CN116935010A - Method, device and equipment for marking inner and outer walls and readable storage medium - Google Patents
Method, device and equipment for marking inner and outer walls and readable storage medium Download PDFInfo
- Publication number
- CN116935010A CN116935010A CN202210330345.8A CN202210330345A CN116935010A CN 116935010 A CN116935010 A CN 116935010A CN 202210330345 A CN202210330345 A CN 202210330345A CN 116935010 A CN116935010 A CN 116935010A
- Authority
- CN
- China
- Prior art keywords
- wall
- primitive
- primitives
- identified
- closed area
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T19/00—Manipulating 3D models or images for computer graphics
- G06T19/20—Editing of 3D images, e.g. changing shapes or colours, aligning objects or positioning parts
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Geometry (AREA)
- Computer Hardware Design (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Architecture (AREA)
- Civil Engineering (AREA)
- Computer Graphics (AREA)
- Structural Engineering (AREA)
- Computational Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Pure & Applied Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Processing Or Creating Images (AREA)
Abstract
The invention discloses a method, a device, equipment and a readable storage medium for marking an inner wall and an outer wall, wherein the method comprises the following steps: acquiring a target building three-dimensional model, and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model; determining the maximum closed area which can be formed by all wall primitives in the set to be identified, and acquiring the outer contour line of the maximum closed area; inputting the set to be identified and the outer contour line into a preset feature extraction algorithm to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line; the position characteristic information of all the wall primitives and the outer contour line are input into a preset classification algorithm, so that all the wall primitives are divided into an inner wall set and an outer wall set based on the outer contour line; marking each wall primitive in the inner wall set as an inner wall, and marking each wall primitive in the outer wall set as an outer wall; and the automatic labeling of the inner wall and the outer wall is realized.
Description
Technical Field
The present invention relates to the field of computer aided design, and in particular, to a method, an apparatus, a device and a readable storage medium for labeling an inner wall and an outer wall.
Background
When a user uses building modeling software to create a building three-dimensional model, the situation that wall components such as a shear wall, a masonry wall, a heat-insulating wall or a preset wall are drawn often exists, and the properties of the inner wall and the outer wall of the wall components need to be marked; the accuracy of the labeling of the inner wall and the outer wall is very important, and the calculation of the related civil engineering quantity can be directly influenced, for example: the engineering amount of the inner and outer walls needs to be distinguished when calculating the scaffold area. In the prior art, most building modeling software does not have a function module for automatically marking the inner wall and the outer wall, and a user is required to manually mark, so that a large amount of human resources are wasted; although part of building modeling software is provided with functional modules for automatically marking the inner wall and the outer wall, the functional modules can not accurately identify the inner wall and the outer wall especially for the condition of three-wall intersection and four-wall intersection because the position relation between wall members in actual engineering is very complex, and users are required to manually adjust the marking of the inner wall and the outer wall, so that the use efficiency of the users is seriously affected. Therefore, how to efficiently and accurately label the inner wall and the outer wall of the wall body member in the three-dimensional building model is a technical problem to be solved by the technicians in the field.
Disclosure of Invention
The invention aims to provide a method, a device, equipment and a readable storage medium for marking inner and outer walls, which can automatically mark the inner and outer walls of wall body members in a three-dimensional building model, and has the advantages of good marking accuracy and higher marking efficiency.
According to one aspect of the present invention, there is provided a method of labelling an interior and exterior wall, the method comprising:
acquiring a target building three-dimensional model, and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model;
determining the maximum closed area which can be formed by all wall primitives in the set to be identified, and acquiring the outer contour line of the maximum closed area;
inputting the set to be identified and the outer contour line into a preset feature extraction algorithm to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line;
the position characteristic information of all the wall primitives and the outer contour line are input into a preset classification algorithm, so that all the wall primitives are divided into an inner wall set and an outer wall set based on the outer contour line;
and marking each wall primitive in the inner wall set as an inner wall, and marking each wall primitive in the outer wall set as an outer wall.
Optionally, the determining the maximum enclosed area that can be formed by all the wall primitives in the set to be identified includes:
selecting a target wall primitive from the set to be identified, determining all closed areas formed by taking the target wall primitive as a starting point according to the relative position information among the wall primitives in the set to be identified, and taking the closed area with the largest area in the formed closed areas as a representative closed area corresponding to the target wall primitive;
and respectively calculating the area of the representative closed area of each wall primitive in the set to be identified, and taking the representative closed area with the largest area as the maximum closed area.
Optionally, before the obtaining the three-dimensional model of the target building and forming the set to be identified by all the wall primitives of the target floor in the three-dimensional model of the target building, the method further includes:
acquiring a training sample set; the training sample set comprises a plurality of training samples, and each training sample represents all wall primitives of a floor in a three-dimensional building model;
determining a maximum sample closed area which can be formed by all wall primitives in the training sample, and acquiring a sample outline of the maximum sample closed area;
Inputting the training sample and the sample outline into a preset feature extraction algorithm to calculate sample position feature information of each wall primitive in the training sample relative to the sample outline;
sample position characteristic information of all wall primitives in the training sample is input into a preset clustering algorithm, so that all sample position characteristic information is clustered into an inner wall training set, an outer wall training set and an unknown wall training set;
and calculating the common characteristic information of the inner wall according to the position characteristic information of all samples in the inner wall training set, and calculating the common characteristic information of the outer wall according to the position characteristic information of all samples in the outer wall training set.
Optionally, the labeling each wall primitive in the inner wall set as an inner wall and labeling each wall primitive in the outer wall set as an outer wall includes:
sequentially calculating a first similarity value of the position characteristic information and the inner wall commonality characteristic information of each wall primitive in the inner wall set, judging whether the first similarity value is larger than a first preset threshold value, if so, marking the corresponding wall primitive as an inner wall, and if not, marking the corresponding wall primitive as an unknown wall;
And sequentially calculating a second similarity value of the position characteristic information of each wall primitive in the outer wall set and the outer wall commonality characteristic information, judging whether the second similarity value is larger than a second preset threshold value, if so, marking the corresponding wall primitive as an outer wall, and if not, marking the corresponding wall primitive as an unknown wall.
Optionally, after labeling each wall primitive in the set of inner walls as an inner wall and labeling each wall primitive in the set of outer walls as an outer wall, the method further comprises:
and sending the wall graphic elements marked as the unknown walls to a designated terminal so as to manually mark the inner walls and the outer walls by a user.
Optionally, the feature extraction algorithm is a scale invariant feature transform SIFT algorithm, the classification algorithm is a support vector machine SVM algorithm, and the clustering algorithm is a K-means clustering algorithm.
In order to achieve the above purpose, the invention also provides a device for marking the inner and outer walls, which comprises the following components:
the acquisition module is used for acquiring a target building three-dimensional model and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model;
The determining module is used for determining the maximum closed area which can be formed by all wall primitives in the set to be identified and acquiring the outer contour line of the maximum closed area;
the extraction module is used for inputting the set to be identified and the outer contour line into a preset feature extraction algorithm so as to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line;
the classification module is used for inputting the position characteristic information of all the wall primitives and the outer contour line into a preset classification algorithm so as to divide all the wall primitives into an inner wall set and an outer wall set based on the outer contour line;
and the labeling module is used for labeling each wall primitive in the inner wall set as an inner wall and labeling each wall primitive in the outer wall set as an outer wall.
Optionally, the determining module is configured to:
selecting a target wall primitive from the set to be identified, determining all closed areas formed by taking the target wall primitive as a starting point according to the relative position information among the wall primitives in the set to be identified, and taking the closed area with the largest area in the formed closed areas as a representative closed area corresponding to the target wall primitive;
And respectively calculating the area of the representative closed area of each wall primitive in the set to be identified, and taking the representative closed area with the largest area as the maximum closed area.
In order to achieve the above object, the present invention further provides a computer device, which specifically includes: the system comprises a memory, a processor and a computer program stored in the memory and capable of running on the processor, wherein the processor realizes the steps of the method for marking the inner wall and the outer wall when executing the computer program.
To achieve the above object, the present invention further provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method for labeling an interior and exterior wall described above.
The method, the device, the equipment and the readable storage medium for marking the inner wall and the outer wall provided by the invention are characterized in that firstly, the closed area with the largest area formed by all wall primitives in a target floor is determined by traversing all wall primitives in the target floor, the closed area is used as a reference for judging the inner wall and the outer wall, then, each wall primitive in the target floor is traversed, and whether the currently traversed wall primitive is the inner wall or the outer wall is determined according to the position relation between the currently traversed wall primitive and the closed area, so that the inner wall and the outer wall are marked. The method can automatically label the inner wall and the outer wall of the wall body component in the building three-dimensional model, and has the advantages of good labeling accuracy and higher efficiency. In addition, the SIFT algorithm, the K-means algorithm and the SVM algorithm are applied to the scene of intelligent identification of the inner and outer walls, and the model capable of realizing intelligent labeling of the inner and outer walls is obtained through a large amount of training, so that a large amount of human resources are not consumed any more, and the efficiency of labeling the inner and outer walls is improved.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to designate like parts throughout the figures. In the drawings:
FIG. 1 is a schematic flow chart of an alternative method for labeling an interior and exterior wall according to the first embodiment;
FIG. 2 is a schematic flow chart of training a model for labeling an inner wall and an outer wall according to the first embodiment;
FIG. 3 is a schematic view of an alternative composition of the device for labeling an inner and outer wall according to the second embodiment;
fig. 4 is a schematic diagram of an alternative hardware architecture of a computer device according to the third embodiment.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Example 1
The embodiment of the invention provides a method for marking an inner wall and an outer wall, as shown in fig. 1, which specifically comprises the following steps:
step S101: and obtaining a target building three-dimensional model, and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model.
Wherein, the target building three-dimensional model is created by a user through building modeling software, and one building three-dimensional model can comprise one or more floors, and in the embodiment, the labeling of the inner wall and the outer wall is performed by taking one floor as the minimum labeling unit; various types of primitives such as wall primitives, beam primitives, plate primitives and the like are included in the building three-dimensional model.
Specifically, the set to be identified includes: positional information for each wall primitive and relative positional information between the individual wall primitives.
Step S102: and determining the maximum closed area which can be formed by all the wall primitives in the set to be identified, and acquiring the outer contour line of the maximum closed area.
Specifically, the determining the maximum enclosed area that can be formed by all the wall primitives in the set to be identified includes:
step A1: selecting a target wall primitive from the set to be identified, determining all closed areas formed by taking the target wall primitive as a starting point according to the relative position information among the wall primitives in the set to be identified, and taking the closed area with the largest area in the formed closed areas as a representative closed area corresponding to the target wall primitive;
In this embodiment, determining a representative closed area of each wall primitive according to the manner of the step A1 for all the wall primitives of the target floor; in the process of determining the representative closed area, one wall graphic element is used as a starting point to traverse other wall graphic elements one by one, so that all closed areas which can be formed by taking the wall graphic element as the starting point are determined, and finally, the closed area with the largest area is used as the representative closed area corresponding to the closed area.
Step A2: calculating the area of the representative closed area of each wall primitive in the set to be identified, and taking the representative closed area with the largest area as the maximum closed area;
in this embodiment, the closed area with the largest area that can be formed by all the wall primitives in the target floor is determined by traversing all the wall primitives in the target floor, and the closed area is used as a reference for determining the inner and outer walls.
Step S103: inputting the set to be identified and the outer contour line into a preset feature extraction algorithm to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line.
Wherein the location feature information includes: the distance from the center point of the wall primitive to the center point of the outer contour line, the position relation between the wall primitive and the outer contour line and the nearest distance from the center of the wall primitive to the outer contour line.
Preferably, the feature extraction algorithm is SIFT (Scale Invariant Feature Transform, scale-invariant feature transform) algorithm. The SIFT algorithm is a local feature descriptor, and has good stability and invariance: the device can adapt to rotation, scale scaling and brightness variation; good distinguishing property: the rapid and accurate distinguishing information can be matched in a mass characteristic database; multiple-quantum: a large number of feature vectors can be generated even with a single object; high speed performance: feature vector matching can be rapidly carried out; scalability: can be combined with other forms of feature vectors; etc.
Step S104: and inputting the position characteristic information of all the wall primitives and the outer contour line into a preset classification algorithm so as to divide all the wall primitives into an inner wall set and an outer wall set based on the outer contour line.
Specifically, the inner wall set includes all wall primitives located within the outer contour, and the outer wall set includes all wall primitives located outside the outer contour.
Preferably, the classification algorithm is an SVM (Support Vector Machine ) algorithm. SVM is a supervised learning model commonly used for pattern recognition, classification, and regression analysis. The SVM algorithm is a two-class algorithm, and determines a dividing line for dividing the discrete points into two classes based on a positional relationship between the discrete points, and in this embodiment, the positional characteristic information of each wall primitive corresponds to the discrete points, and the outer contour line corresponds to the dividing line.
Step S105: and marking each wall primitive in the inner wall set as an inner wall, and marking each wall primitive in the outer wall set as an outer wall.
In this embodiment, the inner and outer walls may be marked on each wall primitive of each floor in the three-dimensional model of the target building according to the steps S101 to S105.
Specifically, before step S101, the method further includes:
step B1: acquiring a training sample set; the training sample set comprises a plurality of training samples, and each training sample represents all wall primitives of a floor in a three-dimensional building model;
wherein, the training sample can be actual data in an actual building project.
Step B2: determining a maximum sample closed area which can be formed by all wall primitives in the training sample, and acquiring a sample outline of the maximum sample closed area;
it should be noted that, the maximum sample sealing area may be determined according to the above-mentioned methods from step A1 to step A2, which is not described herein.
Step B3: inputting the training sample and the sample outline into a preset feature extraction algorithm to calculate sample position feature information of each wall primitive in the training sample relative to the sample outline;
The feature extraction algorithm adopts a SIFT algorithm, and the sample position feature information comprises: the distance from the center point of the wall primitive to the center point of the sample outer contour line, the positional relationship between the wall primitive and the sample outer contour line, and the nearest distance from the center of the wall primitive to the sample outer contour line.
Step B4: sample position characteristic information of all wall primitives in the training sample is input into a preset clustering algorithm, so that all sample position characteristic information is clustered into an inner wall training set, an outer wall training set and an unknown wall training set;
preferably, the clustering algorithm is a K-means clustering algorithm (K-means Clustering Algorithm); the K-means algorithm model is one of the classic algorithms in the Clustering, and the data mining is one of ten classic algorithms, the core idea of the algorithm model is to cluster by taking K points in space as the centers, classify the objects closest to the K points, and successively update the values of the Clustering centers by an iterative method until the best Clustering result is obtained. In this example the value of K is 3.
The interior wall training set comprises: sample position characteristic information of wall primitives belonging to the inner wall in the training samples, wherein the outer wall training set comprises: sample position characteristic information of wall primitives belonging to an outer wall in a training sample, wherein the unknown wall training set comprises: and sample position characteristic information of wall primitives except for the inner wall and the outer wall in the training sample. In practical applications, there may be wall primitives that cannot be identified as an inner wall or an outer wall, and therefore, in this embodiment, wall primitives that belong to neither an inner wall nor an outer wall are set as unknown walls.
Step B5: and calculating the common characteristic information of the inner wall according to the position characteristic information of all samples in the inner wall training set, and calculating the common characteristic information of the outer wall according to the position characteristic information of all samples in the outer wall training set.
In practical application, the intersection of all sample position feature information in the inner wall training set can be used as inner wall common feature information, and the intersection of all sample position feature information in the outer wall training set can be used as outer wall common feature information, or the inner wall common feature information and the outer wall common feature information can be calculated according to other information fusion algorithms, and the method is not limited specifically.
Further, step S105 specifically includes:
sequentially calculating a first similarity value of the position characteristic information and the inner wall commonality characteristic information of each wall primitive in the inner wall set, judging whether the first similarity value is larger than a first preset threshold value, if so, marking the corresponding wall primitive as an inner wall, and if not, marking the corresponding wall primitive as an unknown wall;
and sequentially calculating a second similarity value of the position characteristic information of each wall primitive in the outer wall set and the outer wall commonality characteristic information, judging whether the second similarity value is larger than a second preset threshold value, if so, marking the corresponding wall primitive as an outer wall, and if not, marking the corresponding wall primitive as an unknown wall.
In practical applications, the first preset threshold and the second preset threshold may be set to the same value or different values.
Still further, the method further comprises:
and sending the wall graphic elements marked as the unknown walls to a designated terminal so as to manually mark the inner walls and the outer walls by a user.
In the embodiment, the SIFT algorithm, the K-means algorithm and the SVM algorithm are applied to the scene of intelligent identification of the inner and outer walls, and a model capable of realizing intelligent labeling of the inner and outer walls is obtained through a large amount of training, so that a large amount of human resources are not consumed any more, and the efficiency of labeling the inner and outer walls is improved.
The steps S101 to S105 are execution logic of a model for labeling the inner and outer walls, which is trained in advance, and in this embodiment, model training is performed based on a machine learning algorithm, so as to achieve the purpose of automatically judging the inner and outer walls by machine learning. As shown in fig. 2, the model training process mainly comprises the following parts: data acquisition analysis, model training, model evaluation and model use. The model acquisition analysis is to collect specific data in the actual building engineering of a user to serve as training data set data, the model training process is to extract characteristic information through SIFT algorithm through the collected training data set data, then clustering and grouping all the characteristic information by using a K-means algorithm model to serve as training result set, and finally, the SVM algorithm is utilized to compare the similarity between the data to be processed in the actual building engineering and the training result set to judge whether the data to be processed belongs to an inner wall or an outer wall.
Specifically, the data acquisition and analysis stage comprises the following steps:
(1) And (3) data acquisition: through first-line customer service, the design drawings and actual projects of users are collected, and actual project files which are drawn and only need quantity are mainly collected, wherein the actual project files are different from the actual project files of users in the middle process of drawing, the model data and the inner wall and outer wall identification attributes of the actual projects which are drawn and completed approach or are equivalent to the positions of walls and inner wall and outer wall identifications in the actual construction scenes and the design drawings, and the actual projects represent scenes which can be really met in actual use.
(2) Data processing: after the actual projects and drawings of the user are collected, the actual projects are further processed, all wall primitive data in the scenes in the actual projects of the user are collected, primitive IDs, primitive coordinates, associated primitives, internal and external wall marks and other data are arranged, and the data can be used as algorithm training original data to obtain an algorithm training data set.
(3) Data classification: by means of exploratory data analysis of the dimensions including the graphic primitive orders, the number of the wall graphic primitives and the like of the actual engineering, the actual engineering including the different wall graphic primitive orders is classified according to the wall graphic primitive attribute inner and outer wall marks which are already made by a user in the actual engineering, so that the actual engineering is used for algorithm training and learning reference data. When data classification is performed, data skew cannot be too serious, and the number of data in different categories is not subject to a gap of several orders of magnitude. And the magnitude of the data is also evaluated, the consumption degree of the memory can be evaluated by the number of samples and the number of characteristics, and whether the memory can be put down in the training process is judged. If the actual engineering with larger graphic element magnitude exists in the collected data after classification, the part of the collected data needing dimension reduction and distributed processing is firstly considered to analyze how large the magnitude proportion of the other graphic elements which are irrelevant to the wall graphic elements in the actual engineering is, and the other graphic elements which are irrelevant to the wall are deleted in batches, because the wall graphic elements judge that the inner wall and outer wall marks are only relevant to the wall type graphic elements, the graphic elements can not influence the judgment of the inner wall and outer wall marks, the dimension reduction of the graphic element magnitude is carried out by deleting the irrelevant graphic elements, and if the wall type graphic elements are still large after the irrelevant graphic elements are deleted, the data segmentation operation processing such as the actual engineering can be split according to floors.
(4) Data preprocessing: for the collected data set, some engineering scenes are special, and more or less data loss, illegal data, uneven distribution, abnormal data and the like exist, so that the collected data needs to be further processed, which is called as data preprocessing. The data preprocessing is mainly aimed at the problems, whether a support algorithm can be used for learning is judged through scene analysis of missing data, illegal and abnormal scene data are removed from a data set, and the illegal and abnormal scene data are removed from an illegal and abnormal scene, so that the data are distributed unevenly. After the data preprocessing, the data normalization is ensured, and the subsequent processing on the aspects of feature extraction, data dimension reduction and the like of the data set is facilitated.
(5) Data set segmentation: after a series of processing, the collected data set can be divided into three independent parts: training set (train set), validation set (validation set) and test set (test set). Wherein the training set is used to estimate the model, the verification set is used to adjust model parameters to obtain an optimal model, and the test set is used to verify how the optimal model performs. One typical division is 50% of the total sample for the training set and 25% for each of the other, all three being randomly extracted from the sample.
Further, the model training stage comprises the following steps:
prior to model training, suitable algorithms such as linear regression, decision trees, random forests, logistic regression, gradient boosting, SVM, etc. are determined. The best method when selecting an algorithm is to test various algorithms and then select the best one by cross-validation. In the embodiment, a K-means algorithm model is selected as a training model of the time, a SIFT algorithm and an SVM algorithm are utilized to assist in executing a machine algorithm learning process, in the training process, feature point extraction and description are carried out on a training data set through the SIFT algorithm, after data feature extraction and description extraction are completed, a K-means core algorithm is used for clustering, an optimal solution is found, and finally, an SVM machine learning model is used for learning training.
Further, the model evaluation stage comprises the following steps:
in the model evaluation stage, different evaluation indexes can be selected according to different concerns such as classification, regression, sequencing and the like, and the loss is different from that in model selection. In the process of model evaluation, judging the over fitting and under fitting of the model, wherein under fitting refers to the stage of training starting, and the model has larger difference with the expected result, and in the process, the model algorithm is required to be continuously optimized and adjusted to improve the accuracy; after the model is trained to a certain degree, the difference between the result and the training result is larger when the test is started, and the situation is that the training model excessively depends on the existing training data set and more training data needs to be added for training. If the phenomenon of over fitting exists, the problem of over fitting can be solved by increasing the proportion of the training set or regularization; if the data fitting is not in place, the data training is not in place, the general characteristics of the data cannot be extracted, and the problem of the lack of fitting is solved by the methods of increasing the dimension of the polynomial, reducing the regularization parameters and the like. The model is evaluated through the split training data, and the quality of the model is judged through comparing the real data with the predicted data. The five common methods of model evaluation: confusion matrix, lifting map & lorentz map, coefficient of base, ks curve, roc curve. The confusion matrix cannot be used as the only standard of the evaluation model, and is the basis for calculating other indexes of the model. In addition, model evaluation should also take into account temporal, spatial complexity, stability, mobility, etc.
Further, the model using stage comprises the following steps:
(1) Model use process: packaging the trained model into a dynamic library form called by a product, calling a packaged model library when the product executes standard functions of the inner wall and the outer wall, inputting the geometric shape, the position relation and the association primitive of the wall forming a closed area into a model library interface according to the set of the data, comparing the data to be used as a reference, and feeding back the position relation of the interface to judge the result set of the inner wall and the outer wall to be used as a reference; the method comprises the steps that a SIFT algorithm, a K-means algorithm and an SVM algorithm are specifically adopted, the parameter input and output relation is as follows, the wall geometric position relation, the subsection region composition and the like of collected training set engineering model data are used as parameter input of feature factors extracted by the SIFT algorithm, the extracted feature set is used as parameter input of the K-means algorithm, a feature subset is refined again to complete model algorithm training, when actual engineering is compared, the parameter input is that an interface is input into an engineering wall data set, and the training feature subset is used for judging the attribution of an inner wall and an outer wall through the SVM algorithm comparing the engineering data and the feature set similarity; the purpose of identifying the inner wall and the outer wall is achieved by completing model comparison.
(2) Release and use: gray scale distribution is a distribution mode that can smoothly transition between black and white. AB test is a gray level distribution mode, which allows a part of users to continue to use a, a part of users to start to use B, and if the users have no objection to B, the range is gradually expanded, and all users are migrated to the top of B. The gray level release can ensure the stability of the whole system, and the problems can be found and adjusted during the initial gray level so as to ensure the influence degree.
In this embodiment, the stability of the model is gradually improved through a large amount of training, and then the processing efficiency and accuracy are improved by identifying the inner and outer walls through the model.
Example two
The embodiment of the invention provides a device for marking an inner wall and an outer wall, as shown in fig. 3, which specifically comprises the following components:
the acquisition module 301 is configured to acquire a three-dimensional model of a target building, and form a set to be identified from all wall primitives of a target floor in the three-dimensional model of the target building;
the determining module 302 is configured to determine a maximum closed area that can be formed by all wall primitives in the set to be identified, and obtain an outer contour line of the maximum closed area;
the extracting module 303 is configured to input the set to be identified and the outer contour line into a preset feature extracting algorithm, so as to calculate position feature information of each wall primitive in the set to be identified relative to the outer contour line;
the classification module 304 is configured to input the position feature information of all the wall primitives and the outer contour line into a preset classification algorithm, so as to divide all the wall primitives into an inner wall set and an outer wall set based on the outer contour line;
and the labeling module 305 is configured to label each wall primitive in the set of inner walls as an inner wall, and label each wall primitive in the set of outer walls as an outer wall.
Specifically, the determining module 302 is configured to:
selecting a target wall primitive from the set to be identified, determining all closed areas formed by taking the target wall primitive as a starting point according to the relative position information among the wall primitives in the set to be identified, and taking the closed area with the largest area in the formed closed areas as a representative closed area corresponding to the target wall primitive; and respectively calculating the area of the representative closed area of each wall primitive in the set to be identified, and taking the representative closed area with the largest area as the maximum closed area.
Further, the device further comprises:
the training module is used for acquiring a training sample set; the training sample set comprises a plurality of training samples, and each training sample represents all wall primitives of a floor in a three-dimensional building model; determining a maximum sample closed area which can be formed by all wall primitives in the training sample, and acquiring a sample outline of the maximum sample closed area; inputting the training sample and the sample outline into a preset feature extraction algorithm to calculate sample position feature information of each wall primitive in the training sample relative to the sample outline; sample position characteristic information of all wall primitives in the training sample is input into a preset clustering algorithm, so that all sample position characteristic information is clustered into an inner wall training set, an outer wall training set and an unknown wall training set; and calculating the common characteristic information of the inner wall according to the position characteristic information of all samples in the inner wall training set, and calculating the common characteristic information of the outer wall according to the position characteristic information of all samples in the outer wall training set.
Further, the labeling module 305 is specifically configured to:
sequentially calculating a first similarity value of the position characteristic information and the inner wall commonality characteristic information of each wall primitive in the inner wall set, judging whether the first similarity value is larger than a first preset threshold value, if so, marking the corresponding wall primitive as an inner wall, and if not, marking the corresponding wall primitive as an unknown wall;
and sequentially calculating a second similarity value of the position characteristic information of each wall primitive in the outer wall set and the outer wall commonality characteristic information, judging whether the second similarity value is larger than a second preset threshold value, if so, marking the corresponding wall primitive as an outer wall, and if not, marking the corresponding wall primitive as an unknown wall.
Further, the method further comprises:
and the sending module is used for sending the wall primitives marked as the unknown walls to the appointed terminal so as to manually mark the inner walls and the outer walls by a user.
Furthermore, the feature extraction algorithm is a scale-invariant feature transform SIFT algorithm, the classification algorithm is a support vector machine SVM algorithm, and the clustering algorithm is a K-means clustering algorithm.
Example III
The present embodiment also provides a computer device, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a rack-mounted server, a blade server, a tower server, or a rack-mounted server (including an independent server or a server cluster formed by a plurality of servers) that can execute a program. As shown in fig. 4, the computer device 40 of the present embodiment includes at least, but is not limited to: a memory 401 and a processor 402 which can be communicatively connected to each other via a system bus. It should be noted that FIG. 4 only shows computer device 40 having components 401-402, but it should be understood that not all of the illustrated components are required to be implemented and that more or fewer components may be implemented instead.
In this embodiment, the memory 401 (i.e., readable storage medium) includes flash memory, a hard disk, a multimedia card, a card memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, and the like. In some embodiments, the memory 401 may be an internal storage unit of the computer device 40, such as a hard disk or a memory of the computer device 40. In other embodiments, the memory 401 may also be an external storage device of the computer device 40, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card) or the like, which are provided on the computer device 40. Of course, memory 401 may also include both internal storage units of computer device 40 and external storage devices. In this embodiment, the memory 401 is typically used to store an operating system and various types of application software installed on the computer device 40. In addition, the memory 401 can also be used to temporarily store various types of data that have been output or are to be output.
The processor 402 may be a central processing unit (Central Processing Unit, CPU), controller, microcontroller, microprocessor, or other data processing chip in some embodiments. The processor 402 is generally used to control the overall operation of the computer device 40.
Specifically, in this embodiment, the processor 402 is configured to execute a program for labeling an inner wall and an outer wall, where the program for labeling an inner wall and an outer wall is stored in the memory 401, and when executed, implement the following steps:
acquiring a target building three-dimensional model, and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model;
determining the maximum closed area which can be formed by all wall primitives in the set to be identified, and acquiring the outer contour line of the maximum closed area;
inputting the set to be identified and the outer contour line into a preset feature extraction algorithm to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line;
the position characteristic information of all the wall primitives and the outer contour line are input into a preset classification algorithm, so that all the wall primitives are divided into an inner wall set and an outer wall set based on the outer contour line;
And marking each wall primitive in the inner wall set as an inner wall, and marking each wall primitive in the outer wall set as an outer wall.
The specific embodiment process of the above method steps can be referred to as embodiment one, and the description of this embodiment is not repeated here.
Example IV
The present embodiment also provides a computer readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, a server, an App application store, etc., having stored thereon a computer program that when executed by a processor performs the following method steps:
acquiring a target building three-dimensional model, and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model;
determining the maximum closed area which can be formed by all wall primitives in the set to be identified, and acquiring the outer contour line of the maximum closed area;
inputting the set to be identified and the outer contour line into a preset feature extraction algorithm to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line;
The position characteristic information of all the wall primitives and the outer contour line are input into a preset classification algorithm, so that all the wall primitives are divided into an inner wall set and an outer wall set based on the outer contour line;
and marking each wall primitive in the inner wall set as an inner wall, and marking each wall primitive in the outer wall set as an outer wall.
The specific embodiment process of the above method steps can be referred to as embodiment one, and the description of this embodiment is not repeated here.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one … …" does not exclude the presence of other like elements in a process, method, article, or apparatus that comprises the element.
The foregoing embodiment numbers of the present invention are merely for the purpose of description, and do not represent the advantages or disadvantages of the embodiments.
From the above description of the embodiments, it will be clear to those skilled in the art that the above-described embodiment method may be implemented by means of software plus a necessary general hardware platform, but of course may also be implemented by means of hardware, but in many cases the former is a preferred embodiment.
The foregoing description is only of the preferred embodiments of the present invention, and is not intended to limit the scope of the invention, but rather is intended to cover any equivalents of the structures or equivalent processes disclosed herein or in the alternative, which may be employed directly or indirectly in other related arts.
Claims (10)
1. A method of labeling an interior and exterior wall, the method comprising:
acquiring a target building three-dimensional model, and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model;
determining the maximum closed area which can be formed by all wall primitives in the set to be identified, and acquiring the outer contour line of the maximum closed area;
inputting the set to be identified and the outer contour line into a preset feature extraction algorithm to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line;
The position characteristic information of all the wall primitives and the outer contour line are input into a preset classification algorithm, so that all the wall primitives are divided into an inner wall set and an outer wall set based on the outer contour line;
and marking each wall primitive in the inner wall set as an inner wall, and marking each wall primitive in the outer wall set as an outer wall.
2. The method for labeling an interior and exterior wall according to claim 1, wherein said determining a maximum enclosed area that can be formed by all wall primitives in the set to be identified comprises:
selecting a target wall primitive from the set to be identified, determining all closed areas formed by taking the target wall primitive as a starting point according to the relative position information among the wall primitives in the set to be identified, and taking the closed area with the largest area in the formed closed areas as a representative closed area corresponding to the target wall primitive;
and respectively calculating the area of the representative closed area of each wall primitive in the set to be identified, and taking the representative closed area with the largest area as the maximum closed area.
3. The method of labeling interior and exterior walls according to claim 1, wherein prior to said obtaining a target building three-dimensional model and constructing all wall primitives for a target floor in the target building three-dimensional model into a set to be identified, the method further comprises:
Acquiring a training sample set; the training sample set comprises a plurality of training samples, and each training sample represents all wall primitives of a floor in a three-dimensional building model;
determining a maximum sample closed area which can be formed by all wall primitives in the training sample, and acquiring a sample outline of the maximum sample closed area;
inputting the training sample and the sample outline into a preset feature extraction algorithm to calculate sample position feature information of each wall primitive in the training sample relative to the sample outline;
sample position characteristic information of all wall primitives in the training sample is input into a preset clustering algorithm, so that all sample position characteristic information is clustered into an inner wall training set, an outer wall training set and an unknown wall training set;
and calculating the common characteristic information of the inner wall according to the position characteristic information of all samples in the inner wall training set, and calculating the common characteristic information of the outer wall according to the position characteristic information of all samples in the outer wall training set.
4. A method of labeling interior and exterior walls as in claim 3 wherein labeling each wall primitive in the set of interior walls as an interior wall and labeling each wall primitive in the set of exterior walls as an exterior wall comprises:
Sequentially calculating a first similarity value of the position characteristic information and the inner wall commonality characteristic information of each wall primitive in the inner wall set, judging whether the first similarity value is larger than a first preset threshold value, if so, marking the corresponding wall primitive as an inner wall, and if not, marking the corresponding wall primitive as an unknown wall;
and sequentially calculating a second similarity value of the position characteristic information of each wall primitive in the outer wall set and the outer wall commonality characteristic information, judging whether the second similarity value is larger than a second preset threshold value, if so, marking the corresponding wall primitive as an outer wall, and if not, marking the corresponding wall primitive as an unknown wall.
5. The method of labeling an interior and exterior wall according to claim 4, wherein after labeling each wall primitive in the set of interior walls as an interior wall and labeling each wall primitive in the set of exterior walls as an exterior wall, the method further comprises:
and sending the wall graphic elements marked as the unknown walls to a designated terminal so as to manually mark the inner walls and the outer walls by a user.
6. The method for labeling an interior and an exterior wall according to claim 3, wherein the feature extraction algorithm is a scale-invariant feature transform SIFT algorithm, the classification algorithm is a support vector machine SVM algorithm, and the clustering algorithm is a K-means clustering algorithm.
7. An apparatus for marking an interior and exterior wall, the apparatus comprising:
the acquisition module is used for acquiring a target building three-dimensional model and forming a set to be identified by all wall primitives of a target floor in the target building three-dimensional model;
the determining module is used for determining the maximum closed area which can be formed by all wall primitives in the set to be identified and acquiring the outer contour line of the maximum closed area;
the extraction module is used for inputting the set to be identified and the outer contour line into a preset feature extraction algorithm so as to calculate the position feature information of each wall primitive in the set to be identified relative to the outer contour line;
the classification module is used for inputting the position characteristic information of all the wall primitives and the outer contour line into a preset classification algorithm so as to divide all the wall primitives into an inner wall set and an outer wall set based on the outer contour line;
and the labeling module is used for labeling each wall primitive in the inner wall set as an inner wall and labeling each wall primitive in the outer wall set as an outer wall.
8. The apparatus for labeling an interior and exterior wall according to claim 7, wherein the determining module is configured to:
Selecting a target wall primitive from the set to be identified, determining all closed areas formed by taking the target wall primitive as a starting point according to the relative position information among the wall primitives in the set to be identified, and taking the closed area with the largest area in the formed closed areas as a representative closed area corresponding to the target wall primitive;
and respectively calculating the area of the representative closed area of each wall primitive in the set to be identified, and taking the representative closed area with the largest area as the maximum closed area.
9. A computer device, the computer device comprising: memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the method according to any one of claims 1 to 6 when the computer program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210330345.8A CN116935010A (en) | 2022-03-30 | 2022-03-30 | Method, device and equipment for marking inner and outer walls and readable storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210330345.8A CN116935010A (en) | 2022-03-30 | 2022-03-30 | Method, device and equipment for marking inner and outer walls and readable storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116935010A true CN116935010A (en) | 2023-10-24 |
Family
ID=88381285
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210330345.8A Pending CN116935010A (en) | 2022-03-30 | 2022-03-30 | Method, device and equipment for marking inner and outer walls and readable storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116935010A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117909838A (en) * | 2024-03-15 | 2024-04-19 | 合肥坤颐建筑科技合伙企业(有限合伙) | Basement exterior wall labeling method, device, equipment and storage medium |
-
2022
- 2022-03-30 CN CN202210330345.8A patent/CN116935010A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117909838A (en) * | 2024-03-15 | 2024-04-19 | 合肥坤颐建筑科技合伙企业(有限合伙) | Basement exterior wall labeling method, device, equipment and storage medium |
| CN117909838B (en) * | 2024-03-15 | 2024-06-04 | 合肥坤颐建筑科技合伙企业(有限合伙) | Basement exterior wall marking method, device, equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110825899B (en) | Clothing Image Retrieval Method Fused with Color Feature and Residual Network Depth Feature | |
| CN109325538B (en) | Object detection method, device and computer-readable storage medium | |
| US9330341B2 (en) | Image index generation based on similarities of image features | |
| CN110991465B (en) | Object identification method, device, computing equipment and storage medium | |
| CN101103373B (en) | improved image segmentation method | |
| Shabanzade et al. | Combination of local descriptors and global features for leaf recognition | |
| Yacob et al. | Anomaly detection in Skin Model Shapes using machine learning classifiers | |
| CN111178196B (en) | Cell classification method, device and equipment | |
| CN108764361B (en) | Working condition identification method of indicator diagram of beam-pumping unit based on integrated learning | |
| CN111292377B (en) | Target detection method, device, computer equipment and storage medium | |
| CN117636421A (en) | Face deep pseudo detection method based on edge feature acquisition | |
| CN111783910A (en) | Building project management method, electronic equipment and related products | |
| Liu et al. | Planogram compliance checking using recurring patterns | |
| CN116935010A (en) | Method, device and equipment for marking inner and outer walls and readable storage medium | |
| CN112115996A (en) | Image data processing method, device, equipment and storage medium | |
| CN106610977A (en) | Data clustering method and device | |
| CN118761591B (en) | Project matching method and system based on automatic data identification | |
| CN117909333B (en) | Screening method and system for realizing data based on big data combined with artificial intelligence | |
| Takaishi et al. | Free-Form Feature Classification for Finite Element Meshing based on Shape Descriptors and Machine Learning. | |
| CN118230284A (en) | A method and device for quickly identifying road elements using a mobile laser radar | |
| CN119173915A (en) | Training method, segmentation method, computer program product and segmentation device for training machine learning algorithms | |
| CN112734337B (en) | Article monitoring management method and device based on information acquisition | |
| CN120147826B (en) | Ore grade identification method and system | |
| CN119478922B (en) | Method, system and storage medium for evaluating silkworm pupa status for making silkworm pupa | |
| WO2020255223A1 (en) | Identification result explanation device, identification result explanation method, and identification result explanation program |
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 |