A data generation method and system for building information model (BIM)

Abstract

The present invention provides a data generation method and system for a building information model (BIM), which relates to the technical field of data processing. In the present invention, based on the exemplary building information model data included in the exemplary building information model data cluster, the constructed building information analysis neural network is optimized in at least two stages to form corresponding target building information model data; the target building information model data is extracted; the target building information model data is loaded to analyze and generate model description data corresponding to the target building information model data using the first building information analysis neural network, and the model description data is used as identification data of the target building information model data. Based on the above method, the reliability of the identification data of the target building information model data can be improved to a certain extent.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a data generation method and system for a Building Information Model (BIM). Background Art

[0002] BIM (Building Information Modeling) technology is a digital tool used in engineering design, construction, and management. By integrating the digital and information models of buildings, it is shared and transmitted throughout the entire life cycle of project planning, operation, and maintenance, enabling engineering and technical personnel to correctly understand and efficiently respond to various building information, providing a basis for collaborative work for the design team and all construction entities including construction and operation units, and playing an important role in improving production efficiency, saving costs, and shortening construction periods. Among them, in existing applications, it is necessary to determine the identification data of the target building information model data to describe the target building information model data. However, in the existing technology, labeling is based on manual experience, which makes its reliability poor. Summary of the invention

[0003] In view of this, an object of the present invention is to provide a data generation method and system for a building information model (BIM), so as to improve the reliability of identification data for determining target building information model data to a certain extent.

[0004] To achieve the above objectives, the embodiments of the present invention adopt the following technical solutions:

[0005] A data generation method for a building information model (BIM), comprising:

[0006] According to the exemplary building information model data included in the exemplary building information model data cluster, the constructed building information analysis neural network is optimized in at least two stages to form corresponding target building information model data;

[0007] Extracting target building information model data;

[0008] The target building information model data is loaded to analyze and generate model description data corresponding to the target building information model data using a first building information analysis neural network, and the model description data is used as identification data of the target building information model data.

[0009] In some preferred embodiments, in the data generation method of the building information model BIM, the step of optimizing the constructed building information analysis neural network in at least two stages according to the exemplary building information model data included in the exemplary building information model data cluster to form the corresponding target building information model data includes:

[0010] Extracting an exemplary building information model data cluster, wherein the exemplary building information model data cluster includes exemplary building information model data corresponding to at least two optimization stages;

[0011] Extracting first exemplary building information model data corresponding to the first optimization stage, and optimizing the feature mining subnetwork included in the constructed building information analysis neural network according to the first exemplary building information model data to form a corresponding candidate building information analysis neural network;

[0012] extracting second exemplary building information model data corresponding to a subsequent optimization stage, and optimizing the feature mining subnetwork included in the candidate building information analysis neural network based on the second exemplary building information model data to form a corresponding optimized building information analysis neural network;

[0013] Loading the used building information model data corresponding to the second exemplary building information model data into the feature mining subnetworks included in the candidate building information analysis neural network and the optimized building information analysis neural network, respectively, to output corresponding candidate building information feature mining results and optimized building information feature mining results;

[0014] According to the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data, the data conversion output subnetwork included in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network;

[0015] The step of reverting to the step of extracting the second exemplary building information model data corresponding to the next optimization stage, and optimizing the feature mining subnetwork included in the candidate building information analysis neural network based on the second exemplary building information model data, and marking the corresponding candidate building information analysis neural network at the end of the optimization to mark and form the corresponding first building information analysis neural network.

[0016] In some preferred embodiments, in the above-mentioned method for generating data of a building information model (BIM), the step of extracting an exemplary building information model data cluster comprises:

[0017] Extracting a plurality of initial exemplary building information model data, each of the initial exemplary building information model data having initial identification data;

[0018] Performing a data aggregation operation on each initial exemplary building information model data corresponding to the same initial identification data to form a building information model data aggregation result corresponding to each initial identification data;

[0019] Analyzing the phase order relationship information corresponding to each initial exemplary building information model data according to the initial identification data;

[0020] Marking the building information model data aggregation results to form corresponding optimization sub-stages, performing data aggregation operations on each optimization sub-stage corresponding to the same stage sorting relationship information to form an optimization stage aggregation result corresponding to each stage sorting relationship information, wherein the optimization stage aggregation result includes each exemplary building information model data corresponding to the same optimization stage;

[0021] According to each of the optimization stage aggregation results, a corresponding exemplary building information model data cluster is formed.

[0022] In some preferred embodiments, in the data generation method of the building information model BIM, the exemplary building information model data in the exemplary building information model data cluster has optimization identification data, the building information analysis neural network to be optimized belongs to the constructed building information analysis neural network or the candidate building information analysis neural network, and the feature mining subnetwork included in the building information analysis neural network to be optimized is optimized, including the following steps:

[0023] Using the building information analysis neural network to be optimized, the loaded exemplary building information model data is subjected to a building information model analysis operation to analyze and generate analysis identification data corresponding to the loaded exemplary building information model data;

[0024] Based on the identification data distinction information between the optimization identification data and the corresponding analysis identification data corresponding to the loaded exemplary building information model data, the sub-network parameters of the feature mining sub-network included in the building information analysis neural network to be optimized are optimized, so as to complete the optimization of the building information analysis neural network to be optimized.

[0025] In some preferred embodiments, in the data generation method of the above-mentioned building information model BIM, the target building information analysis neural network belongs to the constructed building information analysis neural network, the candidate building information analysis neural network, the optimized building information analysis neural network or the first building information analysis neural network, the feature mining subnetwork included in the target building information analysis neural network is used to perform feature mining operations on the loaded exemplary building information model data to output the corresponding first building information feature mining results, and the data conversion output subnetwork included in the target building information analysis neural network is used to perform data conversion and output operations on the first building information feature mining results to output the corresponding second building information feature mining results.

[0026] In some preferred embodiments, in the data generation method of the building information model BIM, the feature mining subnetwork includes at least two feature mining units, the data conversion output subnetwork includes at least two data conversion output units, there is a one-to-one correspondence between the at least two feature mining units and the at least two data conversion output units, the first building information feature mining result includes a local first building information feature mining result mined by each feature mining unit, and the data conversion output operation of the second building information feature mining result includes:

[0027] The local first building information feature mining results mined by each of the feature mining units are loaded into the corresponding data conversion and output units, so as to use the data conversion and output units to convert and output the local second building information feature mining results corresponding to each data conversion and output unit;

[0028] According to each of the local second building information feature mining results, a corresponding second building information feature mining result is determined.

[0029] In some preferred embodiments, in the data generation method of the building information model BIM, the feature mining subnetwork includes at least two feature mining units, the data conversion output subnetwork includes at least two data conversion output units, and there is a one-to-one correspondence between at least two of the feature mining units and at least two of the data conversion output units, the candidate building information feature mining results include the local candidate building information feature mining results mined by each feature mining unit of the candidate building information analysis neural network, and the optimized building information feature mining results include the local optimized building information feature mining results mined by each feature mining unit of the optimized building information analysis neural network;

[0030] The step of optimizing the data conversion output subnetwork included in the optimized building information analysis neural network based on the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data to form a corresponding optimized candidate building information analysis neural network comprises:

[0031] Respectively loading each local optimized building information feature mining result corresponding to the used building information model data into the corresponding data conversion output unit in the optimized building information analysis neural network to convert and output each corresponding local optimized building information feature conversion result;

[0032] Marking the corresponding feature mining unit and data conversion output unit to mark them as the corresponding first feature mining unit and first data conversion output unit;

[0033] According to the local candidate building information feature mining results mined by the first feature mining unit and the local optimized building information feature conversion results converted by the first data conversion output unit corresponding to the same exemplary building information model data, the corresponding data learning cost index is analyzed to determine the data learning cost index corresponding to each data conversion output unit;

[0034] According to each of the data learning cost indicators, the corresponding data conversion output unit in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network.

[0035] The embodiment of the present invention further provides a data generation system for a building information model (BIM), comprising:

[0036] A building information analysis neural network optimization module is used to optimize the constructed building information analysis neural network in at least two stages according to the exemplary building information model data included in the exemplary building information model data cluster to form corresponding target building information model data;

[0037] A model data extraction module is used to extract target building information model data;

[0038] The model description data analysis module is used to load the target building information model data to analyze and generate model description data corresponding to the target building information model data using a first building information analysis neural network, and the model description data serves as identification data of the target building information model data.

[0039] In some preferred embodiments, in the data generation system of the building information model BIM, the building information analysis neural network optimization module is specifically used to:

[0040] Extracting an exemplary building information model data cluster, wherein the exemplary building information model data cluster includes exemplary building information model data corresponding to at least two optimization stages;

[0041] Extracting first exemplary building information model data corresponding to the first optimization stage, and optimizing the feature mining subnetwork included in the constructed building information analysis neural network according to the first exemplary building information model data to form a corresponding candidate building information analysis neural network;

[0042] extracting second exemplary building information model data corresponding to a subsequent optimization stage, and optimizing the feature mining subnetwork included in the candidate building information analysis neural network based on the second exemplary building information model data to form a corresponding optimized building information analysis neural network;

[0043] Loading the used building information model data corresponding to the second exemplary building information model data into the feature mining subnetworks included in the candidate building information analysis neural network and the optimized building information analysis neural network, respectively, to output corresponding candidate building information feature mining results and optimized building information feature mining results;

[0044] According to the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data, the data conversion output subnetwork included in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network;

[0045] The extraction of the second exemplary building information model data corresponding to the next optimization stage is executed in a reverse manner, and the feature mining subnetwork included in the candidate building information analysis neural network is optimized based on the second exemplary building information model data. When the optimization is completed, the corresponding candidate building information analysis neural network at the end of the optimization is marked to form the corresponding first building information analysis neural network.

[0046] In some preferred embodiments, in the data generation system of the building information model BIM, the extracting of the exemplary building information model data cluster includes:

[0047] Extracting a plurality of initial exemplary building information model data, each of the initial exemplary building information model data having initial identification data;

[0048] Performing a data aggregation operation on each initial exemplary building information model data corresponding to the same initial identification data to form a building information model data aggregation result corresponding to each initial identification data;

[0049] Analyzing the phase order relationship information corresponding to each initial exemplary building information model data according to the initial identification data;

[0050] Marking the building information model data aggregation results to form corresponding optimization sub-stages, performing data aggregation operations on each optimization sub-stage corresponding to the same stage sorting relationship information to form an optimization stage aggregation result corresponding to each stage sorting relationship information, wherein the optimization stage aggregation result includes each exemplary building information model data corresponding to the same optimization stage;

[0051] According to each of the optimization stage aggregation results, a corresponding exemplary building information model data cluster is formed.

[0052] The embodiment of the present invention provides a data generation method and system for a building information model (BIM), which can optimize the constructed building information analysis neural network in at least two stages according to the exemplary building information model data included in the exemplary building information model data cluster to form corresponding target building information model data; extract the target building information model data; load the target building information model data to analyze and generate model description data corresponding to the target building information model data using the first building information analysis neural network, and the model description data serves as identification data of the target building information model data. Based on the above content, since the target building information model data is formed through at least two stages of optimization, its analysis accuracy is higher, thereby improving the reliability of the identification data of the target building information model data to a certain extent.

[0053] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A structural block diagram of a data generation platform for a building information model (BIM) provided in an embodiment of the present invention.

[0055] Figure 2 A flow chart of the steps of a method for generating data of a building information model (BIM) provided in an embodiment of the present invention.

[0056] Figure 3 A schematic diagram of various modules included in a data generation system for a building information model (BIM) provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0057] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0058] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0059] like Figure 1As shown, an embodiment of the present invention provides a data generation platform for a building information model (BIM). The data generation platform for a building information model (BIM) may include a memory and a processor, and may also include other devices or components.

[0060] In detail, the memory and the processor are electrically connected directly or indirectly to achieve data transmission or interaction. For example, they can be electrically connected to each other through one or more communication buses or signal lines. The memory can store at least one software function module (computer program) that can exist in the form of software or firmware. The processor can be used to execute the executable computer program stored in the memory, thereby realizing the data generation method of the building information model BIM provided by the embodiment of the present invention (as described later).

[0061] It can be selected that, in some embodiments, the memory can be, but not limited to, random access memory (RAM), read only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electric erasable programmable read-only memory (EEPROM), etc. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a system on chip (SoC), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0062] Optionally, in some implementations, the data generation platform of the building information model (BIM) may be a server with data processing capabilities.

[0063] Combination Figure 2 The embodiment of the present invention also provides a data generation method for a building information model (BIM), which can be applied to the data generation platform for the building information model (BIM). The method steps defined in the process related to the data generation method for the building information model (BIM) can be implemented by the data generation platform for the building information model (BIM).

[0064] The following will Figure 2 The specific process shown is explained in detail.

[0065] Step S110, based on the exemplary building information model data included in the exemplary building information model data cluster, the constructed building information analysis neural network is optimized in at least two stages to form corresponding target building information model data.

[0066] In an embodiment of the present invention, the data generation platform of the building information model BIM can optimize the constructed building information analysis neural network in at least two stages based on the exemplary building information model data included in the exemplary building information model data cluster to form corresponding target building information model data (so as to have a reliable identification data generation function).

[0067] Step S120: extracting target building information model data.

[0068] In the embodiment of the present invention, the data generation platform of the building information model (BIM) can extract target building information model data (ie, model data to be processed).

[0069] Step S130: loading the target building information model data to analyze and generate model description data corresponding to the target building information model data using a first building information analysis neural network.

[0070] In an embodiment of the present invention, the data generation platform of the building information model BIM can load the target building information model data to analyze and generate model description data corresponding to the target building information model data using the first building information analysis neural network. The model description data serves as identification data of the target building information model data (the identification data can be used to reflect the label or general description data of the target building information model data).

[0071] Based on the foregoing, since the target building information model data is formed through at least two stages of optimization, its analysis accuracy is higher, thereby improving the reliability of the identification data of the target building information model data to a certain extent, thereby improving the deficiencies of the existing technology.

[0072] Optionally, in some embodiments, the step of optimizing the constructed building information analysis neural network in at least two stages based on the exemplary building information model data included in the exemplary building information model data cluster to form corresponding target building information model data may further include the following achievable sub-steps:

[0073] Extracting an exemplary building information model data cluster, wherein the exemplary building information model data cluster includes exemplary building information model data corresponding to at least two optimization stages;

[0074] Extracting first exemplary building information model data corresponding to a first optimization stage (the first optimization stage may refer to the previous optimization stage of a subsequent optimization stage in the text below, that is, in the process of performing optimization of at least two stages, the previous stage of any two adjacent stages is the first optimization stage, and the subsequent stage is the subsequent optimization stage), and optimizing the feature mining subnetwork included in the constructed building information analysis neural network according to the first exemplary building information model data to form a corresponding candidate building information analysis neural network;

[0075] extracting second exemplary building information model data corresponding to a subsequent optimization stage, and optimizing the feature mining subnetwork included in the candidate building information analysis neural network based on the second exemplary building information model data to form a corresponding optimized building information analysis neural network;

[0076] Loading the used building information model data corresponding to the second exemplary building information model data into the feature mining subnetworks included in the candidate building information analysis neural network and the optimized building information analysis neural network, respectively, to output corresponding candidate building information feature mining results and optimized building information feature mining results;

[0077] According to the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data, the data conversion output subnetwork included in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network;

[0078] The step of reverting to the step of extracting the second exemplary building information model data corresponding to the next optimization stage, and optimizing the feature mining subnetwork included in the candidate building information analysis neural network based on the second exemplary building information model data, and marking the corresponding candidate building information analysis neural network at the end of the optimization to mark and form the corresponding first building information analysis neural network.

[0079] Optionally, in some embodiments, the step of extracting an exemplary building information model data cluster may further include the following achievable sub-steps:

[0080] Extracting a plurality of initial exemplary building information model data, each of the initial exemplary building information model data having initial identification data (i.e., corresponding actual model description data);

[0081] Performing a data aggregation operation on each initial exemplary building information model data corresponding to the same initial identification data to form a building information model data aggregation result corresponding to each initial identification data;

[0082] Analyzing the phase order relationship information (which may be pre-defined or configured) corresponding to each initial exemplary building information model data according to the initial identification data;

[0083] Marking the building information model data aggregation results to form corresponding optimization sub-stages, performing data aggregation operations on each optimization sub-stage corresponding to the same stage sorting relationship information to form an optimization stage aggregation result corresponding to each stage sorting relationship information, wherein the optimization stage aggregation result includes each exemplary building information model data corresponding to the same optimization stage;

[0084] According to each of the optimization stage aggregation results, a corresponding exemplary building information model data cluster is formed.

[0085] Alternatively, in some embodiments, the exemplary building information model data in the exemplary building information model data cluster has optimization identification data, and the building information analysis neural network to be optimized belongs to the constructed building information analysis neural network or the candidate building information analysis neural network. Based on this, optimizing the feature mining subnetwork included in the building information analysis neural network to be optimized may further include the following achievable sub-steps:

[0086] Using the building information analysis neural network to be optimized, the loaded exemplary building information model data is subjected to a building information model analysis operation to analyze and generate analysis identification data corresponding to the loaded exemplary building information model data;

[0087] Based on the identification data distinction information between the optimization identification data and the corresponding analysis identification data corresponding to the loaded exemplary building information model data, the sub-network parameters of the feature mining sub-network included in the building information analysis neural network to be optimized are optimized to complete the optimization of the building information analysis neural network to be optimized.

[0088] It can be selected that, in some embodiments, the target building information analysis neural network belongs to the constructed building information analysis neural network, the candidate building information analysis neural network, the optimized building information analysis neural network or the first building information analysis neural network. Based on this, the feature mining subnetwork included in the target building information analysis neural network is used to perform feature mining operations on the loaded exemplary building information model data (exemplarily, the exemplary building information model data can be first mapped to the feature space to be represented by a corresponding continuous vector, and then processed by a convolution kernel or a filter matrix to mine the key information therein, that is, to obtain the corresponding first building information feature mining result) to output the corresponding first building information feature. The target building information analysis neural network includes a data conversion output subnetwork for performing data conversion and output operations on the first building information feature mining results (exemplarily, the first building information feature mining results can be processed linearly and nonlinearly to obtain the corresponding second building information feature mining results, such as through a fully connected network) to output the corresponding second building information feature mining results (on this basis, the target building information analysis neural network can also include an analysis output unit, which can process the second building information feature mining results to obtain analysis identification data corresponding to the exemplary building information model data, and the analysis output unit can include an incentive mapping output function).

[0089] Optionally, in some embodiments, the feature mining subnetwork includes at least two feature mining units, the data conversion output subnetwork includes at least two data conversion output units, there is a one-to-one correspondence between the at least two feature mining units and the at least two data conversion output units, the first building information feature mining result includes the local first building information feature mining result mined by each feature mining unit, based on this, the data conversion output operation of the second building information feature mining result may include the following achievable sub-steps:

[0090] The local first building information feature mining results mined by each of the feature mining units are loaded into the corresponding data conversion and output units, so as to use the data conversion and output units to convert and output the local second building information feature mining results corresponding to each data conversion and output unit;

[0091] According to each of the local second building information feature mining results, a corresponding second building information feature mining result is determined.

[0092] Optionally, in some embodiments, the feature mining subnetwork includes at least two feature mining units, the data conversion output subnetwork includes at least two data conversion output units, and there is a one-to-one correspondence between at least two of the feature mining units and at least two of the data conversion output units. The candidate building information feature mining results include local candidate building information feature mining results mined by each feature mining unit of the candidate building information analysis neural network, and the optimized building information feature mining results include local optimized building information feature mining results mined by each feature mining unit of the optimized building information analysis neural network. Based on this, the step of optimizing the data conversion output subnetwork included in the optimized building information analysis neural network based on the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data to form a corresponding optimized candidate building information analysis neural network may further include the following achievable sub-steps:

[0093] Respectively loading each local optimized building information feature mining result corresponding to the used building information model data into the corresponding data conversion output unit in the optimized building information analysis neural network to convert and output each corresponding local optimized building information feature conversion result;

[0094] Marking the corresponding feature mining unit and data conversion output unit to mark them as the corresponding first feature mining unit and first data conversion output unit;

[0095] According to the local candidate building information feature mining results mined by the first feature mining unit corresponding to the same exemplary building information model data and the local optimized building information feature conversion results converted by the first data conversion output unit, the corresponding data learning cost index is analyzed to determine the data learning cost index corresponding to each data conversion output unit (exemplarily, the data learning cost index can be calculated by cross entropy error. For example, one feature mining unit corresponds to one data conversion output unit, so that one local candidate building information feature mining result corresponds to one local optimized building information feature conversion result. Based on this, the data learning cost index can be calculated based on the local candidate building information feature mining results output by the first feature mining unit corresponding to the same exemplary building information model data and the local optimized building information feature conversion results output by the first data conversion output unit, so as to obtain the data learning cost indexes corresponding to each data conversion output unit. Among them, the first feature mining unit and the first data conversion output unit refer to the feature mining unit and the data conversion output unit that have a corresponding relationship);

[0096] According to each of the data learning cost indicators, the corresponding data conversion output unit in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network.

[0097] Exemplarily, the optimized building information analysis neural network includes a feature mining unit A and a data conversion output unit A, which have a corresponding relationship, and a data conversion output unit B and a data conversion output unit B, which have a corresponding relationship. Based on this, the exemplary building information model data can be loaded into the feature mining subnetworks of the candidate building information analysis neural network and the optimized building information analysis neural network respectively, the feature mining unit A of the candidate building information analysis neural network outputs a local candidate building information feature mining result X1, the data conversion output unit B of the candidate building information analysis neural network outputs a local candidate building information feature mining result X2, the feature mining unit A of the optimized building information analysis neural network outputs a local optimized building information feature mining result Y1, and the data conversion output unit B of the optimized building information analysis neural network outputs a local optimized building information feature mining result Y2. Afterwards, the local optimized building information feature mining result Y1 can be loaded into the data conversion output unit A of the optimized building information analysis neural network to obtain the first local optimized building information feature conversion result Z1 output by the data conversion output unit A, the data learning cost index W1 is calculated based on the local candidate building information feature mining result X1 and the first local optimized building information feature conversion result Z1, and the parameters of the data conversion output unit A of the optimized building information analysis neural network are adjusted based on the data learning cost index W1. The local optimized building information feature mining result Y2 can be loaded into the data conversion output unit B of the optimized building information analysis neural network to obtain the first local optimized building information feature conversion result Z2, and the data learning cost index W2 is calculated based on the local candidate building information feature mining result X2 and the first local optimized building information feature conversion result Z2, and the parameters of the data conversion output unit B of the optimized building information analysis neural network are adjusted based on the data learning cost index W2. When the parameters of the data conversion output unit are optimized, the above steps can be repeated to calculate new data learning cost indexes W1 and data learning cost indexes W2, and when the updated data learning cost indexes W1 and data learning cost indexes W2 are less than the threshold, the optimized candidate building information analysis neural network can be obtained.

[0098] Combination Figure 3 The embodiment of the present invention also provides a data generation system for a building information model (BIM), which can be applied to the data generation platform of the building information model (BIM). The data generation system for the building information model (BIM) may include the following modules:

[0099] A building information analysis neural network optimization module is used to optimize the constructed building information analysis neural network in at least two stages according to the exemplary building information model data included in the exemplary building information model data cluster to form corresponding target building information model data;

[0100] A model data extraction module is used to extract target building information model data;

[0101] The model description data analysis module is used to load the target building information model data to analyze and generate model description data corresponding to the target building information model data using a first building information analysis neural network, and the model description data serves as identification data of the target building information model data.

[0102] Optionally, in some embodiments, the building information analysis neural network optimization module is specifically used to:

[0103] An exemplary building information model data cluster is extracted, wherein the exemplary building information model data cluster includes exemplary building information model data corresponding to at least two optimization stages; first exemplary building information model data corresponding to the first optimization stage is extracted, and, based on the first exemplary building information model data, a feature mining subnetwork included in the constructed building information analysis neural network is optimized to form a corresponding candidate building information analysis neural network; second exemplary building information model data corresponding to the next optimization stage is extracted, and, based on the second exemplary building information model data, a feature mining subnetwork included in the candidate building information analysis neural network is optimized to form a corresponding optimized building information analysis neural network; used building information model data corresponding to the second exemplary building information model data is loaded into the candidate building information analysis neural network and The feature mining subnetwork included in the optimized building information analysis neural network outputs corresponding candidate building information feature mining results and optimized building information feature mining results; based on the candidate building information feature mining results and optimized building information feature mining results corresponding to the same exemplary building information model data, the data conversion output subnetwork included in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network; the second exemplary building information model data corresponding to the extraction in the next optimization stage is executed in reverse, and, based on the second exemplary building information model data, the feature mining subnetwork included in the candidate building information analysis neural network is optimized; when the optimization is completed, the corresponding candidate building information analysis neural network at the end of the optimization is marked to mark and form a corresponding first building information analysis neural network.

[0104] Optionally, in some embodiments, the extracting of the exemplary building information model data cluster includes: extracting a plurality of initial exemplary building information model data, each of the initial exemplary building information model data having initial identification data; performing a data aggregation operation on each initial exemplary building information model data corresponding to the same initial identification data to form a building information model data aggregation result corresponding to each of the initial identification data; analyzing the stage sorting relationship information corresponding to each of the initial exemplary building information model data based on the initial identification data; marking the building information model data aggregation result to form a corresponding optimization sub-stage, performing a data aggregation operation on each of the optimization sub-stages corresponding to the same stage sorting relationship information to form an optimization stage aggregation result corresponding to each stage sorting relationship information, the optimization stage aggregation result including each exemplary building information model data corresponding to the same optimization stage; and forming a corresponding exemplary building information model data cluster based on each of the optimization stage aggregation results.

[0105] In summary, the data generation method and system of a building information model (BIM) provided by the present invention can optimize the constructed building information analysis neural network in at least two stages according to the exemplary building information model data included in the exemplary building information model data cluster to form corresponding target building information model data; extract the target building information model data; load the target building information model data to analyze and generate model description data corresponding to the target building information model data using the first building information analysis neural network, and the model description data serves as identification data of the target building information model data. Based on the foregoing, since the target building information model data is formed through at least two stages of optimization, its analysis accuracy is higher, thereby improving the reliability of the identification data of the target building information model data to a certain extent.

[0106] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A method for generating data of a building information model (BIM), characterized in that: include: According to the exemplary building information model data included in the exemplary building information model data cluster, the constructed building information analysis neural network is optimized in at least two stages to form a corresponding first building information analysis neural network; Extracting target building information model data; The target building information model data is loaded to analyze and generate model description data corresponding to the target building information model data using a first building information analysis neural network, wherein the model description data serves as identification data of the target building information model data; The step of optimizing the constructed building information analysis neural network in at least two stages according to the exemplary building information model data included in the exemplary building information model data cluster to form a corresponding first building information analysis neural network includes: Extracting an exemplary building information model data cluster, wherein the exemplary building information model data cluster includes exemplary building information model data corresponding to at least two optimization stages; Extracting first exemplary building information model data corresponding to the first optimization stage, and optimizing the feature mining subnetwork included in the constructed building information analysis neural network according to the first exemplary building information model data to form a corresponding candidate building information analysis neural network; extracting second exemplary building information model data corresponding to a subsequent optimization stage, and optimizing the feature mining subnetwork included in the candidate building information analysis neural network based on the second exemplary building information model data to form a corresponding optimized building information analysis neural network; Loading the used building information model data corresponding to the second exemplary building information model data into the feature mining subnetworks included in the candidate building information analysis neural network and the optimized building information analysis neural network, respectively, to output corresponding candidate building information feature mining results and optimized building information feature mining results; According to the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data, the data conversion output subnetwork included in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network; The extraction of the second exemplary building information model data corresponding to the next optimization stage is performed in a reverse manner, and the feature mining subnetwork included in the candidate building information analysis neural network is optimized based on the second exemplary building information model data. When the optimization is completed, the corresponding candidate building information analysis neural network at the end of the optimization is marked to form the corresponding first building information analysis neural network.

2. The method for generating data of a building information model (BIM) according to claim 1, characterized in that: The step of extracting an exemplary building information model data cluster comprises: Extracting a plurality of initial exemplary building information model data, each of the initial exemplary building information model data having initial identification data; Performing a data aggregation operation on each initial exemplary building information model data corresponding to the same initial identification data to form a building information model data aggregation result corresponding to each initial identification data; Analyzing the phase order relationship information corresponding to each initial exemplary building information model data according to the initial identification data; Marking the building information model data aggregation results to form corresponding optimization sub-stages, performing data aggregation operations on each optimization sub-stage corresponding to the same stage sorting relationship information to form an optimization stage aggregation result corresponding to each stage sorting relationship information, wherein the optimization stage aggregation result includes each exemplary building information model data corresponding to the same optimization stage; According to each of the optimization stage aggregation results, a corresponding exemplary building information model data cluster is formed.

3. The method for generating data of a building information model (BIM) according to claim 1, characterized in that: The exemplary building information model data in the exemplary building information model data cluster has optimization identification data, the building information analysis neural network to be optimized belongs to the constructed building information analysis neural network or the candidate building information analysis neural network, and optimizing the feature mining subnetwork included in the building information analysis neural network to be optimized comprises the following steps: Using the building information analysis neural network to be optimized, the loaded exemplary building information model data is subjected to a building information model analysis operation to analyze and generate analysis identification data corresponding to the loaded exemplary building information model data; Based on the identification data distinction information between the optimization identification data and the corresponding analysis identification data corresponding to the loaded exemplary building information model data, the sub-network parameters of the feature mining sub-network included in the building information analysis neural network to be optimized are optimized, so as to complete the optimization of the building information analysis neural network to be optimized.

4. The method for generating data of a building information model (BIM) according to claim 1, characterized in that: The target building information analysis neural network belongs to the constructed building information analysis neural network, the candidate building information analysis neural network, the optimized building information analysis neural network or the first building information analysis neural network. The feature mining subnetwork included in the target building information analysis neural network is used to perform feature mining operations on the loaded exemplary building information model data to output the corresponding first building information feature mining results. The data conversion output subnetwork included in the target building information analysis neural network is used to perform data conversion and output operations on the first building information feature mining results to output the corresponding second building information feature mining results.

5. The method for generating data of a building information model (BIM) according to claim 4, characterized in that: The feature mining subnetwork includes at least two feature mining units, the data conversion output subnetwork includes at least two data conversion output units, there is a one-to-one correspondence between the at least two feature mining units and the at least two data conversion output units, the first building information feature mining result includes a local first building information feature mining result mined by each feature mining unit, and the data conversion output operation of the second building information feature mining result includes: The local first building information feature mining results mined by each of the feature mining units are loaded into the corresponding data conversion and output units, so as to use the data conversion and output units to convert and output the local second building information feature mining results corresponding to each data conversion and output unit; According to each of the local second building information feature mining results, a corresponding second building information feature mining result is determined.

6. The method for generating data of a building information model (BIM) according to claim 1, characterized in that: The feature mining subnetwork includes at least two feature mining units, the data conversion output subnetwork includes at least two data conversion output units, there is a one-to-one correspondence between the at least two feature mining units and the at least two data conversion output units, the candidate building information feature mining results include the local candidate building information feature mining results mined by each feature mining unit of the candidate building information analysis neural network, and the optimized building information feature mining results include the local optimized building information feature mining results mined by each feature mining unit of the optimized building information analysis neural network; The step of optimizing the data conversion output subnetwork included in the optimized building information analysis neural network based on the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data to form a corresponding optimized candidate building information analysis neural network comprises: Respectively loading each local optimized building information feature mining result corresponding to the used building information model data into the corresponding data conversion output unit in the optimized building information analysis neural network to convert and output each corresponding local optimized building information feature conversion result; Marking the corresponding feature mining unit and data conversion output unit to mark them as the corresponding first feature mining unit and first data conversion output unit; According to the local candidate building information feature mining results mined by the first feature mining unit and the local optimized building information feature conversion results converted by the first data conversion output unit corresponding to the same exemplary building information model data, the corresponding data learning cost index is analyzed to determine the data learning cost index corresponding to each data conversion output unit; According to each of the data learning cost indicators, the corresponding data conversion output unit in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network.

7. A data generation system for a building information model (BIM), characterized in that: include: A building information analysis neural network optimization module is used to optimize the constructed building information analysis neural network in at least two stages according to the exemplary building information model data included in the exemplary building information model data cluster to form a corresponding first building information analysis neural network; A model data extraction module is used to extract target building information model data; A model description data analysis module is used to load the target building information model data, and use a first building information analysis neural network to analyze and generate model description data corresponding to the target building information model data, wherein the model description data serves as identification data of the target building information model data; Wherein, the building information analysis neural network optimization module is specifically used for: Extracting an exemplary building information model data cluster, wherein the exemplary building information model data cluster includes exemplary building information model data corresponding to at least two optimization stages; Extracting first exemplary building information model data corresponding to the first optimization stage, and optimizing the feature mining subnetwork included in the constructed building information analysis neural network according to the first exemplary building information model data to form a corresponding candidate building information analysis neural network; extracting second exemplary building information model data corresponding to a subsequent optimization stage, and optimizing the feature mining subnetwork included in the candidate building information analysis neural network based on the second exemplary building information model data to form a corresponding optimized building information analysis neural network; Loading the used building information model data corresponding to the second exemplary building information model data into the feature mining subnetworks included in the candidate building information analysis neural network and the optimized building information analysis neural network, respectively, to output corresponding candidate building information feature mining results and optimized building information feature mining results; According to the candidate building information feature mining results and the optimized building information feature mining results corresponding to the same exemplary building information model data, the data conversion output subnetwork included in the optimized building information analysis neural network is optimized to form a corresponding optimized candidate building information analysis neural network; The extraction of the second exemplary building information model data corresponding to the next optimization stage is performed in a reverse manner, and the feature mining subnetwork included in the candidate building information analysis neural network is optimized based on the second exemplary building information model data. When the optimization is completed, the corresponding candidate building information analysis neural network at the end of the optimization is marked to form the corresponding first building information analysis neural network.

8. The data generation system of the building information model (BIM) according to claim 7, characterized in that: The extracting of an exemplary building information model data cluster includes: Extracting a plurality of initial exemplary building information model data, each of the initial exemplary building information model data having initial identification data; Performing a data aggregation operation on each initial exemplary building information model data corresponding to the same initial identification data to form a building information model data aggregation result corresponding to each initial identification data; Analyzing the phase order relationship information corresponding to each initial exemplary building information model data according to the initial identification data; Marking the building information model data aggregation results to form corresponding optimization sub-stages, performing data aggregation operations on each optimization sub-stage corresponding to the same stage sorting relationship information to form an optimization stage aggregation result corresponding to each stage sorting relationship information, wherein the optimization stage aggregation result includes each exemplary building information model data corresponding to the same optimization stage; According to each of the optimization stage aggregation results, a corresponding exemplary building information model data cluster is formed.

Images