such as K-means, in combination with the attribute values of the actual data, determine which areas represented by clusters have higher activity or importance, and thus determine them as hot spot areas. It is also possible to calculate the central position, boundary range, etc. of each cluster to further clarify the location and scale of the hot spot area, analyze the distribution law, observe the distribution positions, shapes, sizes of different clusters, as well as the distances and mutual relationships between them, and summarize the distribution law of spatial data. For example, whether there are multiple obvious aggregation areas, whether these areas are concentrated within a specific geographical range, or show a scattered distribution pattern; whether there is a certain spatial correlation or gradient change between different types of clusters, etc.

[0034] Preferably, the intelligent spatial regression global distribution model adopts the spatial Durbin model SDM, and the relationship is as follows:

[0035]

[0036] β: The spatial regression coefficient of the independent variable, a 1×k-dimensional vector;

[0037] ξ: The random error vector;

[0038] p: The spatial interaction coefficient, 1×0;

[0039] λ: Spatial error term parameter 1 - 0;

[0040] I: Identity matrix;

[0041] W: Spatial weight matrix, showing the relationship between y and adjacent regions;

[0042] where y is an n×1 explained vector, x is an n×k exogenous explanatory variable design matrix, β is a k×1 regression coefficient vector, μ is an n×1 error vector, p is the spatial correlation coefficient, λ is the residual spatial correlation coefficient, ∈ is an n×1 random error vector, the elements of which are independent and identically distributed, with zero mean and variance σ 2 finite, and W1, W2 are n×n non - random spatial weight matrices;

[0043] By imposing different restrictions on the model parameters, different models can be derived:

[0044] S1. When p≠0, β = λ = 0, it is a pure spatial autoregressive model or a first - order spatial autoregressive model;

[0045] S2. When p≠0, β≠0, λ = 0, it is a spatial lag model;

[0046] S3. When p≠0, β≠0, λ≠0, it is a spatial error model.

[0047] Another technical problem to be solved by the present invention is to provide an intelligent spatial distribution method based on layout constraints, including the following steps:

[0048] 1) Planning and design process;

[0049] 2) Image processing and computer vision recognition process;

[0050] 3) Combining with the mathematical model analysis process;

[0051] 4) Finding out the distribution law and hot spots of spatial data through cluster analysis;

[0052] 5) Intelligent spatial regression global distribution model;

[0053] (III) Beneficial effects

[0054] Compared with the prior art, the present invention provides an intelligent spatial distribution method based on layout constraints, having the following beneficial effects:

[0055] 1. The intelligent spatial distribution method based on layout constraints finds the distribution law and hot spots of spatial data through clustering analysis, which is divided into four steps: data preparation, selection of clustering algorithm, clustering calculation, and identification and analysis of distribution law of hot spots. When selecting a clustering algorithm, the K-means clustering algorithm is adopted, which is applicable to the situation where the data distribution is relatively uniform and the cluster shape is relatively regular. The number of clusters K needs to be specified in advance. Then, for clustering calculation, first determine the parameters. According to the selected algorithm, determine the corresponding parameters. For example, for the K value in the K-means algorithm, appropriate parameter values can be determined through methods such as experience, multiple trials, or cross-validation. Then perform clustering. Input the prepared data into the selected clustering algorithm for clustering calculation. The algorithm will divide the data points into different clusters according to metrics such as the distance or similarity between data points. Finally, evaluate the clustering results. Use evaluation metrics such as silhouette coefficient, Davies-Bouldin index, and within-cluster sum of squared deviations to evaluate the quality of the clustering results. The closer the silhouette coefficient is, the better the clustering effect. The smaller the Davies-Bouldin index, the better the clustering effect.

[0056] 2. For the intelligent spatial distribution method based on layout constraints, by identifying hot spots, for density-based clustering algorithms, clusters with higher density can usually be regarded as hot spots. For algorithms such as K-means, the attribute values of the actual data can be combined to determine which clusters represent areas with higher activity or importance, and thus determine them as hot spots. The central position, boundary range, etc. of each cluster can also be calculated to further clarify the location and scale of the hot spots. Analyze the distribution law, observe the distribution positions, shapes, sizes of different clusters, as well as the distances and mutual relationships between them, and summarize the distribution law of spatial data. For example, whether there are multiple obvious aggregation areas, whether these areas are concentrated within a specific geographical range, or show a scattered distribution pattern, and whether there is a certain spatial association or gradient change between different types of clusters. Brief Description of the Drawings

[0057] Figure 1 It is a schematic diagram of the bridge layout. Detailed Implementation Manner

[0058] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0059] Embodiment 1: An intelligent spatial distribution method based on layout constraints includes the following steps:

[0060] 1) Planning and design process, the planning and design process includes the following steps:

[0061] S1. Strengthen communication and collaboration. Before determining the space layout requirements, the design party or IE engineer communicates fully with relevant personnel such as the construction team and the user side to jointly discuss possible problems and requirements. For example, in a construction project, the design team communicates with the construction side in advance about the construction technology and difficulty, and communicates with future users about the actual usage requirements to avoid adjustments due to unclear requirements in the later stage.

[0062] S2. Conduct in-depth on-site research. Designers should conduct a detailed survey of the construction site to fully understand the actual situation of the space, including building structure, circuit routing, gas pipeline distribution, etc. For example, before office decoration, designers should measure the space size on-site, understand the location of the original circuit interfaces and the routing of gas pipelines, etc., to provide a basis for precise design.

[0063] S3. Establish a digital model. Use digital technology to establish a three-dimensional model of the space, visually present the space layout and the routing of various equipment and pipelines, and discover possible conflicts and problems in advance. For example, in hospital construction, the layout of medical equipment, gas pipelines, circuits, etc. is simulated in advance through a BIM model to avoid collisions and rework during construction.

[0064] S4. Develop a flexible design plan. Consider a certain degree of flexibility and scalability during design to adapt to possible future changes and requirements, rather than being overly backward compatible. For example, movable partitions are used in office spaces to facilitate re-dividing the space according to personnel changes or business adjustments in the later stage.

[0065] 2) Image processing and computer vision recognition process, the image processing and computer vision recognition process, including image acquisition and recognition. Use devices such as cameras to collect space images, and use image processing technology to analyze and recognize the images to extract information such as objects, boundaries, and regions in the space. For example, by recognizing the contours of buildings and the routing of roads, etc., the structure of urban space is determined. In addition, space segmentation and understanding. With the help of computer vision algorithms, the space in the image is segmented and understood, and it is divided into different functional areas, such as distinguishing residential areas, commercial areas, industrial areas, etc. In an indoor environment, the position and space layout of furniture can also be recognized through image processing to achieve intelligent indoor space management.

[0066] 3) Combined mathematical model analysis process, the combined mathematical model analysis process, adopts spatial statistical analysis methods, uses spatial statistical methods such as spatial autocorrelation analysis and hotspot analysis to study the distribution characteristics and correlations of spatial data, and determines whether there is a tendency of aggregation or dispersion of spatial data by calculating statistical indicators such as Moran's index, and determines the hotspot areas and cold spot areas in the space;

[0067] The formula for calculating the Lan index is as follows:

[0068]

[0069] where n is the number of samples, x i and x j are the observed values of the i-th and j-th sample points, is the average observed value of all sample points, and ω ij is the spatial connection weight between sample points i and j.

[0070] 4) Finding the distribution law and hot spots of spatial data through cluster analysis. The process of finding the distribution law and hot spots of spatial data through cluster analysis is divided into four steps: data preparation, selection of clustering algorithm, clustering calculation, and identification and analysis of hot spots and distribution law. The data preparation includes the following steps:

[0071] S1. Data collection: Collect various data related to spatial locations, such as geographical coordinates, population density, commercial activity frequency, traffic flow, etc. These data can come from sensors, GPS devices, questionnaires, statistical reports, remote sensing images, etc.;

[0072] S2. Data cleaning: Check the integrity and accuracy of the data, remove duplicate, incorrect, or incomplete data records, fill in missing values, and handle outliers that deviate significantly from other data according to the actual situation, such as deletion or reasonable correction;

[0073] S3. Data transformation: According to the characteristics of the data and the requirements of the clustering algorithm, transform the data, which may include converting categorical data into numerical data and performing standardization or normalization on numerical data to ensure comparability between different features;

[0074] Among them, when selecting the clustering algorithm, the K-means clustering algorithm is adopted, which is suitable for situations where the data distribution is relatively uniform and the cluster shapes are relatively regular. The number of clusters K needs to be specified in advance. Then, for the clustering calculation, first determine the parameters. According to the selected algorithm, determine the corresponding parameters. For example, for the K value in the K-means algorithm, appropriate parameter values can be determined through methods such as experience, multiple trials, or cross-validation. Then perform the clustering. Input the prepared data into the selected clustering algorithm for clustering calculation. The algorithm will divide the data points into different clusters according to metrics such as the distance or similarity between data points. Finally, evaluate the clustering results. Use evaluation indicators such as the silhouette coefficient, Davies-Bouldin index, and within-cluster sum of squared deviations to evaluate the quality of the clustering results. The closer the silhouette coefficient is, the better the clustering effect. The smaller the Davies-Bouldin index, the better the clustering effect. The calculation formula of the K-means clustering algorithm:

[0075]

[0076] where x j is the j-th feature of sample x, and μi j is the j-th feature of the clustering center μi;

[0077] Among them, to identify hot spots, for density-based clustering algorithms, clusters with higher density can usually be regarded as hot spots. For algorithms such as K-means, the attribute values of the actual data can be combined to determine which clusters represent regions with higher activity or importance, and thus determined as hot spots. The central position, boundary range, etc. of each cluster can also be calculated to further clarify the location and scale of the hot spots, analyze the distribution law, observe the distribution positions, shapes, sizes of different clusters, as well as the distances and mutual relationships between them, and summarize the distribution law of spatial data. For example, whether there are multiple obvious aggregation regions, whether these regions are concentrated within a specific geographical range, or show a scattered distribution pattern; whether there is a certain spatial correlation or gradient change between different types of clusters, etc.

[0078] 5) Intelligent spatial regression global distribution model, the intelligent spatial regression global distribution model adopts the spatial Durbin model SDM, where the relationship is as follows:

[0079]

[0080] β: Spatial regression coefficient of independent variable, a 1×k vector;

[0081] ξ: Random error vector;

[0082] p: Spatial interaction coefficient, 1×0;

[0083] λ: Spatial error term parameter, 1×0;

[0084] I: Identity matrix;

[0085] W: Spatial weight matrix, showing the relationship between y and adjacent regions;

[0086] where y is an n×1 explained vector, x is an n×k exogenous explanatory variable design matrix, β is a k×1 regression coefficient vector, μ is an n×1 error vector, p is the spatial correlation coefficient, λ is the residual spatial correlation coefficient, ∈ is an n×1 random error vector, the elements of which are independent and identically distributed, and have zero mean and variance σ 2 finite, W1, W2 are n×n non-random spatial weight matrices;

[0087] By imposing different restrictions on the model parameters, different models can be derived:

[0088] S1. When p≠0, β = λ = 0, it is a pure spatial autoregressive model or a first-order spatial autoregressive model;

[0089] S2. When p≠0, β≠0, λ = 0, it is a spatial lag model;

[0090] S3. When p≠0, β≠0, λ≠0, it is a spatial error model.

[0091] As Figure 1 shown, through the existing design, the layout of the cable tray is generally in the form shown by the red line. In order to cover a variety of terminal combinations, generally, the standard of using device A uniformly is adopted for layout. In fact, multiple device Bs can be placed, but because the standard of A is used, the number of terminals is placed less. Secondly, the cable tray shown by the red line will be configured with water, electricity and gas interfaces to meet the needs of A, but during the use process, it will be switched to the needs of BCDE, etc. Therefore, during the automatic configuration process, suggestions will be made based on the historical situation of the installed devices, giving priority to ensuring that devices with the same water, electricity and gas requirements are in the same position, followed by ensuring the adaptation of terminal combinations with relatively small changes in water, electricity and gas, and then, in order to facilitate the management of the same type of devices, they will be placed together as much as possible. For example, in terms of electricity, the load capacity and source of each cable tray are not necessarily the same, so the general distribution will result in situations such as power outages of the devices corresponding to the entire cable tray due to insufficient load, causing many losses. Therefore, a guiding configuration is needed, and it can be manually allocated, but this method is used for inspection before implementation to reduce the occurrence of many risk problems and rework.

[0092] The beneficial effects of the present invention are as follows: First, the distribution law and hot spots of spatial data are found through clustering analysis, which are divided into four steps: data preparation, selection of clustering algorithm, clustering calculation, and identification and distribution law analysis of hot spots. When selecting a clustering algorithm, the K-means clustering algorithm is adopted, which is suitable for the case where the data distribution is relatively uniform and the cluster shape is relatively regular. The number of clusters K needs to be specified in advance. Then, for clustering calculation, first determine the parameters. According to the selected algorithm, determine the corresponding parameters. For example, for the K value in the K-means algorithm, appropriate parameter values can be determined by methods such as experience, multiple experiments, or cross-validation. Then perform clustering. Input the prepared data into the selected clustering algorithm for clustering calculation. The algorithm will divide the data points into different clusters according to metrics such as the distance or similarity between data points. Finally, evaluate the clustering results. Use evaluation indicators such as the silhouette coefficient, Davies-Bouldin index, and within-cluster sum of squared deviations to evaluate the quality of the clustering results. The closer the silhouette coefficient is, the better the clustering effect. The smaller the Davies-Bouldin index, the better the clustering effect. Then, by identifying hot spots, for density-based clustering algorithms, clusters with higher density can usually be regarded as hot spots. For algorithms such as K-means, the attribute values of the actual data can be combined to determine which clusters represent regions with higher activity or importance, and thus determine them as hot spots. The central position, boundary range, etc. of each cluster can also be calculated to further clarify the location and scale of the hot spots. Analyze the distribution law, observe the distribution positions, shapes, sizes of different clusters, as well as the distances and mutual relationships between them, and summarize the distribution law of spatial data. For example, whether there are multiple obvious aggregation regions, whether these regions are concentrated within a specific geographical range, or show a scattered distribution pattern, and whether there is a certain spatial correlation or gradient change between different types of clusters, etc.

[0093] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principle and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent spatial distribution method based on layout constraints, characterized in that: The following steps are involved: 1) Planning and design process; 2) Image processing and computer vision recognition process; 3) Combine the mathematical model analysis process; 4) Find out the distribution patterns and hot spots of spatial data through cluster analysis; 5) Intelligent spatial regression global distribution model.

2. The intelligent space distribution method based on layout constraints according to claim 1, characterized in that: The planning and design process includes the following steps: S1. Strengthen communication and collaboration. Before determining the space layout requirements, the designer or IE engineer should fully communicate with the construction team, users and other relevant personnel to jointly discuss possible problems and requirements. For example, in a construction project, the design team should communicate with the construction party in advance about the construction process and difficulty, and communicate with future users about actual usage requirements to avoid adjustments later due to unclear requirements. S2. Conduct in-depth on-site investigation. Designers should conduct detailed surveys of the construction site to fully understand the actual situation of the space, including the building structure, circuit direction, gas distribution, etc. For example, before decorating an office, designers should measure the space dimensions on site to understand the original circuit interface locations, gas pipeline directions, etc., to provide a basis for accurate design. S3. Establish a digital model. Use digital technology to establish a three-dimensional model of the space to intuitively present the spatial layout and the directions of various equipment and pipelines, and discover possible conflicts and problems in advance. For example, in hospital construction, the BIM model can be used to simulate the layout of medical equipment, gas circuits, circuits, etc. in advance to avoid collisions and rework during construction. S4. Develop a flexible design plan, taking into account a certain degree of flexibility and scalability when designing to adapt to possible changes and needs in the future, rather than being overly backward compatible. For example, movable partitions can be used in office spaces to facilitate the redivision of space based on personnel changes or business adjustments in the future.

3. The intelligent space distribution method based on layout constraints according to claim 1, characterized in that: The image processing and computer vision recognition process, including image acquisition and recognition, uses cameras and other equipment to collect spatial images, uses image processing technology to analyze and recognize images, and extracts information such as objects, boundaries, and regions in the space. For example, by identifying the outline of buildings, the direction of roads, etc., the structure of urban space can be determined. In addition, spatial segmentation and understanding, with the help of computer vision algorithms, the space in the image is segmented and understood, and divided into different functional areas, such as distinguishing residential areas, commercial areas, industrial areas, etc. In indoor environments, image processing can also be used to identify the location and spatial layout of furniture to achieve intelligent indoor space management.

4. The intelligent space distribution method based on layout constraints according to claim 1, characterized in that: The above-mentioned combined mathematical model analysis process adopts spatial statistical analysis methods, uses spatial autocorrelation analysis, hot spot analysis and other spatial statistical methods, studies the distribution characteristics and correlation of spatial data, and determines whether there is a trend of aggregation or dispersion in spatial data by calculating statistical indicators such as Moran's index, and determines the hot spot area and cold spot area in space; The formula for calculating the Lan Index is as follows: Where n is the number of samples, x i and x j are the observed values ​​of the i-th and j-th sample points, is the average observation value of all sample points, ω ij is the spatial connection weight between sample points i and j.

5. The intelligent space distribution method based on layout constraints according to claim 1, characterized in that: The method of finding the distribution law and hot spots of spatial data through cluster analysis is divided into four steps: data preparation, selection of clustering algorithm, clustering calculation, hot spot identification and distribution law analysis.

6. The intelligent space distribution method based on layout constraints according to claim 5, characterized in that: The data preparation includes the following steps: S1. Data collection: collect various data related to spatial location, such as geographic coordinates, population density, commercial activity frequency, traffic flow, etc. These data can come from sensors, GPS equipment, questionnaires, statistical reports, remote sensing images, etc. S2. Data cleaning: Check the completeness and accuracy of data, remove duplicate, erroneous or incomplete data records, fill in missing values, and handle outliers that are obviously deviated from other data according to actual conditions, such as deleting or making reasonable corrections; S3. Data transformation: transform the data according to the data characteristics and the requirements of the clustering algorithm, which may include converting categorical data into numerical data and standardizing or normalizing the numerical data to ensure comparability between different features.

7. The intelligent space distribution method based on layout constraints according to claim 5, characterized in that: The clustering algorithm selected adopts the K-means clustering algorithm, which is suitable for the situation where the data distribution is relatively uniform and the cluster shape is relatively regular. The number of clusters K needs to be specified in advance. Then the clustering calculation first determines the parameters. According to the selected algorithm, the corresponding parameters are determined, such as the K value in the K-means algorithm. The appropriate parameter value can be determined through experience, multiple experiments or cross-validation. Then clustering is performed, and the prepared data is input into the selected clustering algorithm for clustering calculation. The algorithm will divide the data points into different clusters according to metrics such as the distance or similarity between the data points. Finally, the clustering results are evaluated. The quality of the clustering results is evaluated using evaluation indicators such as the silhouette coefficient, Davies-Bouldin index, and the sum of squares of intra-cluster deviations. The closer the silhouette coefficient is, the better the clustering effect is. The smaller the Davies-Bouldin index is, the better the clustering effect is.

8. The intelligent space distribution method based on layout constraints according to claim 7, characterized in that: The calculation formula of the K-means clustering algorithm is: where x j is the jth feature of sample x, μi j is the j-th feature of cluster center μi.

9. The intelligent space distribution method based on layout constraints according to claim 5, characterized in that: For the identification of hotspot areas, for density-based clustering algorithms, clusters with higher density can usually be regarded as hotspot areas. For algorithms such as K-means, the attribute values ​​of the actual data can be combined to determine which clusters represent areas with higher activity or importance, thereby determining them as hotspot areas. The center position, boundary range, etc. of each cluster can also be calculated to further clarify the position and scale of the hotspot area, analyze the distribution pattern, observe the distribution position, shape, size of different clusters and the distance and relationship between them, and summarize the distribution pattern of spatial data, for example, whether there are multiple obvious clustering areas, whether these areas are concentrated in a specific geographical range, or present a dispersed distribution pattern; whether there is a certain spatial correlation or gradient change between different types of clusters, etc.

10. The intelligent space distribution method based on layout constraints according to claim 1, characterized in that: The intelligent spatial regression global distribution model adopts the spatial Durbin model SDM, where the relationship is as follows: β: spatial regression coefficient of the independent variable 1xk dimensional vector; ξ: random error vector; p: spatial mutual coefficient 1-0; λ: spatial error term parameter 1-0; I: identity matrix; W: spatial weight matrix, showing the relationship between y and neighboring areas; Where y is the n×1 explained vector, x is the n×k exogenous explanatory variable design matrix, β is the k×1 regression coefficient vector, μ is the n×1 error vector, p is the spatial correlation coefficient, λ is the residual spatial correlation coefficient, ∈ is the n×1 random error vector, whose elements are independent and identically distributed, with zero mean and variance σ 2 Finite, W1, W2 are n×n non-random spatial weight matrices; By placing different restrictions on the model parameters, different models can be derived: S1, when p≠0, β=λ=0, it is a pure spatial autoregressive model or a -th order spatial autoregressive model; S2, when p≠0, β≠0, λ=0, it is the spatial lag model; S3. When p≠0, β≠0, λ≠0, it is a spatial error model.

Images