Intelligent building management system

Abstract

The invention provides a building intelligent management system comprising a modeling data acquisition module used for acquiring building basic data through freehand sketch input or multi-sensor fusion scanning; the intelligent processing module is used for sequentially executing sketch regularization, feature extraction and three-dimensional modeling functions; the decoration generation module is used for automatically generating a design scheme based on deep learning and realizing visual display of virtual-real fusion; the data management module is used for carrying out classified management and authority control on the whole-process data of the building; efficient management of the whole life cycle of the building is realized through multi-dimensional technical innovation; a laser scanning technology and a high-resolution imaging technology are combined, the industry problem that geometric accuracy and texture details are difficult to consider in the prior art is effectively solved through an adaptive fusion algorithm, data characteristics of different sensors can be intelligently balanced, integrity and accuracy of building information collection are ensured, and the construction quality is improved. And the modeling efficiency and quality are greatly improved.

Description

Technical Field

[0001] The present invention relates to a building intelligent management system, in particular to the technical field of building intelligence. Background Art

[0002] With the digital transformation and intelligent development of the construction industry, the traditional building design, construction and operation and maintenance management models are difficult to meet the needs of modern building complexity; the current building information management mainly faces the following technical bottlenecks:

[0003] In terms of data acquisition: The traditional surveying method relies on equipment such as total stations, which has problems such as low efficiency and single data (only obtaining geometric information). Although laser scanning technology has been widely used, a single sensor cannot meet the dual requirements of geometric accuracy and texture details at the same time.

[0004] In terms of design processing: Existing BIM software relies on manual modeling. Designers need to spend a lot of time building the basic model. During the digitalization process of hand-drawn sketches, the line recognition accuracy is low, and there is a lack of automatic verification ability for building codes.

[0005] In terms of decorative design: The traditional rendering tool has a long scheme generation cycle, and there is a significant deviation between the rendered image and the actual space. The application of augmented reality technology (AR) in the building scene is still limited by the real-time registration accuracy and dynamic light adaptation problems.

[0006] In terms of data management: The heterogeneous data generated during the whole life cycle of the building lacks a unified indexing system, and the permission management is extensive, and it is impossible to achieve fine control at the component level (such as a single beam or column). Summary of the Invention

[0007] Object of the Invention: To propose a building intelligent management system to solve the above problems existing in the prior art.

[0008] Technical Solution: A building intelligent management system includes:

[0009] A modeling data acquisition module for obtaining building basic data through hand-drawn sketch input or multi-sensor fusion scanning;

[0010] An intelligent processing module that sequentially executes functions of sketch regularization, feature extraction and 3D modeling;

[0011] A decoration generation module that automatically generates a design scheme based on deep learning and realizes a visual display of virtual-real fusion;

[0012] A data management module for classifying and managing the whole process data of the building and controlling permissions.

[0013] In a further embodiment, the modeling data acquisition module includes a hand-drawn input unit. Drawing a sketch gives designers an independent creative space and provides a way to obtain basic information for building modeling.

[0014] A three-dimensional scanning unit that scans the three dimensions of a building to obtain the structural data of the actual building. Scanning the three dimensions of a building is achieved by a LiDAR and an RGB camera for rapid building structure modeling. The LiDAR can obtain the three-dimensional spatial information of the building, and the RGB camera provides the texture and color information of the building. The two are combined to complete the building structure modeling.

[0015] A data fusion unit that uses an adaptive algorithm to compensate for the measurement differences of different sensors. The scanning accuracy compensation algorithm compensates for sensor errors and multi-source data in the scenario of LiDAR and RGB camera fusion modeling. The expression is:

[0016]

[0017] In the formula, Pout is the compensated three-dimensional coordinate; w is the LiDAR data weight coefficient; Plidar is the original point cloud coordinate of the LiDAR; Prgb is the back-projected coordinate of the RGB camera; λ is the geometric smoothing coefficient; is the spatial Laplacian operator; the weight calculation formula:

[0018] w = σrgb / (σrgb + σlidar)

[0019] In the formula, w is the LiDAR data weight coefficient; σrgb is the noise variance of the RGB camera; σlidar is the noise variance of the LiDAR. According to the real-time noise levels of the LiDAR and the RGB camera, the contribution weights of the two are dynamically adjusted.

[0020] When the LiDAR noise is low and approaches 1, LiDAR data is used;

[0021] When the RGB noise is low and approaches 0, the weight of the RGB data is enhanced.

[0022] In a further embodiment, the intelligent processing module includes a sketch regularization unit. The specific steps are as follows:

[0023] Step 211, stroke classification processing: Perform stroke classification through a neural network algorithm to accurately distinguish basic stroke types such as straight lines and curves;

[0024] Step 212, geometric figure conversion: For straight strokes, use the Gestalt principle to perceive intersection points to obtain straight line segments; for curved strokes, use a five-point sampling method including endpoints to determine ellipse parameters;

[0025] Step 213, topological relationship reconstruction: Analyze the connection relationships between geometric elements and reconstruct the spatial topological structure of the sketch.

[0026] Step 214, Automatic dimension calibration: Automatically calculate and mark the actual dimensions of each component according to a preset ratio;

[0027] The feature extraction unit, the specific steps are as follows:

[0028] Step 221, Component recognition: Extract the building component features from the regularized graphics; Automatically recognize the basic building components such as walls, doors, windows, beams and columns; Retrieve the feature points and related line segments in the sketch; Based on the feature points, determine the component contour through neighborhood analysis;

[0029] Step 222, Spatial analysis: Obtain the stretching plane information and stretching length according to the extracted graphic features; Analyze the spatial position relationship of each component and establish the spatial division of floors and rooms;

[0030] Step 223, Attribute annotation: Deduce the corresponding entity coordinate system and local three-dimensional model, and add material and dimension attribute information to the recognized components;

[0031] Step 224, Specification check: Automatically check the compliance of the design scheme against building codes;

[0032] The 3D modeling unit, the specific steps are as follows:

[0033] Step 231, Basic model generation: Construct a complete three-dimensional space model under the entity coordinate system and local three-dimensional model;

[0034] Step 232, Detail optimization: Add necessary building details and construction joints;

[0035] Step 233, Model verification: Check the integrity and consistency of the model;

[0036] Step 234, Format output: Generate a BIM model file that meets industry standards.

[0037] In a further embodiment, the decoration generation module includes: a scheme generation unit, which automatically generates a decoration scheme based on building features, and the specific steps are as follows:

[0038] Step 311, Feature analysis stage: Receive the building three-dimensional model data from the intelligent processing module, and extract the key parameters such as spatial dimensions, structural features, and lighting conditions;

[0039] Step 312, Style matching stage: According to the user's preset design style preference, screen and match the decorative elements from the material library;

[0040] Step 313, Scheme generation stage: Adopt a deep learning algorithm to automatically generate multiple feasible decoration schemes;

[0041] Step 314, Scheme Optimization Phase: Receive user feedback through the human-computer interaction interface and iteratively optimize the scheme;

[0042] The virtual-real fusion unit superimposes the design scheme onto the actual space for display, and the specific steps are as follows:

[0043] Step 321, Space Registration Phase: Establish an accurate correspondence between the virtual decoration scheme and the actual building space through visual positioning technology;

[0044] Step 322, Real-time Rendering Phase: Dynamically adjust the display attributes of the decoration effect according to the current perspective and lighting conditions;

[0045] Step 323, Interactive Display Phase: Support users to interact with the virtual decoration through touch, gestures or voice;

[0046] Step 324, Effect Evaluation Phase: Provide effect simulations from multiple angles and at multiple time periods to assist in decision-making.

[0047] In a further embodiment, the data management module includes:

[0048] Step 41, Data Organization Unit, which is used to intelligently classify and store the whole-process building data, and the specific steps are as follows:

[0049] Step 42, Spatial Dimension Organization: Based on the BIM coordinate system, spatially index the collected building data by floor, room, and component, and establish the spatial topological relationship of the building digital twin;

[0050] Step 43, Temporal Dimension Organization: Adopt the time axis of the whole building life cycle to record the data versions of each stage from design, construction to operation and maintenance,

[0051] Step 44, Data Association: Associate heterogeneous data such as 3D models, decoration schemes, and equipment information through unique identifiers to form a complete building information network;

[0052] The permission control unit is used to implement refined permission management of building data, and the specific steps are as follows:

[0053] Step 45, Role Definition: Set multiple levels of roles such as designers, constructors, owners, and operation and maintenance personnel according to the responsibilities of the participants in the building project;

[0054] Step 46, Spatial Permission: Combine the spatial dimension of data organization to achieve refined access control by building area;

[0055] Step 47, Temporal Permission: Control data access permissions based on project stages, such as restricting the modification permission of design drawings during the construction stage;

[0056] Step 48, Operation Audit: Record the access logs of all users and associate them with the time versions of the building data to achieve complete operation traceability.

[0057] In a further embodiment, the data organization unit and the permission control unit work together, where: the spatial classification of data organization provides a regional management basis for permission control; the time version record provides a time benchmark for permission audit; through a unified data identification system, seamless connection from data storage to access control is achieved.

[0058] In a further embodiment, the data flow of each module is as follows: the raw data collected by the modeling data acquisition module is transmitted to the intelligent processing module;

[0059] the three-dimensional model generated by the intelligent processing module is transmitted to the decoration generation module;

[0060] the design scheme generated by the decoration generation module is transmitted to the data management module;

[0061] the data stored in the data management module is available for each module to call;

[0062] A closed-loop building information processing flow is formed.

[0063] Beneficial Effects: The present invention proposes a building intelligent management system, which realizes the efficient management of the entire life cycle of a building through multi-dimensional technological innovation; combines laser scanning technology with high-resolution imaging technology, and through an adaptive fusion algorithm, effectively solves the industry problem in traditional technologies that it is difficult to balance geometric accuracy and texture details, can intelligently balance the data characteristics of different sensors, ensure the integrity and accuracy of building information collection, and greatly improve the modeling efficiency and quality; adopts a neural network architecture to achieve accurate recognition and automatic conversion of design sketches, and the built-in building code knowledge base can detect the compliance of the design scheme in real time, significantly reducing the workload of manual review;

[0064] By integrating real-time positioning and pre-computed lighting technology, high-precision matching between the design scheme and the actual space is achieved; the innovative interactive design platform allows users to intuitively adjust and evaluate the decoration effect, transforming the traditional design process into a dynamic creative process. Description of the Drawings

[0065] Figure 1 It is the flow chart of the modeling data acquisition of the present invention.

[0066] Figure 2 It is the flow chart of the intelligent processing of the present invention.

[0067] Figure 3 It is the flow chart of the decoration generation and virtual-real fusion of the present invention.

[0068] Figure 4This is the data management closed-loop flowchart of the present invention. Detailed implementation manners

[0069] To solve the problems existing in the prior art, the present invention provides a building intelligent management system, which solves the systematic problems of long-term low efficiency, error accumulation, and difficult collaboration in the construction industry, and provides a new technical paradigm for intelligent construction.

[0070] The following further specifically describes the solution through embodiments and with reference to the accompanying drawings.

[0071] In the present application, a building intelligent management system is proposed, including:

[0072] A modeling data acquisition module, configured to acquire building basic data through hand-drawn sketch input or multi-sensor fusion scanning; the modeling data acquisition module includes a hand-drawn input unit, and drawing a sketch gives designers an independent creative space to provide a basic information acquisition method for building modeling.

[0073] A three-dimensional scanning unit, scanning the three dimensions of the building to obtain the structural data of the actual building; scanning the three dimensions of the building is realized by LiDAR (light detection and ranging) and an RGB camera to quickly model the building structure. LiDAR can obtain the three-dimensional space information of the building, and the RGB camera provides the texture and color information of the building. The two are combined to complete the building structure modeling.

[0074] A data fusion unit, adopting an adaptive algorithm to compensate for the measurement differences of different sensors. The scanning accuracy compensation algorithm compensates for sensor errors and multi-source data in the scenario of LiDAR and RGB camera fusion modeling. The expression is:

[0075]

[0076] In the formula, Pout is the compensated three-dimensional coordinate; w is the LiDAR data weight coefficient; Plidar is the LiDAR original point cloud coordinate; Prgb is the RGB camera back-projection coordinate; λ is the geometric smoothing coefficient; is the spatial Laplacian operator; the weight calculation formula:

[0077] w = σrgb / (σrgb + σlidar)

[0078] In the formula, w is the LiDAR data weight coefficient; σrgb is the RGB camera noise variance; σlidar is the LiDAR noise variance; according to the real-time noise levels (σrgb, σlidar) of LiDAR and the RGB camera, the contribution weights of the two are dynamically adjusted;

[0079] When the LiDAR noise (σlidar) is low and approaches 1, LiDAR data (high-precision geometry) is adopted;

[0080] When the RGB noise is low (σrgb) approaching 0, enhance the weight of RGB data (rich color information).

[0081] The intelligent processing module sequentially performs functions of sketch regularization, feature extraction, and 3D modeling; the intelligent processing module includes a sketch regularization unit, and the specific steps are as follows:

[0082] Step 211, stroke classification processing: Perform stroke classification through a neural network algorithm to accurately distinguish basic stroke types of straight lines and curves;

[0083] Step 212, geometric figure conversion: For straight strokes, use the Gestalt principle to perceive intersection points to obtain straight line segments; for curved strokes, use the five-point sampling method including endpoints to determine ellipse parameters;

[0084] Step 213, topological relationship reconstruction: Analyze the connection relationships between geometric elements and reconstruct the spatial topological structure of the sketch;

[0085] Step 214, automatic dimension calibration: Automatically calculate and mark the actual dimensions of each component according to a preset ratio;

[0086] The feature extraction unit, the specific steps are as follows:

[0087] Step 221, component recognition: Extract the characteristics of building components from the regularized graphics; automatically recognize basic building components such as walls, doors, windows, beams, and columns; retrieve the feature points and related line segments in the sketch; based on the feature points, determine the component contour through neighborhood analysis;

[0088] Step 222, spatial analysis: Obtain the stretching plane information and stretching length according to the extracted graphic features; analyze the spatial position relationships of each component and establish floor and room space divisions;

[0089] Step 223, attribute annotation: Deduce the corresponding entity coordinate system and local 3D model, and add material and dimension attribute information to the recognized components;

[0090] Step 224, specification check: Automatically check the compliance of the design scheme against building codes;

[0091] The 3D modeling unit, the specific steps are as follows:

[0092] Step 231, basic model generation: Construct a complete 3D space model under the entity coordinate system and local 3D model;

[0093] Step 232, detail optimization: Add necessary building details and construction joints;

[0094] Step 233, model verification: Check the integrity and consistency of the model;

[0095] Step 234, Format Output: Generate BIM model files that comply with industry standards.

[0096] The decoration generation module automatically generates design solutions based on deep learning and realizes the visual display of virtual-real integration; the decoration generation module includes: a solution generation unit that automatically generates decoration solutions based on building features, and the specific steps are as follows:

[0097] Step 311, Feature Analysis Phase: Receive the building three-dimensional model data from the intelligent processing module, and extract key parameters such as spatial dimensions, structural features, and lighting conditions;

[0098] Step 312, Style Matching Phase: According to the user's preset design style preferences, screen and match decorative elements from the material library;

[0099] Step 313, Solution Generation Phase: Adopt deep learning algorithms, comprehensively consider functionality, aesthetics, and specification requirements, and automatically generate multiple feasible decoration solutions;

[0100] Step 314, Solution Optimization Phase: Receive user feedback through the human-computer interaction interface and iteratively optimize the solution;

[0101] The virtual-real integration unit superimposes the design solution onto the actual space for display, and the specific steps are as follows:

[0102] Step 321, Space Registration Phase: Through visual positioning technology, establish an accurate correspondence between the virtual decoration solution and the actual building space;

[0103] Step 322, Real-time Rendering Phase: Dynamically adjust the display attributes of the decoration effect according to the current perspective and lighting conditions;

[0104] Step 323, Interactive Display Phase: Support users to interact with the virtual decoration through touch, gestures, or voice;

[0105] Step 324, Effect Evaluation Phase: Provide multi-angle and multi-period effect simulations to assist in decision-making.

[0106] The data management module classifies and manages the whole-process building data and controls permissions; the data management module includes:

[0107] Step 41, Data Organization Unit, which is used to intelligently classify and store the whole-process building data, and the specific steps are as follows:

[0108] Step 42, Spatial Dimension Organization: Based on the BIM coordinate system, spatially index the collected building data by floor, room, and component, and establish the spatial topological relationship of the building digital twin;

[0109] Step 43, Time Dimension Organization: Adopt the time axis of the whole life cycle of the building to record the data versions of each stage from design, construction to operation and maintenance, and support quickly tracing back the historical status according to time nodes;

[0110] Step 44, Data Association: Associate heterogeneous data such as 3D models, decoration schemes, and equipment information through unique identifiers to form a complete building information network;

[0111] The permission control unit is used to implement refined permission management of building data. The specific steps are as follows:

[0112] Step 45, Role Definition: Set multiple levels of roles such as designers, constructors, owners, and operation and maintenance personnel according to the responsibilities of the participants in the building project;

[0113] Step 46, Space Permission: Combine the space dimension of data organization to implement refined access control according to building areas (such as floors, rooms);

[0114] Step 47, Time Permission: Control data access permissions based on project stages. For example, restrict the modification permission of design drawings during the construction stage;

[0115] Step 48, Operation Audit: Record the access logs of all users and associate them with the time versions of building data to achieve complete operation traceability.

[0116] The data organization unit and the permission control unit work together. Among them: the spatial classification of data organization provides a regional management basis for permission control; the time version record provides a time benchmark for permission audit; through a unified data identification system, seamless connection from data storage to access control is achieved.

[0117] The data flow of each module is as follows: The original data collected by the modeling data acquisition module is transmitted to the intelligent processing module;

[0118] The 3D model generated by the intelligent processing module is transmitted to the decoration generation module;

[0119] The design scheme generated by the decoration generation module is transmitted to the data management module;

[0120] The data stored in the data management module can be called by each module;

[0121] A closed-loop building information processing flow is formed.

[0122] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as a limitation of the present invention itself. Various changes can be made in its form and details without departing from the spirit and scope of the present invention defined by the appended claims.

Claims

1. An intelligent building management system, characterized in that, Including: A modeling data acquisition module, which is used to obtain building basic data through hand-drawn sketch input or multi-sensor fusion scanning; An intelligent processing module, which sequentially executes functions of sketch regularization, feature extraction, and 3D modeling; A decoration generation module, which automatically generates a design scheme based on deep learning and realizes a visual display of virtual-real fusion; A data management module, which classifies and manages the whole-process data of the building and controls permissions.

2. The building intelligent management system according to claim 1, characterized in that The modeling data acquisition module includes a hand-drawn input unit. Drawing a sketch gives designers an independent creative space and provides a way to obtain basic information for building modeling. A 3D scanning unit, which scans the building in 3D to obtain the structural data of the actual building; Scanning the building in 3D is realized by LiDAR and an RGB camera to quickly model the building structure. LiDAR can obtain the 3D spatial information of the building, and the RGB camera provides the texture and color information of the building. The two are combined to complete the building structure modeling. A data fusion unit, which uses an adaptive algorithm to compensate for the measurement differences of different sensors. The scanning accuracy compensation algorithm compensates for sensor errors and multi-source data in the scenario of LiDAR and RGB camera fusion modeling. The expression is: Wherein, Pout is the compensated three-dimensional coordinate; s is the LiDAR data weight coefficient; Plidar is the LiDAR original point cloud coordinate; Prgb is the RGB camera back-projection coordinate; λ is the geometric smoothing coefficient; is the spatial Laplacian operator; the weight calculation formula: w = σrgb / (σrgb + σlidar) In the formula, w is the weight coefficient of LiDAR data; σrgb is the noise variance of the RGB camera; σlidar is the noise variance of LiDAR; according to the real-time noise levels of LiDAR and the RGB camera, the contribution weights of the two are dynamically adjusted; When the LiDAR noise is low and approaches 1, LiDAR data is adopted; When the RGB noise is low and approaches 0, the weight of RGB data is enhanced.

3. The intelligent building management system according to claim 1, characterized in that, The intelligent processing module includes a sketch regularization unit. The specific steps are as follows: Step 211, stroke classification processing: Classify strokes through a neural network algorithm to accurately distinguish basic stroke types of straight lines and curves; Step 212, geometric figure conversion: For straight strokes, use the Gestalt principle to perceive intersection points to obtain straight line segments; for curved strokes, use a five-point sampling method including endpoints to determine ellipse parameters; Step 213, topological relationship reconstruction: Analyze the connection relationships between geometric elements and reconstruct the spatial topological structure of the sketch; Step 214, automatic dimension calibration: Automatically calculate and mark the actual dimensions of each component according to a preset ratio; A feature extraction unit. The specific steps are as follows: Step 221, component recognition: Extract building component features from the regularized graphics; automatically recognize basic building components such as walls, doors, windows, beams, and columns; retrieve feature points and related line segments in the sketch; based on the feature points, determine the component contour through neighborhood analysis; Step 222, spatial analysis: Obtain the stretching plane information and stretching length according to the extracted graphic features; analyze the spatial position relationships of each component and establish floor and room space divisions; Step 223, attribute annotation: Deduce the corresponding entity coordinate system and local 3D model, and add material and dimension attribute information to the recognized components; Step 224, specification check: Automatically check the compliance of the design scheme against building codes; A 3D modeling unit. The specific steps are as follows: Step 231, basic model generation: Construct a complete 3D space model under the entity coordinate system and local 3D model; Step 232, Detail Optimization: Add necessary building details and construction joints; Step 233, Model Verification: Check the integrity and consistency of the model; Step 234, Format Output: Generate BIM model files that comply with industry standards.

4. A building intelligent management system according to claim 1, characterized in that, The decoration generation module includes: a scheme generation unit that automatically generates decoration schemes based on building features. The specific steps are as follows: Step 311, Feature Analysis Phase: Receive the 3D building model data from the intelligent processing module and extract key parameters such as spatial dimensions, structural features, and lighting conditions; Step 312, Style Matching Phase: Screen and match decorative elements from the material library according to the user's preset design style preferences; Step 313, Scheme Generation Phase: Use deep learning algorithms to automatically generate multiple feasible decoration schemes; Step 314, Scheme Optimization Phase: Receive user feedback through the human-computer interaction interface and iteratively optimize the scheme; A virtual-real fusion unit that superimposes the design scheme onto the actual space for display. The specific steps are as follows: Step 321, Space Registration Phase: Establish an accurate correspondence between the virtual decoration scheme and the actual building space through visual positioning technology; Step 322, Real-time Rendering Phase: Dynamically adjust the display attributes of the decoration effect according to the current perspective and lighting conditions; Step 323, Interactive Display Phase: Support users to interact with the virtual decoration through touch, gestures, or voice; Step 324, Effect Evaluation Phase: Provide multi-angle and multi-time period effect simulations to assist in decision-making.

5. An intelligent building management system according to claim 1, characterized in that The data management module includes: Step 41, Data Organization Unit, which is used to intelligently classify and store the whole-process building data. The specific steps are as follows: Step 42, Spatial Dimension Organization: Based on the BIM coordinate system, spatially index the collected building data by floor, room, and component, and establish the spatial topological relationship of the building digital twin; Step 43, Temporal Dimension Organization: Adopt the time axis of the whole building life cycle to record the data versions of each stage from design, construction to operation and maintenance; Step 44, Data Association: Associate heterogeneous data such as 3D models, decoration schemes, and equipment information through unique identifiers to form a complete building information network; A permission control unit, which is used to implement refined permission management of building data. The specific steps are as follows: Step 45, Role Definition: Set multiple levels of roles such as designers, constructors, owners, and operation and maintenance personnel according to the responsibilities of the participants in the building project; Step 46, Spatial Permission: Combine the spatial dimension of data organization to achieve refined access control by building area; Step 47, Temporal Permission: Control data access permissions based on project stages. For example, restrict the modification permission of design drawings during the construction stage; Step 48, Operation Audit: Record the access logs of all users and associate them with the time versions of building data to achieve complete operation traceability.

6. An intelligent building management system according to claim 5, characterized in that, The data organization unit and the permission control unit work together. Among them: the spatial classification of data organization provides the basis for area management in permission control; the time version record provides the time benchmark for permission auditing; through a unified data identification system, seamless connection from data storage to access control is achieved.

7. An intelligent building management system according to claim 1, characterized in that, The data flow of each module is as follows: the original data collected by the modeling data acquisition module is transmitted to the intelligent processing module; The 3D model generated by the intelligent processing module is transmitted to the decoration generation module; The design scheme generated by the decoration generation module is transmitted to the data management module; The data stored in the data management module can be called by each module; A closed-loop building information processing flow is formed.

Images