Artificial intelligence-based building model identification method and device, computer device

Abstract

This application relates to an artificial intelligence-based method, apparatus, computer device, and storage medium for building model recognition. The method includes: acquiring images of building components corresponding to a building model; wherein the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute; inputting the building component images into a preset recognition network to determine the attribute relationships between the building components in the building model, and obtaining the recognition results corresponding to the attribute relationships between the building components. This method can improve the quality of building models.

Description

Technical Field

[0001] This application relates to the field of building-aided design technology, and in particular to a building model recognition method, device, computer equipment, and storage medium based on artificial intelligence. Background Technology

[0002] With the development of science and technology, the field of architectural design has become increasingly automated.

[0003] Traditional architectural models are designed manually. After the design is completed, it is adjusted and optimized by people based on years of industry experience. For example, some regional features are added or removed to make the design more suitable for the local application, so that the design is more in line with people's needs.

[0004] However, in the traditional design optimization process, relying solely on manual experience to adjust the building model may result in low efficiency and incomplete adjustments, leading to low quality of the building model. Summary of the Invention

[0005] Therefore, it is necessary to provide a method, apparatus, computer equipment, and storage medium for identifying water supply and drainage system models that can improve the quality of building models, in order to address the aforementioned technical problems.

[0006] In a first aspect, embodiments of this application provide an artificial intelligence-based building model recognition method, the method comprising:

[0007] Obtain building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute;

[0008] The images of the building components are input into a preset recognition network to identify the attribute relationships between the building components in the building model, and the recognition results corresponding to the attribute relationships between the building components are obtained.

[0009] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0010] In one embodiment, the step of inputting the building component image into a preset recognition network to determine the attribute relationships between the building components in the building model and obtaining the recognition results corresponding to the attribute relationships between the building components includes:

[0011] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0012] Obtain each relationship in the knowledge graph and the architectural design rule to obtain the recognition result;

[0013] Prompts will be provided based on the recognition results.

[0014] In one embodiment, the association relationship includes at least one of the following: water supply and drainage system association relationship, electrical system association relationship, HVAC system association relationship, functional site association relationship, and structural body association relationship.

[0015] In one embodiment, before obtaining the water supply and drainage image corresponding to the design model, the method further includes:

[0016] Obtain the building model;

[0017] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components.

[0018] In one embodiment, the process of acquiring the identification network includes:

[0019] The architectural sample model is converted from three-dimensional to two-dimensional to obtain multiple initial sample images of architectural components;

[0020] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0021] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the classification method of clustering algorithm to obtain the recognition network including the building design rules.

[0022] Secondly, embodiments of this application provide an artificial intelligence-based building model recognition method, the method comprising:

[0023] The building sample model with pre-labeled component attributes is converted from three-dimensional to two-dimensional to obtain multiple initial sample images of building components;

[0024] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0025] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the clustering algorithm classification method to obtain the recognition network including the building design rules;

[0026] Obtain the building model;

[0027] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components;

[0028] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0029] The identification result is obtained by acquiring each relationship in the knowledge graph and the building design rule; wherein, the relationship includes at least one of the following: water supply and drainage system relationship, electrical system relationship, HVAC system relationship, functional site relationship, and structural body relationship;

[0030] Prompts will be provided based on the recognition results.

[0031] Thirdly, embodiments of this application provide an artificial intelligence-based building model recognition device, the device comprising:

[0032] The acquisition module is used to acquire images of building components corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute;

[0033] The recognition module is used to input the building component image into a preset recognition network to determine the attribute relationships between the building components in the building model, and to obtain the recognition results corresponding to the attribute relationships between the building components.

[0034] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0035] Fourthly, embodiments of this application provide an artificial intelligence-based building model recognition device, the device comprising:

[0036] The training module is used to convert a pre-annotated architectural sample model of components from three dimensions to two dimensions to obtain multiple initial sample images of architectural components. The initial sample images of architectural components are preprocessed to obtain architectural sample images. The architectural sample images are then input into an initial recognition network, and association relationships are extracted based on a clustering algorithm classification method to obtain the recognition network including the architectural design rules. The preprocessing includes at least one of filtering and stitching.

[0037] The processing module is used to acquire a building model, perform a 3D to 2D conversion on the building model to obtain multiple building component images, input the building component images into the recognition network to identify the attribute relationships between the building components, obtain a knowledge graph of the component attributes, and acquire each association relationship in the knowledge graph and the building design rules to obtain the recognition result.

[0038] The aforementioned relationships include at least one of the following: water supply and drainage system relationships, electrical system relationships, HVAC system relationships, functional site relationships, and structural entity relationships;

[0039] The prompting module is used to provide prompts based on the recognition results.

[0040] Fifthly, embodiments of this application provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0041] Obtain building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute;

[0042] The images of the building components are input into a preset recognition network to identify the attribute relationships between the building components in the building model, and the recognition results corresponding to the attribute relationships between the building components are obtained.

[0043] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0044] Sixthly, embodiments of this application provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0045] The building sample model with pre-labeled component attributes is converted from three-dimensional to two-dimensional to obtain multiple initial sample images of building components;

[0046] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0047] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the clustering algorithm classification method to obtain the recognition network including the building design rules;

[0048] Obtain the building model;

[0049] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components;

[0050] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0051] The identification result is obtained by acquiring each relationship in the knowledge graph and the building design rule; wherein, the relationship includes at least one of the following: water supply and drainage system relationship, electrical system relationship, HVAC system relationship, functional site relationship, and structural body relationship;

[0052] Prompts will be provided based on the recognition results.

[0053] In a seventh aspect, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, performs the following steps:

[0054] Obtain building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute;

[0055] The images of the building components are input into a preset recognition network to identify the attribute relationships between the building components in the building model, and the recognition results corresponding to the attribute relationships between the building components are obtained.

[0056] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0057] Eighthly, embodiments of this application provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, performs the following steps:

[0058] The building sample model with pre-labeled component attributes is converted from three-dimensional to two-dimensional to obtain multiple initial sample images of building components;

[0059] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0060] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the clustering algorithm classification method to obtain the recognition network including the building design rules;

[0061] Obtain the building model;

[0062] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components;

[0063] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0064] The identification result is obtained by acquiring each relationship in the knowledge graph and the building design rule; wherein, the relationship includes at least one of the following: water supply and drainage system relationship, electrical system relationship, HVAC system relationship, functional site relationship, and structural body relationship;

[0065] Prompts will be provided based on the recognition results.

[0066] The aforementioned AI-based building model recognition method, apparatus, computer equipment, and medium acquire images of building components corresponding to the building model through the computer equipment, input these images into a preset recognition network to identify the relationships within the building model, and obtain the recognition results corresponding to these relationships. Since the recognition network is a neural network trained using a clustering algorithm, and it learns different architectural design rules through clustering and learning from different types of building component images, it can automatically identify the relationships within the building model from the building component images and compare them with the aforementioned architectural design rules. This yields recognition results characterizing the degree of matching between the relationships within the building model and the architectural design rules. Therefore, it avoids the problems of low efficiency, low accuracy, incomplete adjustments, and high human learning costs that can result from traditional manual adjustments and optimizations of building models based on experience. This method automatically identifies the relationships within the building model using a recognition network and obtains recognition results based on the mined architectural design rules. Therefore, the recognition efficiency is greatly improved, and the relationships within the current building model can be identified more accurately and comprehensively. The accuracy rate is also significantly improved, while the human learning cost is reduced, greatly saving time and manpower. Attached Figure Description

[0067] Figure 1 This is an internal structural diagram of a computer device in one embodiment;

[0068] Figure 2 A flowchart illustrating an AI-based building model recognition method provided in one embodiment;

[0069] Figure 3 A flowchart illustrating an artificial intelligence-based building model recognition method provided for another embodiment;

[0070] Figure 4 A flowchart illustrating an AI-based building model recognition method as another embodiment;

[0071] Figure 5 A flowchart illustrating an AI-based building model recognition method as another embodiment;

[0072] Figure 6 A schematic diagram of an artificial intelligence-based building model recognition device provided in one embodiment;

[0073] Figure 7 A schematic diagram of the structure of an artificial intelligence-based building model recognition device provided in yet another embodiment. Detailed Implementation

[0074] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0075] The artificial intelligence-based building model recognition method provided in this application embodiment can be applied to... Figure 1 The computer device shown includes a processor, memory, network interface, database, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage medium. The database stores the identification network described in the following embodiments; a detailed description of the identification network is provided in the following embodiments. The network interface of the computer device can be used to communicate with other external devices via a network connection. Optionally, the computer device can be a server, a desktop computer, a personal digital assistant, or other terminal devices such as tablets, mobile phones, etc., or a cloud or remote server. This application does not limit the specific form of the computer device. The display screen of the computer device can be a liquid crystal display or an e-ink display. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device casing, or an external keyboard, touchpad, or mouse. Of course, input devices and displays may not be part of the computer equipment; they can be external devices to the computer equipment.

[0076] Those skilled in the art will understand that Figure 1The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0077] The technical solution of this application and how it solves the above-mentioned technical problems will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will be described below with reference to the accompanying drawings.

[0078] It should be noted that the executing entity of the following method embodiments can be an artificial intelligence-based building model recognition device, which can be implemented as part or all of the aforementioned computer equipment through software, hardware, or a combination of software and hardware. The following method embodiments are described using a computer equipment as an example of the executing entity.

[0079] Figure 2 This is a flowchart illustrating an artificial intelligence-based building model recognition method provided in one embodiment. This embodiment relates to the process of using computer equipment and artificial neural networks to assist in the design of building models. Figure 2 As shown, it includes:

[0080] S11. Obtain building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute. The building component images can be two-dimensional planar design drawings or planar screenshots of three-dimensional building models.

[0081] Specifically, the computer device can read pre-stored two-dimensional building component images, receive building component images sent by other devices, and obtain building component images by recognizing the building model; this embodiment does not limit the scope of these methods. It should be noted that the aforementioned building component images are obtained by the computer device converting the building model from three-dimensional to two-dimensional, resulting in two-dimensional building component images with different angles and component attributes. Optionally, the aforementioned component attributes can be entity objects in the building model, including but not limited to objects in the water supply and drainage system model, such as water supply pipes, drainage pipes, and entities related to water supply and drainage, such as sinks, showers, toilets, etc.; they can also include objects in the electrical system, such as circuits, electrical equipment, distribution boxes, etc.; they can also include objects in the heating, ventilation, and air conditioning system, such as ventilation ducts, air vents, etc.; and they can also include objects in the main building structure, such as walls, doors, and windows, etc.

[0082] S12. Input the building component image into a preset recognition network to perform attribute relationship analysis between the building components in the building model, and obtain the recognition result corresponding to the attribute relationship between the building components.

[0083] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0084] Specifically, the computer inputs the aforementioned building component images into a preset recognition network. Since this recognition network is a neural network trained using a clustering algorithm, it learns by clustering and studying the two-dimensional building component images of multiple building models. This network uncovers patterns in the various relationships between different component attributes and uses these patterns as architectural design rules. Therefore, the recognition network can identify the relationships between building component images and determine the degree of matching between these relationships and the discovered architectural design rules. This results in a determination of whether the current building model meets the design requirements. If the degree of matching between the building model's relationships and the discovered architectural design rules is high, the current building model design is considered to meet the design requirements, indicating high design quality. Conversely, if the degree of matching is low, the current building model design may not meet the design requirements, indicating low design quality. Optionally, the above-mentioned relationship can be the relationship between different component attributes, such as the relationship between the daylighting rate of a south-facing bedroom and the window size, or the relationship between the attributes or values ​​of different component attributes. The influence range of doors and windows on indicators such as noise and humidity is not limited in this embodiment.

[0085] In this embodiment, a computer device acquires images of building components corresponding to a building model and inputs these images into a preset recognition network to identify the relationships between the building models, thus obtaining recognition results corresponding to these relationships. Since the aforementioned recognition network is a neural network trained using a clustering algorithm, and this network learns different architectural design rules through clustering and learning from different types of building component images, it can automatically identify the relationships between the building components and compare them with the architectural design rules. This yields recognition results that characterize the degree of matching between the relationships between the building models and the architectural design rules. Therefore, this method avoids the problems of low efficiency, low accuracy, incomplete adjustments, and high human learning costs that may result from traditional manual adjustments and optimizations of building models based on experience. This method automatically identifies the relationships between the building models using a recognition network and obtains recognition results based on the mined architectural design rules. Therefore, the recognition efficiency is greatly improved, and the relationships between the current building models can be identified more accurately and comprehensively, significantly increasing accuracy while reducing human learning costs and greatly saving time and manpower.

[0086] Optionally, based on the above embodiments, one possible implementation of step S12 can be as follows: Figure 2 As shown, it includes:

[0087] S121. Input the building component image into the recognition network to identify the attribute relationships between the building components and obtain a knowledge graph of the component attributes.

[0088] The attributes mentioned are the properties of each building component in the building model. For example, a wall may be a load-bearing wall, and the load-bearing wall is an attribute of that wall. The relationship between that wall and other walls is also called an attribute, i.e., an association relationship.

[0089] Specifically, the computer device inputs the aforementioned water supply and drainage images into the aforementioned recognition network, which can identify multiple water supply and drainage objects and multiple different attribute relationships between different water supply and drainage objects, and then generates a knowledge graph from these relationships.

[0090] S122. Based on each association in the knowledge graph and the architectural design rules, the identification result is obtained.

[0091] Specifically, the computer device evaluates each relationship in the knowledge graph according to architectural design rules, obtaining the degree of matching between each relationship and the corresponding architectural design rule as the recognition result. This can include multiple matching degree levels, such as complete match, partial match, and no match. Optionally, the matching degree can also be quantified, which is not limited in this embodiment. For example, if the architectural design rule states that the window area of ​​a north-facing bedroom is between A square meters and B square meters, and the computer device obtains a relationship that includes a window in a north-facing bedroom, but the window area is less than A square meters, then the computer device can determine that the current window design does not meet the lighting requirements, and the resulting recognition result can include that the window area of ​​an inward-facing bedroom is less than the minimum area requirement.

[0092] S123. Provide prompts based on the recognition results.

[0093] Specifically, when the correlation in the above identification results does not match the building design rules at all, the computer device can output a mismatch prompt to remind the user to pay attention to the component attributes involved in the correlation. If the building design rules state that the window area of ​​the north-facing bedroom is between A square meters and B square meters, and the correlation obtained by the computer device includes the window of the north-facing bedroom, but the area of ​​the window is less than A square meters, then the computer device can determine that the current window design does not meet the lighting requirements. The obtained identification results can include that the window area of ​​the inward-facing bedroom does not meet the design requirements, and therefore a prompt can be output to ask whether the window size needs to be increased.

[0094] In this embodiment, the computer device inputs building component images into a recognition network for relational identification, obtaining a knowledge graph of component attributes. It then acquires each relation and architectural design rule from the knowledge graph, generating a recognition result, and provides prompts based on this result. This method can obtain recognition results based on the matching degree between each relation and architectural design rule, and provide prompts accordingly. This enables prompts for relations that do not match architectural design rules, achieving intelligent assisted design. Therefore, it makes the design of building models more accurate, reasonable, and intelligent.

[0095] Optionally, the above architectural design rules are derived from learning from multiple architectural models. In a northern architectural model, the living room balcony is designed to be 2 square meters, resulting in insufficient sunlight. If the balcony were designed to be 15 square meters, the living room would be smaller, receiving more sunlight, but this would negatively impact indoor temperature and increase outdoor noise intrusion. Therefore, an "optimal solution" needs to be selected from various factors. Transmission tools can be used to assist designers in obtaining the impact of changes in balcony and living room areas on temperature and noise. By comparing the technical indicators (temperature and noise) before and after the transformation, the optimal solution is one that maximizes the balcony area while keeping both temperature and noise levels within acceptable ranges. Optionally, large-scale data collection can be conducted for different regions, such as collecting architectural data for a specific northern location, including temperature, humidity, atmospheric pressure, solar radiation intensity, ice storage air conditioning equipment power, and control signals. Then, a suitable clustering algorithm, such as the density-based clustering algorithm DBSCAN, can be determined based on these characteristics. The association rules of architectural models with similar characteristics can be statistically analyzed to obtain the above architectural design rules.

[0096] Optionally, the relationships include at least one of the following: water supply and drainage system relationships, electrical system relationships, HVAC system relationships, functional site relationships, and structural system relationships. Specifically, water supply and drainage system relationships can include relationships between water supply and drainage objects, such as outlets and drains, and relationships between water supply and drainage objects and other objects, such as the relationship of water pipes attached to walls. HVAC system relationships can include relationships between pipes and cavities, hierarchical relationships between different pipes, and positional relationships. Functional site relationships can include relationships between different functional points of a building, such as the relationship between a community power distribution room and the community entrance. Structural system relationships can also include relationships between the main floor structure and other floors.

[0097] Optionally, before S11 above, the process may further include: acquiring a building model; performing a three-dimensional to two-dimensional conversion on the water supply and drainage system model to obtain multiple building component images. Specifically, the computer device can traverse the design model, perform a three-dimensional to two-dimensional conversion based on the attributes of the traversed components, and obtain multiple two-dimensional target images. Optionally, the computer device can also filter the obtained two-dimensional building component images, deleting some unclear images, or stitching together some related building component images to obtain a complete image of the component attributes. In this embodiment, the computer device acquires the building model in the design model, performs a three-dimensional to two-dimensional conversion on the building model, and obtains multiple building component images, thereby enabling automatic acquisition of building component images from the complete design model. Therefore, the degree of automation is higher, further improving recognition efficiency and accuracy.

[0098] Optionally, based on the above embodiments, the training process of the recognition network described above is a learning process of architectural design rules for the building model, such as... Figure 4 As shown, it includes:

[0099] S131. Convert the building sample model from three-dimensional to two-dimensional to obtain multiple initial sample images of building components.

[0100] Specifically, computer equipment can acquire multiple architectural sample models as learning samples. These architectural sample models include different types of architectural models, such as southern residential buildings, northern residential buildings, southern office buildings, northern high-rise buildings, etc.

[0101] S132. Preprocess the initial sample image of the building component to obtain a building sample image; wherein the preprocessing includes at least one of filtering and stitching.

[0102] Specifically, the computer equipment preprocesses the initial sample images of the aforementioned building components. This preprocessing may include data cleaning and filtering, removing unclear or unrecognizable images, such as images containing only a water pipe head or images from which related objects cannot be identified. It may also include stitching, for example, stitching together different images of the same component attribute to obtain a complete component attribute, such as stitching together images of water pipes and pools. Furthermore, the initial sample images may be normalized to obtain water supply and drainage sample images with consistent size or pixel count. This embodiment does not limit the specific processes of filtering and stitching included in the above preprocessing, as long as valid building component images are obtained.

[0103] S133. Input the water supply and drainage sample image into the initial recognition network, extract the association relationship based on the classification method of clustering algorithm, and obtain the recognition network including the water supply and drainage design rules.

[0104] Optionally, the aforementioned associations can broadly include the following four categories: Boolean association rules, quantification rules, one-dimensional and multi-dimensional rules, and single-layer and multi-layer association rules. For example, a Boolean association rule could be that if X exists, Y must also exist, meaning that if Y does not exist, the requirement is not met; a quantification rule could be a rule for the quantified value of a certain indicator of a water supply and drainage object, such as the diameter of a water pipe being above 3cm and below 25cm; one-dimensional and multi-dimensional rules could be two-dimensional associations between a faucet and the water supply point and the drain; and one-layer and multi-layer association rules could be a single-layer association between the main inlet pipe and the next-level water pipe, and a multi-layer association between the main inlet pipe and the end pipe.

[0105] Specifically, the computer device inputs the aforementioned building sample images into an initial recognition network. This recognition network can use a clustering algorithm to classify the aforementioned building sample images and extract the association relationships of different component attributes based on the classified images. Optionally, it can also obtain frequent itemsets based on these association relationships to obtain the recognition network that includes building design rules.

[0106] Clustering algorithms, such as the density-based clustering algorithm DBSCAN, are used for data collection. This allows analysis to reveal a close correlation between energy utilization efficiency and human activity, resulting in n energy consumption time distribution classes. Time periods with similar energy consumption are grouped together, and the energy consumption distribution rate is obtained through energy distribution weights. Computer equipment can then input the collected data into a recognition network for calculations and analysis, generating a load allocation chart.

[0107] Alternatively, object detection and segmentation techniques from Mast R-CNN and TensorFlow can be used to perform deep learning on water supply and drainage objects. Based on the labeled images of water supply and drainage, the water pipes and water appliances can be identified to obtain a trained model. For example, 50,000 labeled JPG images can be selected for training, achieving a recognition rate of 95.1%.

[0108] In this embodiment, the computer device can convert the architectural sample model from three-dimensional to two-dimensional to obtain multiple initial sample images of architectural components. These initial sample images are then preprocessed to obtain architectural sample images, which are then input into an initial recognition network. Based on a clustering algorithm, association relationships are extracted to obtain a recognition network that includes architectural design rules. This method, by mining and learning the relationships within the design sample model, can acquire more implicit relationships and patterns, resulting in more comprehensive architectural design rules. Therefore, it can comprehensively and effectively identify architectural models based on artificial intelligence, better supplementing existing manual design methods and further improving the design quality of architectural models.

[0109] To describe the technical solution provided in this application in more detail, a specific embodiment is used to illustrate this application, such as... Figure 5 As shown, it includes:

[0110] S21. Convert the pre-annotated building sample model with component attributes from three-dimensional to two-dimensional to obtain multiple initial sample images of building components;

[0111] S22. Preprocess the initial sample image of the building component to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0112] S23. Input the building sample image into the initial recognition network, extract the association relationship based on the clustering algorithm classification method, and obtain frequent itemsets to obtain the recognition network including the building design rules;

[0113] S24. Obtain the building model;

[0114] S25. Perform a three-dimensional to two-dimensional conversion on the building model to obtain multiple images of the building components;

[0115] S26. Input the building component image into the recognition network to identify the attribute relationships between the building components and obtain a knowledge graph of the component attributes;

[0116] S27. Obtain each association in the knowledge graph and the building design rule to obtain the identification result; wherein, the association includes at least one of the following: water supply and drainage system association, electrical system association, HVAC system association, functional site association, and structural main body association;

[0117] S28. Provide prompts based on the recognition results.

[0118] For a detailed description of the steps and technical effects in this embodiment, please refer to the foregoing embodiments, which will not be repeated here.

[0119] It should be understood that, although Figure 2-5 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 2-5 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0120] In one embodiment, such as Figure 6 As shown, a device for recognizing building system models is provided, comprising:

[0121] The acquisition module 100 is used to acquire building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute;

[0122] The recognition module 200 is used to input the building component image into a preset recognition network to determine the attribute relationships between the building components in the building model, and to obtain the recognition results corresponding to the attribute relationships between the building components.

[0123] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0124] In one embodiment, the recognition module 200 is specifically used to input the building component image into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes; to obtain each association relationship and the building design rule in the knowledge graph, thereby obtaining the recognition result; and to provide prompts based on the recognition result.

[0125] In one embodiment, the association relationship includes at least one of the following: water supply and drainage system association relationship, electrical system association relationship, HVAC system association relationship, functional site association relationship, and structural body association relationship.

[0126] In one embodiment, the acquisition module 100 is specifically used to acquire the building model; and to perform a three-dimensional to two-dimensional conversion on the building model to obtain multiple images of the building components.

[0127] In one embodiment, the recognition module 200 is specifically used to convert the building sample model from three-dimensional to two-dimensional to obtain multiple initial sample images of building components; preprocess the initial sample images of building components to obtain building sample images; input the building sample images into an initial recognition network, and extract the association relationship based on the classification method of clustering algorithm to obtain the recognition network including the building design rules; wherein, the preprocessing includes at least one of screening and stitching.

[0128] In one embodiment, such as Figure 7 As shown, an artificial intelligence-based building model recognition device is provided, comprising:

[0129] The training module 300 is used to convert the building sample model from three dimensions to two dimensions to obtain multiple initial sample images of building components; to preprocess the initial sample images of building components to obtain building sample images; and to input the building sample images into an initial recognition network to extract the association relationship based on a clustering algorithm classification method to obtain the recognition network including the building design rules; wherein, each building sample model includes multiple component attributes, and the preprocessing includes at least one of filtering and stitching;

[0130] The processing module 400 is used to acquire a building model, perform a 3D-to-2D conversion on the building model to obtain multiple building component images, input the building component images into the recognition network to identify the attribute relationships between the building components, obtain a knowledge graph of the component attributes, and acquire each association relationship in the knowledge graph and the building design rules to obtain the recognition result; wherein, the association relationship includes at least one of the following: water supply and drainage system association relationship, electrical system association relationship, HVAC system association relationship, functional site association relationship, and structural body association relationship;

[0131] The prompting module 500 is used to provide prompts based on the recognition results.

[0132] Specific limitations regarding the AI-based building model recognition device can be found in the above description of the limitations of the AI-based building model recognition method, and will not be repeated here. The modules in the aforementioned AI-based building model recognition device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0133] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0134] Obtain building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute;

[0135] The images of the building components are input into a preset recognition network to identify the attribute relationships between the building components in the building model, and the recognition results corresponding to the attribute relationships between the building components are obtained.

[0136] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0137] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0138] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0139] Obtain each relationship in the knowledge graph and the architectural design rule to obtain the recognition result;

[0140] Prompts will be provided based on the recognition results.

[0141] In one embodiment, the association relationship includes at least one of the following: water supply and drainage system association relationship, electrical system association relationship, HVAC system association relationship, functional site association relationship, and structural body association relationship.

[0142] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0143] Obtain the building model;

[0144] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components.

[0145] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0146] The architectural sample model is converted from three-dimensional to two-dimensional to obtain multiple initial sample images of architectural components;

[0147] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0148] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the classification method of clustering algorithm to obtain the recognition network including the building design rules.

[0149] It should be clear that the process of the processor executing the computer program in the embodiments of this application is consistent with the execution process of each step in the above method, as can be seen in the description above.

[0150] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the following steps:

[0151] The architectural sample model is converted from 3D to 2D to obtain multiple initial sample images of architectural components; each architectural sample model includes multiple component attributes.

[0152] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0153] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the clustering algorithm classification method to obtain the recognition network including the building design rules;

[0154] Obtain the building model;

[0155] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components;

[0156] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0157] The identification result is obtained by acquiring each relationship in the knowledge graph and the building design rule; wherein, the relationship includes at least one of the following: water supply and drainage system relationship, electrical system relationship, HVAC system relationship, functional site relationship, and structural body relationship;

[0158] Prompts will be provided based on the recognition results.

[0159] It should be clear that the process of the processor executing the computer program in the embodiments of this application is consistent with the execution process of each step in the above method, as can be seen in the description above.

[0160] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0161] Obtain building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute;

[0162] The images of the building components are input into a preset recognition network to identify the attribute relationships between the building components in the building model, and the recognition results corresponding to the attribute relationships between the building components are obtained.

[0163] The identification results are used to characterize the degree of matching between the attribute relationships of each building component and the building design rules, and the identification network is a neural network obtained by mining building design rules using a clustering algorithm.

[0164] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0165] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0166] Obtain each relationship in the knowledge graph and the architectural design rule to obtain the recognition result;

[0167] Prompts will be provided based on the recognition results.

[0168] In one embodiment, the association relationship includes at least one of the following: water supply and drainage system association relationship, electrical system association relationship, HVAC system association relationship, functional site association relationship, and structural body association relationship.

[0169] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0170] Obtain the building model;

[0171] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components.

[0172] In one embodiment, when the computer program is executed by a processor, it further performs the following steps:

[0173] The architectural sample model is converted from three-dimensional to two-dimensional to obtain multiple initial sample images of architectural components;

[0174] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0175] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the classification method of clustering algorithm to obtain the recognition network including the building design rules.

[0176] It should be clear that the process of the processor executing the computer program in the embodiments of this application is consistent with the execution process of each step in the above method, as can be seen in the description above.

[0177] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0178] The architectural sample model is converted from 3D to 2D to obtain multiple initial sample images of architectural components; each architectural sample model includes multiple component attributes.

[0179] The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching;

[0180] The building sample image is input into the initial recognition network, and the association relationship is extracted based on the clustering algorithm classification method to obtain the recognition network including the building design rules;

[0181] Obtain the building model;

[0182] The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components;

[0183] The images of the building components are input into the recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes.

[0184] The identification result is obtained by acquiring each relationship in the knowledge graph and the building design rule; wherein, the relationship includes at least one of the following: water supply and drainage system relationship, electrical system relationship, HVAC system relationship, functional site relationship, and structural body relationship;

[0185] Prompts will be provided based on the recognition results.

[0186] It should be clear that the process of the processor executing the computer program in the embodiments of this application is consistent with the execution process of each step in the above method, as can be seen in the description above.

[0187] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0188] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0189] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.

Claims

1. A building model recognition method based on artificial intelligence, characterized in that, The method includes: Obtain building component images corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute; The images of the building components are input into a preset recognition network to identify the attribute relationships between the building components in the building model, and the recognition results corresponding to the attribute relationships between the building components are obtained. The identification results are used to characterize the degree of matching between the attribute relationships between various building components and the building design rules. The preset identification network is a neural network obtained by mining building design rules using a clustering algorithm. The process of acquiring the preset recognition network includes: converting the building sample model from three dimensions to two dimensions to obtain multiple initial sample images of building components; preprocessing the initial sample images of building components to obtain building sample images; wherein, the preprocessing includes at least one of filtering and stitching; inputting the building sample images into the initial recognition network, extracting the association relationship based on the classification method of clustering algorithm, and obtaining the preset recognition network including the building design rules.

2. The method according to claim 1, characterized in that, The step of inputting the building component images into a preset recognition network to identify the attribute relationships between the building components in the building model and obtaining the recognition results corresponding to the attribute relationships between the building components includes: The images of the building components are input into the preset recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes; Obtain each relationship in the knowledge graph and the architectural design rule to obtain the recognition result; Prompts will be provided based on the recognition results.

3. The method according to claim 2, characterized in that, The relationships mentioned include at least one of the following: water supply and drainage system relationships, electrical system relationships, HVAC system relationships, functional site relationships, and structural entity relationships.

4. The method according to claim 1, characterized in that, The process of obtaining the building component images corresponding to the building model includes: Obtain the building model; The building model is converted from three-dimensional to two-dimensional to obtain multiple images of the building components.

5. A building model recognition method based on artificial intelligence, characterized in that, The method includes: The architectural sample model is converted from 3D to 2D to obtain multiple initial sample images of architectural components; each architectural sample model includes multiple component attributes. The initial sample image of the building component is preprocessed to obtain a building sample image; wherein, the preprocessing includes at least one of filtering and stitching; The building sample images are input into the initial recognition network, and the association relationship is extracted based on the clustering algorithm classification method to obtain a preset recognition network including building design rules; Obtain the building model; The architectural model is converted from three-dimensional to two-dimensional to obtain multiple images of architectural components; The images of the building components are input into the preset recognition network to identify the attribute relationships between the building components, thereby obtaining a knowledge graph of the component attributes; Obtain each association in the knowledge graph and the building design rule to obtain the recognition result; wherein, the association includes at least one of the following: water supply and drainage system association, electrical system association, HVAC system association, functional site association, and structural body association; Prompts will be provided based on the recognition results.

6. A building model recognition device based on artificial intelligence, characterized in that, The device includes: The acquisition module is used to acquire images of building components corresponding to the building model; wherein, the building component images are images obtained by converting the building model from three-dimensional to two-dimensional, and each building component image represents at least one component attribute; The recognition module is used to input the building component image into a preset recognition network to recognize the attribute relationships between the building components in the building model, and to obtain the recognition results corresponding to the attribute relationships between the building components. The identification results are used to characterize the degree of matching between the attribute relationships between various building components and the building design rules. The preset identification network is a neural network obtained by mining building design rules using a clustering algorithm. The process of acquiring the preset recognition network includes: converting the building sample model from three dimensions to two dimensions to obtain multiple initial sample images of building components; preprocessing the initial sample images of building components to obtain building sample images; wherein, the preprocessing includes at least one of filtering and stitching; inputting the building sample images into the initial recognition network, extracting the association relationship based on the classification method of clustering algorithm, and obtaining the preset recognition network including the building design rules.

7. A building model recognition device based on artificial intelligence, characterized in that, The device includes: The training module is used to convert the building sample model from three-dimensional to two-dimensional to obtain multiple initial sample images of building components; to preprocess the initial sample images of building components to obtain building sample images; and to input the building sample images into an initial recognition network to extract the association relationship based on the clustering algorithm classification method to obtain a preset recognition network including building design rules; wherein, each building sample model includes multiple component attributes, and the preprocessing includes at least one of filtering and stitching; The processing module is used to acquire a building model, perform a 3D to 2D conversion on the building model to obtain multiple building component images, input the building component images into the preset recognition network to identify the attribute relationships between the building components, obtain a knowledge graph of the component attributes, and obtain each association relationship in the knowledge graph and the building design rules to obtain the recognition result; The aforementioned relationships include at least one of the following: water supply and drainage system relationships, electrical system relationships, HVAC system relationships, functional site relationships, and structural entity relationships; The prompting module is used to provide prompts based on the recognition results.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.

Images