CN114840900A - Derivative BIM component automatic generation method based on i-GBDT technology - Google Patents
Derivative BIM component automatic generation method based on i-GBDT technology Download PDFInfo
- Publication number
- CN114840900A CN114840900A CN202210540316.4A CN202210540316A CN114840900A CN 114840900 A CN114840900 A CN 114840900A CN 202210540316 A CN202210540316 A CN 202210540316A CN 114840900 A CN114840900 A CN 114840900A
- Authority
- CN
- China
- Prior art keywords
- construction drawing
- component
- layer
- bim
- data set
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/13—Architectural design, e.g. computer-aided architectural design [CAAD] related to design of buildings, bridges, landscapes, production plants or roads
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Geometry (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Computer Hardware Design (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Health & Medical Sciences (AREA)
- Architecture (AREA)
- Civil Engineering (AREA)
- Structural Engineering (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a derivative BIM component automatic generation method based on an i-GBDT technology, which comprises the following steps: establishing a basic data set based on a standard construction drawing; identifying and extracting image features of a proposed construction drawing through a trained convolutional neural network, performing feature processing by matching with a ReLU nonlinear activation function, and outputting component image information in the construction drawing; establishing an initial learner through a gradient decision tree algorithm, calculating a gradient value of a loss function, performing continuous iteration, and performing linear optimization to calculate the optimal learning rate so as to obtain an optimal classifier; taking the component size information in the basic data set as an attribute parameter, entering an optimal classifier to perform predictive analysis on the identified construction drawing component image information, and outputting the component image information with high fitting degree; taking the predicted component image information as output data, and performing parameter matching with a standard construction drawing in a basic data set to generate IFC format information; and importing the IFC format information into BIM modeling software, and automatically generating a corresponding BIM model.
Description
Technical Field
The invention relates to the technical field of BIM modeling, in particular to an automatic generation method of a derivative BIM component based on an i-GBDT technology.
Background
As a representative example, the BIM technology is widely used in the stages of planning, design, construction, and operation in the engineering construction field, and exerts energy efficiency in a plurality of industrial fields such as hydropower and transportation. The BIM model is used as a bottom information carrier for implementing the BIM technology, and the modeling efficiency of the BIM model directly influences the popularization and application depth of the BIM technology.
At present, the matching and compatibility of the algorithm, the function and the design of the conventional mainstream BIM modeling software in China and construction drawings are poor, and the problem of influence of drawing precision errors caused by poor drawing standardization is further amplified, so that a more effective automatic modeling method is needed to be combined with an artificial intelligence recognition algorithm, the influence of the drawing precision on the modeling software is reduced, the 'die rollover' in the true sense can be realized, and a new opportunity is brought to the BIM technical development.
Therefore, the invention provides an automatic generation method of a derivative BIM component based on an i-GBDT (internet-Gradient Boosting Decision Tree) technology.
Disclosure of Invention
In order to solve the problems, the invention provides an automatic generation method of a derivative BIM component based on an i-GBDT technology.
The invention provides the following technical scheme.
An automatic generation method of a derivative BIM component based on an i-GBDT technology comprises the following steps:
establishing a basic data set based on a standard construction drawing;
identifying and extracting image characteristics of a proposed construction drawing, and outputting component image information in the construction drawing;
establishing an initial learning device through a gradient decision tree algorithm, and obtaining the optimal learning rate through linear optimization so as to obtain an optimal classifier;
taking the component size information in the basic data set as an attribute parameter, entering an optimal classifier to perform predictive analysis on the component image information in the identified construction drawing, and outputting the component image information with high fitting degree;
taking the predicted component image information as output data, and performing parameter matching with a standard construction drawing in a basic data set to generate IFC format information;
and importing the IFC format information into BIM modeling software, and automatically generating a corresponding BIM model.
Preferably, the acquisition of the construction drawing member image data includes the steps of:
establishing a basic data set, and inputting a building standard construction drawing;
constructing a VGG16 network model, and training the VGG16 network model according to the basic data set to obtain a trained convolutional layer;
extracting image characteristics of the construction drawing by using convolution kernels in the convolution layer, and performing characteristic processing by matching with a ReLU nonlinear activation function;
the image features output by the activation layer enter the maximum pooling layer, and the image is subjected to multiple times of feature map size reduction processing through a plurality of convolution checks in the pooling layer;
and integrating and outputting the convolution layer and the pooling layer through the full-connection layer, and then classifying the convolution layer and the pooling layer as a classifier to obtain complete standard component image data.
Preferably, the basic data set is a MySQL database created by Navicat, and a standard construction drawing is recorded.
Preferably, the image feature of the construction drawing includes: a numbering layer, a marking layer, a sideline layer, a filling layer and the like.
Preferably, the obtaining of the optimal classifier specifically includes the following steps:
establishing an initial learner;
selecting a sample A from the image characteristics identified by the convolutional neural network, naming the sample A as a characteristic A, selecting a segmentation point B for the characteristic A, and creating a decision tree by creating nodes, wherein a basic function formula is as follows:
in which T (x; theta) m ) Represents a decision tree, θ m As a decision tree parameter, M is the training times;
calculating a loss function gradient value for each sample, setting the iteration number as m, and fitting residual errors;
and taking the obtained residual error as a new true value of the sample, and continuously iterating for m times, linearly optimizing and calculating the optimal learning rate rho to obtain the optimal classifier.
Preferably, the obtaining of the attribute parameters includes:
and normalizing the size of the component and the large sample graph of the component in the standard construction graph of the basic data set as attribute parameters.
The invention has the beneficial effects that:
the invention provides an automatic generation method of a derivative BIM component based on an i-GBDT technology, which adopts a nonlinear activation function to improve the extraction efficiency of a convolutional layer. The invention improves the gradient decision tree algorithm, carries out image recognition and characteristic value extraction on the proposed construction drawing, automatically generates the IFC standard format and improves the efficiency of drawing error correction and recognition.
Drawings
FIG. 1 is a flow chart of a method for automatically generating a derivative BIM component based on i-GBDT technology;
FIG. 2 is a CAD drawing of a CAD of a component of an embodiment of the present invention;
FIG. 3 illustrates an embodiment of a convolutional neural network identifying image information;
FIG. 4 is a schematic diagram of i-GBDT fitting residual analysis according to an embodiment of the present invention;
FIG. 5 is a partial IFC information automatically generated by a component of an embodiment of the present invention;
FIG. 6 is a three-dimensional model of a BIM component automatically generated according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
The invention discloses a derivative BIM component automatic generation method based on an i-GBDT technology. As shown in fig. 1-6:
the invention uses BIM structural component automatic generation example based on Revit:
a certain building structure construction drawing is taken as a rollover object, a data set is established, a convolutional neural network identifies an image, a decision tree is established by i-GBDT, and finally image data are exported. The specific steps are implemented by referring to the flow shown in fig. 1, and specifically comprise the following steps:
the method comprises the following steps: and (3) creating a MySQL database by adopting Navicat, inputting a beam leveling method construction drawing as a data set identified by the beam leveling method. And inputting a VGG16 source code for training based on a Python3-TensorFlow open source machine learning platform to obtain a trained convolutional layer.
Step two: and identifying and extracting characteristics of primitive information such as a numbered layer, a labeled layer, a side line layer, a filling layer and the like in the beam leveling construction drawing through a convolutional neural network. By extracting image features such as color, shape and image topological structure of image blocks and image edges in the drawing, the robustness of the model is improved through the ReLU activation function, and the disappearance of gradients is slowed down. After a series of image matrix processing such as output from the active layer to the full connection layer, standard component image information in the data set can be obtained, as shown in fig. 3.
Step three: establishing an initial learner:
selecting a sample A from the image data identified by the convolutional neural network, naming the sample A as a feature A, selecting a segmentation point B for the feature A, and if the feature value of the sample A is smaller than B, classifying the sample A into one class, otherwise classifying the sample A into another class. A node of the decision tree is constructed through the scheme, and then the decision tree 1 is created, wherein the basic function formula is as follows:
wherein T (x; theta) m ) Represents a decision tree, θ m For decision tree parameter M as training times, the residual is calculated for M1, 2,3 … with each sample i 1,2,3 … N:
will (x) i ，r im ) As a basis for the next decision tree classification. In the second prediction analysis, the linear analysis conditions are fitted to the residuals, and the prediction range is narrowed down to obtain new residuals again, as shown in fig. 4. The analogy is carried out in each prediction analysis, the optimal classifier is obtained through repeated iteration and a large amount of training, and the basic formula is as follows:
f m (x)＝f m-1 (x)+ρy(x i ,a m )
step four: and distinguishing the component size and reinforcement information of the middle beam of the beam-leveling construction drawing, and entering an optimal classifier as an attribute parameter to perform predictive analysis on the image information obtained by convolutional neural network recognition until component image information with higher fitting degree with the beam component in the beam-leveling construction drawing is finally obtained.
Step five: the image information is used as final output data, and parameter matching is carried out on the image information and information such as beam size, reinforcing bars and the like in a beam leveling construction drawing, so that an IFC standard format of the member is generated, as shown in FIG. 5. The IFC standard format is imported into Revit and the software automatically generates a three-dimensional component of the beam as shown in fig. 6.
The present invention is not limited to the above preferred embodiments, and any modifications, equivalent substitutions and improvements made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (6)
1. An automatic generation method of a derivative BIM component based on an i-GBDT technology is characterized by comprising the following steps:
establishing a basic data set based on a standard construction drawing;
identifying and extracting image characteristics of a proposed construction drawing, and outputting component image information in the construction drawing;
establishing an initial learner through a gradient decision tree algorithm, and obtaining an optimal learning rate through linear optimization so as to obtain an optimal classifier;
taking the component size information in the basic data set as an attribute parameter, entering an optimal classifier to perform predictive analysis on the component image information in the identified construction drawing, and outputting the component image information with high fitting degree;
taking the predicted component image information as output data, and performing parameter matching with a standard construction drawing in a basic data set to generate IFC format information;
and importing the IFC format information into BIM modeling software, and automatically generating a corresponding BIM model.
2. The method for automatically generating the derivative BIM component based on the i-GBDT technology as claimed in claim 1, wherein the obtaining of the construction drawing component image data comprises the following steps:
establishing a basic data set, and inputting a building standard construction drawing;
constructing a VGG16 network model, and training the VGG16 network model according to the basic data set to obtain a trained convolutional layer;
extracting image characteristics of the construction drawing by using convolution kernels in the convolution layer, and performing characteristic processing by matching with a ReLU nonlinear activation function;
the image features output by the activation layer enter the maximum pooling layer, and the image is subjected to multiple times of feature map size reduction processing through a plurality of convolution checks in the pooling layer;
and integrating and outputting the convolution layer and the pooling layer through the full-connection layer, and then classifying the convolution layer and the pooling layer as a classifier to obtain complete standard component image data.
3. The method for automatically generating the derived BIM component based on the i-GBDT technology as claimed in claim 2, wherein the basic data set is a MySQL database created by Navicat, and standard construction drawings are entered.
4. The method for automatically generating the derivative BIM component based on the i-GBDT technology as claimed in claim 1, wherein the image features of the construction drawing comprise: the map layer comprises a numbering map layer, a marking map layer, a side line map layer and a filling map layer.
5. The method for automatically generating the derived BIM component based on the i-GBDT technology as claimed in claim 4, wherein the obtaining of the optimal classifier specifically comprises the following steps:
establishing an initial learner;
selecting a sample A from the image characteristics identified by the convolutional neural network, naming the sample A as a characteristic A, selecting a segmentation point B for the characteristic A, and creating a decision tree by creating nodes, wherein a basic function formula is as follows:
in which T (x; theta) m ) Represents a decision tree, θ m As a decision tree parameter, M is the training times;
calculating a loss function gradient value for each sample, setting the iteration number as m, and fitting residual errors;
and taking the obtained residual error as a new true value of the sample, continuously iterating for m times, and linearly optimizing and calculating the optimal learning rate rho to obtain the optimal classifier.
6. The method for automatically generating the derived BIM component based on i-GBDT technology as claimed in claim 1, wherein the obtaining of the attribute parameters comprises:
and normalizing the size of the component and the large sample graph of the component in the standard construction graph of the basic data set as attribute parameters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210540316.4A CN114840900B (en) | 2022-05-18 | 2022-05-18 | Automatic generation method of derivative BIM (building information modeling) component based on i-GBDT (generic block transformation) technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210540316.4A CN114840900B (en) | 2022-05-18 | 2022-05-18 | Automatic generation method of derivative BIM (building information modeling) component based on i-GBDT (generic block transformation) technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114840900A true CN114840900A (en) | 2022-08-02 |
| CN114840900B CN114840900B (en) | 2023-06-20 |
Family
ID=82571053
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210540316.4A Active CN114840900B (en) | 2022-05-18 | 2022-05-18 | Automatic generation method of derivative BIM (building information modeling) component based on i-GBDT (generic block transformation) technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114840900B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116311336A (en) * | 2023-03-20 | 2023-06-23 | 华联世纪工程咨询股份有限公司 | Automatic beam information identification and acquisition method |
| CN117196418A (en) * | 2023-11-08 | 2023-12-08 | 江西师范大学 | Reading teaching quality assessment method and system based on artificial intelligence |
| CN117272819A (en) * | 2023-09-28 | 2023-12-22 | 苏州苏高新数字科技有限公司 | Data processing conversion method for rapidly generating construction diagram |
| CN117494356A (en) * | 2023-10-31 | 2024-02-02 | 成都建工第九建筑工程有限公司 | Assembled construction method based on BIM technology |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108269212A (en) * | 2016-12-30 | 2018-07-10 | 香港科技大学深圳研究院 | The management method and device of underground communal facility based on BIM and GIS |
| CN108921893A (en) * | 2018-04-24 | 2018-11-30 | 华南理工大学 | A kind of image cloud computing method and system based on online deep learning SLAM |
| CN109684714A (en) * | 2018-12-21 | 2019-04-26 | 杭州睿兴栋宇建筑科技有限公司 | A kind of architectural design method based on machine learning and BIM technology |
| CN109711099A (en) * | 2019-01-23 | 2019-05-03 | 河南省交通规划设计研究院股份有限公司 | A kind of BIM automatic modeling system based on the study of image recognition machine |
| CN111476283A (en) * | 2020-03-31 | 2020-07-31 | 上海海事大学 | Glaucoma fundus image identification method based on transfer learning |
| CN111914612A (en) * | 2020-05-21 | 2020-11-10 | 淮阴工学院 | Construction graph primitive self-adaptive identification method based on improved convolutional neural network |
| CN112613097A (en) * | 2020-12-15 | 2021-04-06 | 中铁二十四局集团江苏工程有限公司 | BIM rapid modeling method based on computer vision |
| CN112634340A (en) * | 2020-12-24 | 2021-04-09 | 深圳大学 | Method, device, equipment and medium for determining BIM (building information modeling) model based on point cloud data |
| US10978056B1 (en) * | 2018-04-20 | 2021-04-13 | Facebook, Inc. | Grammaticality classification for natural language generation in assistant systems |
| CN112784346A (en) * | 2021-02-07 | 2021-05-11 | 殿汇空间(上海)信息科技有限公司 | Building structure autonomous design method, system, terminal and storage medium |
| CN114329745A (en) * | 2022-03-03 | 2022-04-12 | 中国铁路设计集团有限公司 | BIM (building information modeling) model automatic coding method based on decision tree analysis |
-
2022
- 2022-05-18 CN CN202210540316.4A patent/CN114840900B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108269212A (en) * | 2016-12-30 | 2018-07-10 | 香港科技大学深圳研究院 | The management method and device of underground communal facility based on BIM and GIS |
| US10978056B1 (en) * | 2018-04-20 | 2021-04-13 | Facebook, Inc. | Grammaticality classification for natural language generation in assistant systems |
| CN108921893A (en) * | 2018-04-24 | 2018-11-30 | 华南理工大学 | A kind of image cloud computing method and system based on online deep learning SLAM |
| CN109684714A (en) * | 2018-12-21 | 2019-04-26 | 杭州睿兴栋宇建筑科技有限公司 | A kind of architectural design method based on machine learning and BIM technology |
| CN109711099A (en) * | 2019-01-23 | 2019-05-03 | 河南省交通规划设计研究院股份有限公司 | A kind of BIM automatic modeling system based on the study of image recognition machine |
| CN111476283A (en) * | 2020-03-31 | 2020-07-31 | 上海海事大学 | Glaucoma fundus image identification method based on transfer learning |
| CN111914612A (en) * | 2020-05-21 | 2020-11-10 | 淮阴工学院 | Construction graph primitive self-adaptive identification method based on improved convolutional neural network |
| CN112613097A (en) * | 2020-12-15 | 2021-04-06 | 中铁二十四局集团江苏工程有限公司 | BIM rapid modeling method based on computer vision |
| CN112634340A (en) * | 2020-12-24 | 2021-04-09 | 深圳大学 | Method, device, equipment and medium for determining BIM (building information modeling) model based on point cloud data |
| CN112784346A (en) * | 2021-02-07 | 2021-05-11 | 殿汇空间(上海)信息科技有限公司 | Building structure autonomous design method, system, terminal and storage medium |
| CN114329745A (en) * | 2022-03-03 | 2022-04-12 | 中国铁路设计集团有限公司 | BIM (building information modeling) model automatic coding method based on decision tree analysis |
Non-Patent Citations (1)
| Title |
|---|
| 肖宏宇: "面向地铁工程设计过程优化的BIM协同技术研究", 《中国优秀硕士学位论文全文数据库》, no. 5, pages 033 - 341 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116311336A (en) * | 2023-03-20 | 2023-06-23 | 华联世纪工程咨询股份有限公司 | Automatic beam information identification and acquisition method |
| CN117272819A (en) * | 2023-09-28 | 2023-12-22 | 苏州苏高新数字科技有限公司 | Data processing conversion method for rapidly generating construction diagram |
| CN117494356A (en) * | 2023-10-31 | 2024-02-02 | 成都建工第九建筑工程有限公司 | Assembled construction method based on BIM technology |
| CN117196418A (en) * | 2023-11-08 | 2023-12-08 | 江西师范大学 | Reading teaching quality assessment method and system based on artificial intelligence |
| CN117196418B (en) * | 2023-11-08 | 2024-02-02 | 江西师范大学 | Reading teaching quality assessment method and system based on artificial intelligence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114840900B (en) | 2023-06-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114840900B (en) | Automatic generation method of derivative BIM (building information modeling) component based on i-GBDT (generic block transformation) technology | |
| US11113431B2 (en) | MBD-based three-dimensional process designing method and platform for typical automobile machined part | |
| CN109829399A (en) | A kind of vehicle mounted road scene point cloud automatic classification method based on deep learning | |
| CN110084817B (en) | Digital elevation model production method based on deep learning | |
| CN111899353A (en) | Three-dimensional scanning point cloud hole filling method based on generation countermeasure network | |
| CN110992473A (en) | Tree branch modeling method and system based on vehicle-mounted laser scanning point cloud | |
| CN115293751A (en) | Method, system and equipment for processing BIM model data of rail transit | |
| CN113312369A (en) | Multi-scale map database cascading updating method based on unique identification code | |
| CN111597943B (en) | Table structure identification method based on graph neural network | |
| CN106651137A (en) | Project quantity coding and standardization method and system | |
| CN116561327B (en) | Government affair data management method based on clustering algorithm | |
| CN116186864B (en) | Deep foundation pit model rapid modeling method and system based on BIM technology | |
| CN117389542A (en) | Industrial software construction method and device, electronic equipment and storage medium | |
| CN112837420B (en) | Shape complement method and system for terracotta soldiers and horses point cloud based on multi-scale and folding structure | |
| CN111260783B (en) | Ore body three-dimensional automatic modeling method based on K neighbor and Poisson curved surface | |
| CN116629348B (en) | Intelligent workshop data acquisition and analysis method and device and computer equipment | |
| CN112860903B (en) | Remote supervision relation extraction method integrated with constraint information | |
| CN113255581B (en) | Weak supervision deep learning water body extraction method and device, computer equipment and medium | |
| CN112801133A (en) | Spectrum identification and classification method based on keras model | |
| CN115983204A (en) | Method and system for determining industrial process model | |
| CN116883742A (en) | Laser radar environment sensing method based on contrast learning | |
| CN117541772A (en) | YOLOv 3-based straw burning flame identification and detection method | |
| CN117933030A (en) | Finite element calculation model modeling method, device, equipment and medium for power transmission tower | |
| CN115840829A (en) | Automatic retrieval method and device for difference part based on three-dimensional point cloud and BIM | |
| CN118038032A (en) | Point cloud semantic segmentation model based on super point embedding and clustering and training method thereof |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |