CN110555219B - Three-dimensional CAD model similarity retrieval system and method based on image recognition - Google Patents
Three-dimensional CAD model similarity retrieval system and method based on image recognition Download PDFInfo
- Publication number
- CN110555219B CN110555219B CN201810551210.8A CN201810551210A CN110555219B CN 110555219 B CN110555219 B CN 110555219B CN 201810551210 A CN201810551210 A CN 201810551210A CN 110555219 B CN110555219 B CN 110555219B
- Authority
- CN
- China
- Prior art keywords
- similarity
- dimensional cad
- cad model
- dimensional
- database
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
A three-dimensional CAD model similarity retrieval system and method based on image recognition, the system comprises a database, a feature recognition module, a similarity calculation module, a similarity retrieval module and a result display module; storing the three-dimensional CAD model and the target three-dimensional CAD model in a database; the feature identification module traverses all three-dimensional CAD models in the database, identifies geometric features of each three-dimensional CAD model in at least three view directions, takes the geometric features as view feature values, and stores the view feature values in the database in association with the corresponding three-dimensional CAD models; the similarity calculation module calculates the similarity of any two three-dimensional CAD models in the database according to the view characteristic values of the three-dimensional CAD models, and stores all similarity results obtained by calculation in the database; the similarity retrieval module sorts similarity results related to the target three-dimensional CAD model in the database; and the result display module is used for displaying the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity according to the similarity sequence obtained by the similarity retrieval module.
Description
Technical Field
The invention belongs to the technical field of digital manufacturing.
Background
Although the three-dimensional CAD modeling technology is increasingly adopted in the product design stage, the original two-dimensional drawing is replaced by the three-dimensional CAD model, the reuse of the three-dimensional CAD model information is not sufficient in the process design and product manufacturing links of the product, the manual application of process personnel is mainly used, and the related process design work is not carried out in the mode of automatically identifying and applying the three-dimensional CAD model information, so that the aim of improving the process design efficiency and quality is fulfilled.
The three-dimensional model-based process design is mainly based on the inheritance process design. The craft personnel often need to search whether similar products exist or not, whether process design of the similar products is carried out or not, and then the process design work of new products and new parts is completed by referring to the existing process design content and data.
However, for the three-dimensional CAD model, without a technical means for efficiently searching and retrieving similar models, the method mainly depends on the personal experience of the process personnel, or needs to continuously consult other experienced personnel, so that the working efficiency is low, and the working quality is difficult to ensure.
Therefore, if the product similarity retrieval technology and method based on the three-dimensional CAD model can be realized, the retrieval and reuse efficiency of the existing data in the process design process can be greatly improved, the pushing of the process data and process knowledge based on similar products can be further realized, and the intelligent process design and application based on knowledge can be realized.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the system and the method can automatically retrieve one or more similar three-dimensional CAD models in an existing three-dimensional CAD model library aiming at the specified three-dimensional CAD model, and can perform sequencing and display according to the similarity of the three-dimensional CAD models.
The technical solution of the invention is as follows: a three-dimensional CAD model similarity retrieval system based on image recognition comprises a database, a feature recognition module, a similarity calculation module, a similarity retrieval module and a result display module;
storing the three-dimensional CAD model and the target three-dimensional CAD model in a database;
the feature recognition module traverses all three-dimensional CAD models in the database, recognizes the geometric features of each three-dimensional CAD model in at least three view directions, takes the geometric features as view feature values, and stores the view feature values in the database in association with the corresponding three-dimensional CAD models;
the similarity calculation module calculates the similarity of any two three-dimensional CAD models in the database according to the view characteristic values of the three-dimensional CAD models, and stores all similarity results obtained through calculation in the database;
the similarity retrieval module sorts similarity results related to the target three-dimensional CAD model in the database;
and the result display module is used for displaying the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity according to the similarity sequence obtained by the similarity retrieval module.
A three-dimensional CAD model similarity retrieval system based on image recognition comprises a database, a feature recognition module, a similarity calculation module, a similarity retrieval module and a result display module;
storing the three-dimensional CAD model and the target three-dimensional CAD model in a database;
the feature identification module traverses all three-dimensional CAD models in the database, identifies geometric features of each three-dimensional CAD model in at least three view directions, takes the geometric features as view feature values, and stores the view feature values in the database in association with the corresponding three-dimensional CAD models;
the similarity calculation module firstly preprocesses all three-dimensional CAD models in the database, screens out a group of three-dimensional CAD models, calculates the similarity between a target three-dimensional CAD model and the screened models, and stores all similarity results obtained by calculation in the database;
the similarity retrieval module sorts the similarity results related to the target three-dimensional CAD model in the database;
and the result display module displays the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity thereof according to the similarity sequence obtained by the similarity retrieval module.
Further, the pretreatment is realized by the following modes:
the method comprises the following steps of firstly, giving a group of view characteristic values as view characteristic values of a standard three-dimensional CAD model; firstly, calculating the mean value of all view characteristic values of each three-dimensional CAD model in a database, and then calculating the similarity with a standard three-dimensional CAD model by using the mean value to form standard similarity;
secondly, calculating the mean value of the corresponding view characteristic values of all view directions of the target three-dimensional CAD model, and then calculating the similarity of the standard three-dimensional CAD model and the mean value to form the standard similarity of the target three-dimensional CAD model;
and thirdly, taking the similarity obtained in the second step as a median value within a preset span range, and screening out the three-dimensional CAD model with the similarity within the span range from the first step.
Further, the geometric features include contour perimeter, contour area, and corner features.
Furthermore, the chamfer in the three-dimensional CAD model is identified according to the characteristics of one corner point.
Further, the characteristic identification module also identifies material attribute characteristics and stores the identified material attribute characteristics to a database; and the similarity retrieval module screens the three-dimensional CAD models with the same material attribute according to whether the preset material attribute characteristics are considered or not, if the material attribute characteristics are required to be considered, the screened models are sorted according to the similarity, and if the material attribute characteristics are not required to be considered, the three-dimensional CAD models are directly sorted according to the similarity calculated by the similarity calculation module.
Further, the similarity calculation of the two three-dimensional CAD models is realized by the following steps:
(a) Calculating the similarity of the geometric features of the three-dimensional CAD model i in each view direction and the geometric features of the three-dimensional CAD model b in all view directions according to the following formula, and recording all similarity values;
(b) Sorting all similarity values from large to small;
(c) Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, selecting similarity according to the principle of first-large and second-small in the sequencing of the step (b);
(d) Calculating the mean value of the selected similarity as the similarity of the three-dimensional CAD models i and b;
wherein the similarity calculation formula is as follows:
wherein: tables i, b below identify two three-dimensional CAD models,
a represents the area of the three-dimensional figure in the view direction
P represents the perimeter of the three-dimensional figure in the view direction
C denotes the corner count of the three-dimensional figure in the view direction.
A three-dimensional CAD model similarity retrieval method based on image recognition comprises the following steps:
(1) Traversing all three-dimensional CAD models in the database, wherein each three-dimensional CAD model generates at least 3 three-dimensional views in the upper, lower, left, right, front and back 6 directions;
(2) Recognizing the contour perimeter, area and corner data of each three-dimensional view in each three-dimensional CAD model, using the contour perimeter, area and corner data as view characteristic values and storing the view characteristic values into a database;
(3) And calculating the similarity of any two three-dimensional CAD models in the database, and sequencing the similarity results related to the target three-dimensional CAD model in the database.
A three-dimensional CAD model similarity retrieval method based on image recognition comprises the following steps:
(1) Traversing all three-dimensional CAD models in the database, wherein each three-dimensional CAD model generates at least 3 three-dimensional views in the upper, lower, left, right, front and back 6 directions;
(2) Recognizing the contour perimeter, area and corner data of each three-dimensional view in each three-dimensional CAD model, using the contour perimeter, area and corner data as view characteristic values and storing the view characteristic values into a database;
(3) Giving a group of view characteristic values as view characteristic values of a standard three-dimensional CAD model; calculating the similarity between each three-dimensional CAD model in the database and the standard three-dimensional CAD model;
(4) Calculating the similarity between the target three-dimensional CAD model and the standard three-dimensional CAD model;
(5) Taking the similarity obtained in the step (4) as a median value in a preset span range, and screening out a three-dimensional CAD model with the similarity in the span range from the step (3);
(6) And (5) calculating the similarity between the target three-dimensional CAD model and the model screened in the step (5), and sequencing the similarity results.
Further, if the material property characteristics need to be considered, the following steps are performed on the basis of the similarity median screening result:
and screening the three-dimensional CAD models with the same material attributes from the screening results with the similarity as a median value in a preset span range, and then calculating and sequencing the screened models according to the similarity.
Further, the similarity calculation of the two three-dimensional CAD models is realized by the following steps:
(a) Calculating the similarity of the geometric features of the three-dimensional CAD model i in each view direction and the geometric features of the three-dimensional CAD model b in all view directions according to the following formula, and recording all similarity values;
(b) Sorting all similarity values from large to small;
(c) Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, selecting similarity from the sequence of the step (b) according to the principle of big first and small second;
(d) Calculating the mean value of the selected similarity as the similarity of the three-dimensional CAD models i and b;
wherein the similarity calculation formula is as follows:
wherein: tables i, b below identify two three-dimensional CAD models,
a represents the area of the three-dimensional figure in the view direction
P represents the perimeter of the three-dimensional figure in the view direction
C denotes the corner count of the three-dimensional figure in the view direction.
A rapid process design method of a three-dimensional CAD model is realized by the following modes:
storing an identification corresponding to a process design file corresponding to the three-dimensional CAD model in a database;
obtaining a three-dimensional CAD model similar to the target three-dimensional CAD model and a similarity ranking thereof according to the mode of claim 8 or 9;
according to the similarity sorting result, combining a specific three-dimensional CAD model structure, selecting a three-dimensional CAD model from the similarity sorting and obtaining a corresponding process design file identifier;
and pushing the corresponding process design file to serve as a basic file of the process design of the target three-dimensional CAD model according to the acquired process design file identification, and then combining a specific target three-dimensional CAD model on the basis of the file to further complete the process design file of the target three-dimensional CAD model.
Compared with the prior art, the invention has the beneficial effects that:
(1) The three-dimensional CAD model generally has and can be positioned to views in 6 directions, namely, the upper direction, the lower direction, the left direction, the right direction, the front direction and the back direction, the similarity of the three-dimensional model is compared through the views in the 6 directions, and the three-dimensional CAD model has good universality.
(2) The similarity retrieval aiming at the large-batch three-dimensional CAD models can be realized, and the retrieval efficiency is high.
(3) Interception, feature extraction, similarity calculation, similarity sequencing and the like of the three-dimensional CAD model view can be automatically finished, and manual identification and manual operation are avoided.
(4) Through the similar three-dimensional CAD model, corresponding process data and process knowledge are pushed in a correlated mode, and the efficiency and quality of process design can be effectively improved.
By applying the three-dimensional CAD model similarity retrieval system based on image recognition and the method thereof, the existing and similar product three-dimensional CAD model can be quickly positioned when the process design work based on the three-dimensional CAD model is carried out, and the existing process data and process knowledge are pushed and reused through the incidence relation between the product three-dimensional CAD model and the process scheme, the process rules, the process knowledge and the like, so that the knowledge-based quick process design is realized, the process design capability of process personnel is improved, and the quality of the process design work is improved.
Drawings
FIG. 1 is a system framework diagram of the present invention;
FIG. 2 is a flow chart of the method of the present invention;
FIG. 3 is a schematic diagram of the calculation of the direction similarity of the two relevant views of the model.
Detailed Description
The invention is described in detail below with reference to the figures and examples.
Example one
A three-dimensional CAD model similarity retrieval system based on image recognition comprises a database, a feature recognition module, a similarity calculation module, a similarity retrieval module and a result display module;
storing the three-dimensional CAD model and the target three-dimensional CAD model in a database;
the feature identification module traverses all three-dimensional CAD models in the database, identifies geometric features of each three-dimensional CAD model in at least three view directions, takes the geometric features as view feature values, and stores the view feature values in the database in association with the corresponding three-dimensional CAD models;
the similarity calculation module calculates the similarity of any two three-dimensional CAD models in the database according to the view characteristic values of the three-dimensional CAD models, and stores all similarity results obtained by calculation in the database;
the similarity retrieval module sorts similarity results related to the target three-dimensional CAD model in the database;
and the result display module is used for displaying the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity according to the similarity sequence obtained by the similarity retrieval module.
Example two
A three-dimensional CAD model similarity retrieval system based on image recognition comprises a database, a feature recognition module, a similarity calculation module, a similarity retrieval module and a result display module;
storing the three-dimensional CAD model and the target three-dimensional CAD model in a database;
the feature recognition module traverses all three-dimensional CAD models in the database, recognizes the geometric features of each three-dimensional CAD model in at least three view directions, takes the geometric features as view feature values, and stores the view feature values in the database in association with the corresponding three-dimensional CAD models;
the similarity calculation module is used for preprocessing all three-dimensional CAD models in the database, screening out a group of three-dimensional CAD models, calculating the similarity between a target three-dimensional CAD model and the screened out models, and storing all similarity results obtained through calculation in the database;
the similarity retrieval module sorts the similarity results related to the target three-dimensional CAD model in the database;
and the result display module displays the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity thereof according to the similarity sequence obtained by the similarity retrieval module.
The above-mentioned related models are described in detail below (mainly describing the functions of the modules in the second embodiment, and the modules in the first embodiment are adaptively adjusted according to the following description).
The database mainly stores three-dimensional CAD models and target three-dimensional CAD models. For application convenience, a three-dimensional CAD model information data table, a three-dimensional CAD model view data table and a standard CAD model characteristic value data table can be set in the database.
Establishing model ID, drawing number, material name, material brand, standard similarity and process design file identification attribute fields in a three-dimensional CAD model information data table;
and establishing model ID, view name, perimeter, area and corner attribute fields in a three-dimensional CAD model view data table associated with the model ID.
And establishing a standard CAD model characteristic value data table, establishing 3 attribute fields of perimeter, area and angular point, and filling the perimeter, area and angular point values according to experience or statistical data.
NET and other development tools can be utilized in actual engineering application to establish a Windows window application program, and by integrating a PDM interface, existing and newly-added part three-dimensional models in a PDM system can be obtained in an incremental mode according to a certain time interval, and the models are stored in a database.
(II) feature recognition module 2
And extracting the attribute information of the model ID, the drawing number, the material name and the material brand by traversing the three-dimensional CAD model in the database, and storing the attribute information into a three-dimensional CAD model information data table.
When the method is applied to actual engineering, plug-ins such as PView. Ocx in Creo View Express can be integrated through a Windows window, and then a three-dimensional model in a database is loaded through sourceUrl attribute setting. When the model is loaded, model additional information such as size labels and the like in the model needs to be automatically shielded and hidden; after the model loading is completed, the model color is set to black or other dark color by calling the SetInstanceColor interface. And (4) setting the model into a three-dimensional view in an 'upper' direction by calling interfaces such as SetViewlocation and the like, and automatically adapting to the size of the window. The displayed view is then truncated. And (3) the intercepted three-dimensional view in the 'upper' direction is named according to the 'model ID _ top' (other view directions are: bottom, left, right, front and back in sequence), and the three-dimensional view is stored as a JPG format picture.
By integrating an opencv function module, based on an opencv image recognition function, the perimeter value, the area value and the angular point count of the figure outline of the JPG picture in each direction are automatically recognized and used as view characteristic values, and then the view characteristic values and the model ID are stored in a three-dimensional CAD model view data table.
(III) similarity calculation Module 3
Preprocessing all three-dimensional CAD models in the database: for the data of the direction area, the perimeter and the corner point of 6 views of each three-dimensional CAD model in the database, the standard similarity S is calculated by firstly calculating the mean value of the 6 views and each item of data, then calculating the standard similarity S with the view characteristic value of the standard three-dimensional CAD model according to the following formula is And storing the result to a standard similarity field of a corresponding model ID and three-dimensional CAD model information data table.
Wherein:
A i mean value representing the current model, all view graph areas
P i Mean value representing the perimeter of the current model, all view graphs
C i Mean value representing the current model, all view graph corner counts
A s Representing area values in a standard CAD model characteristic value data sheet
P s Representing perimeter values in a standard CAD model feature value data sheet
C s Representing corner values in a standard CAD model characteristic value data sheet
When a target three-dimensional CAD model is designated, a three-dimensional CAD model information data table is retrieved through model ID, drawing number and the like, the mean value of the view characteristic values of the target three-dimensional CAD model is firstly calculated, and then the similarity S between the target three-dimensional CAD model and a standard three-dimensional CAD model is calculated by utilizing the mean value sb ；
To obtain the similarity S sb As median value within a predetermined span range, from S is And screening out the three-dimensional CAD model with the similarity within the span range.
And then calculating the similarity between the target three-dimensional CAD model and the screened CAD model.
In particular, the above-mentioned calculation of the similarity can be summarized in several steps, wherein two three-dimensional CAD models are identified with indices i, b,
(a) Calculating the similarity of the geometric features of the three-dimensional CAD model i in each view direction and the geometric features of the three-dimensional CAD model b in all view directions according to the following formula, and recording all similarity values;
(b) Sorting all similarity values from large to small;
(c) Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, selecting similarity from the sequence of the step (b) according to the principle of big first and small second;
examples are as follows:
according to the schematic of fig. 3, the similarity calculation is performed on the views of the two models in three directions in a crossed manner, and the obtained similarity results are as follows:
| view of the drawing | View of the drawing | Similarity of the two |
| A top | B top | 0.9 |
| A top | B left | 0.8 |
| A top | B right | 0.7 |
| A left | B top | 0.6 |
| A left | B left | 0.5 |
| A left | B right | 0.4 |
| A right | B top | 0.3 |
| A right | B left | 0.2 |
| A right | B right | 0.1 |
Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, the selection process of selecting the similarity according to the principle of big first and small second in the sequencing of the step (b) is shown as the following table:
the final results were selected as follows:
| view of the drawing | View of the drawing | Degree of similarity |
| A top | B top | 0.9 |
| A left | B left | 0.5 |
| A right | B right | 0.1 |
(d) Calculating the mean value of the selected similarity as the similarity of the three-dimensional CAD models i and b;
wherein the similarity calculation formula is as follows:
wherein: tables i, b below identify two three-dimensional CAD models,
a represents the area of the three-dimensional figure in the view direction
P represents the perimeter of the three-dimensional figure in the view direction
C denotes the corner count of the three-dimensional figure in the view direction.
(IV) similarity retrieval module 4 and result display module 5
And the similarity retrieval module screens the three-dimensional CAD models with the same material attribute according to whether the preset material attribute characteristics are considered or not, if the material attribute characteristics are required to be considered, the screened models are sorted according to the similarity, and if the material attribute characteristics are not required to be considered, the three-dimensional CAD models are directly sorted according to the similarity calculated by the similarity calculation module.
And the result display module displays the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity thereof according to the similarity sequence obtained by the similarity retrieval module. Such as displaying the corresponding three-dimensional CAD model, model ID, drawing number, similarity (sorted from big to small by similarity).
When the system is used for realizing rapid process design, the identification corresponding to the process design file corresponding to the three-dimensional CAD model is stored in the database; selecting a three-dimensional CAD model from the similarity ranking and obtaining a corresponding process design file identifier; and pushing the corresponding process design file to serve as a basic file of the process design of the target three-dimensional CAD model according to the acquired process design file identification, and then combining a specific target three-dimensional CAD model on the basis of the file to further complete the process design file of the target three-dimensional CAD model. The process design capability of process personnel is improved, and the accumulation and reuse of knowledge are realized.
By the technology, the application targets of reducing manual operation and improving the efficiency and quality of process design in the process of process design based on the three-dimensional CAD model can be achieved.
The description of the steps of the method related to the present invention is implicit in the description of the system, and will not be described herein. The invention has not been described in detail in part of the common general knowledge of those skilled in the art.
Claims (9)
1. A three-dimensional CAD model similarity retrieval system based on image recognition is characterized in that: the system comprises a database, a feature recognition module, a similarity calculation module, a similarity retrieval module and a result display module;
storing the three-dimensional CAD model and the target three-dimensional CAD model in a database;
the feature identification module traverses all three-dimensional CAD models in the database, identifies geometric features of each three-dimensional CAD model in at least three view directions, takes the geometric features as view feature values, and stores the view feature values in the database in association with the corresponding three-dimensional CAD models;
the similarity calculation module calculates the similarity of any two three-dimensional CAD models in the database according to the view characteristic values of the three-dimensional CAD models, and stores all similarity results obtained by calculation in the database;
the similarity retrieval module sorts similarity results related to the target three-dimensional CAD model in the database;
the result display module sorts the similarity obtained by the similarity retrieval module and displays the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity;
the similarity calculation of the two three-dimensional CAD models is realized by the following steps:
(a) Calculating the similarity of the geometric features of the three-dimensional CAD model i in each view direction and the geometric features of the three-dimensional CAD model b in all view directions according to the following formula, and recording all similarity values;
(b) Sorting all similarity values from large to small;
(c) Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, selecting similarity according to the principle of first-large and second-small in the sequencing of the step (b);
(d) Calculating the mean value of the selected similarity as the similarity of the three-dimensional CAD models i and b;
wherein the similarity calculation formula is as follows:
wherein: the following tables i and b identify two three-dimensional CAD models;
a represents the area of the three-dimensional figure in the view direction;
p represents the perimeter of the three-dimensional figure in the view direction;
c denotes the corner count of the three-dimensional figure in the view direction.
2. A three-dimensional CAD model similarity retrieval system based on image recognition is characterized in that: the system comprises a database, a feature recognition module, a similarity calculation module, a similarity retrieval module and a result display module;
storing the three-dimensional CAD model and the target three-dimensional CAD model in a database;
the feature identification module traverses all three-dimensional CAD models in the database, identifies geometric features of each three-dimensional CAD model in at least three view directions, takes the geometric features as view feature values, and stores the view feature values in the database in association with the corresponding three-dimensional CAD models;
the similarity calculation module firstly preprocesses all three-dimensional CAD models in the database, screens out a group of three-dimensional CAD models, calculates the similarity between a target three-dimensional CAD model and the screened models, and stores all similarity results obtained by calculation in the database;
the similarity retrieval module sorts the similarity results related to the target three-dimensional CAD model in the database;
the result display module displays the three-dimensional CAD model similar to the target three-dimensional CAD model and the corresponding similarity thereof according to the similarity sequence obtained by the similarity retrieval module;
the pretreatment is realized by the following modes:
firstly, a group of view characteristic values are given to serve as view characteristic values of a standard three-dimensional CAD model; firstly, calculating the mean value of all view characteristic values of each three-dimensional CAD model in a database, and then calculating the similarity with a standard three-dimensional CAD model by using the mean value to form standard similarity;
secondly, calculating the mean value of the corresponding view characteristic values of all view directions of the target three-dimensional CAD model, and then calculating the similarity of the standard three-dimensional CAD model and the mean value to form the standard similarity of the target three-dimensional CAD model;
thirdly, taking the similarity obtained in the second step as a median value to be within a preset span range, and screening out a three-dimensional CAD model with the similarity within the span range from the first step;
the similarity calculation of the two three-dimensional CAD models is realized by the following steps:
(a) Calculating the similarity of the geometric features of the three-dimensional CAD model i in each view direction and the geometric features of the three-dimensional CAD model b in all view directions according to the following formula, and recording all similarity values;
(b) Sorting all similarity numerical values from large to small;
(c) Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, selecting similarity from the sequence of the step (b) according to the principle of big first and small second;
(d) Calculating the mean value of the selected similarity as the similarity of the three-dimensional CAD models i and b;
wherein the similarity calculation formula is as follows:
wherein: tables i, b below identify two three-dimensional CAD models,
a represents the area of the three-dimensional figure in the view direction
P represents the perimeter of the three-dimensional figure in the view direction
C denotes the corner count of the three-dimensional figure in the view direction.
3. The system according to claim 1 or 2, characterized in that: the geometric characteristics comprise contour perimeter, contour area and angular point characteristics.
4. The system of claim 3, wherein: and identifying the chamfer angle in the three-dimensional CAD model according to the characteristic of one corner point.
5. The system according to claim 1 or 2, characterized in that: the characteristic identification module also identifies material attribute characteristics and stores the identified material attribute characteristics to a database; and the similarity retrieval module screens the three-dimensional CAD models with the same material attribute according to whether the preset material attribute characteristics are considered or not, if the material attribute characteristics need to be considered, the screened models are sorted according to the similarity, and if the material attribute characteristics do not need to be considered, the three-dimensional CAD models are directly sorted according to the similarity calculated by the similarity calculation module.
6. A three-dimensional CAD model similarity retrieval method based on image recognition is characterized by comprising the following steps:
(1) Traversing all three-dimensional CAD models in the database, wherein each three-dimensional CAD model generates at least 3 three-dimensional views in the upper, lower, left, right, front and back 6 directions;
(2) Recognizing the contour perimeter, the area and the corner data of each three-dimensional view in each three-dimensional CAD model, using the contour perimeter, the area and the corner data as view characteristic values and storing the view characteristic values in a database;
(3) Calculating the similarity of any two three-dimensional CAD models in the database, and sequencing similarity results related to the target three-dimensional CAD model in the database;
the similarity calculation of the two three-dimensional CAD models is realized by the following steps:
(a) Calculating the similarity of the geometric characteristics of the three-dimensional CAD model i in each view direction and the geometric characteristics of the three-dimensional CAD model b in all view directions according to the following formula, and recording all similarity values;
(b) Sorting all similarity values from large to small;
(c) Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, selecting similarity from the sequence of the step (b) according to the principle of big first and small second;
(d) Calculating the mean value of the selected similarity as the similarity of the three-dimensional CAD models i and b;
wherein the similarity calculation formula is as follows:
wherein: tables i, b below identify two three-dimensional CAD models,
a represents the area of the three-dimensional figure in the view direction
P represents the perimeter of the three-dimensional figure in the view direction
C denotes the corner count of the three-dimensional figure in the view direction.
7. A three-dimensional CAD model similarity retrieval method based on image recognition is characterized by comprising the following steps:
(1) Traversing all three-dimensional CAD models in the database, wherein each three-dimensional CAD model generates at least 3 three-dimensional views in the upper, lower, left, right, front and back 6 directions;
(2) Recognizing the contour perimeter, area and corner data of each three-dimensional view in each three-dimensional CAD model, using the contour perimeter, area and corner data as view characteristic values and storing the view characteristic values into a database;
(3) Giving a group of view characteristic values as view characteristic values of a standard three-dimensional CAD model; calculating the similarity between each three-dimensional CAD model in the database and the standard three-dimensional CAD model;
(4) Calculating the similarity between the target three-dimensional CAD model and the standard three-dimensional CAD model;
(5) Taking the similarity obtained in the step (4) as a median value in a preset span range, and screening out a three-dimensional CAD model with the similarity in the span range from the step (3);
(6) Calculating the similarity between the target three-dimensional CAD model and the model screened in the step (5), and sequencing similarity results;
the similarity calculation of the two three-dimensional CAD models is realized by the following steps:
(a) Calculating the similarity of the geometric features of the three-dimensional CAD model i in each view direction and the geometric features of the three-dimensional CAD model b in all view directions according to the following formula, and recording all similarity values;
(b) Sorting all similarity values from large to small;
(c) Under the condition of ensuring that each view direction of the three-dimensional CAD model i and the three-dimensional CAD model b is only used once, selecting similarity from the sequence of the step (b) according to the principle of big first and small second;
(d) Calculating the mean value of the selected similarity as the similarity of the three-dimensional CAD models i and b;
wherein the similarity calculation formula is as follows:
wherein: tables i, b below identify two three-dimensional CAD models,
a represents the area of the three-dimensional figure in the view direction
P represents the perimeter of the three-dimensional figure in the view direction
C denotes the corner count of the three-dimensional figure in the view direction.
8. The method according to claim 6 or 7, characterized in that: if the material property characteristics need to be considered, the following steps are executed on the basis of the similarity median screening result:
and screening three-dimensional CAD models with the same material attributes from the screening results of the similarity serving as a median value in a preset span range, and then calculating and sorting the screened models according to the similarity.
9. A rapid process design method of a three-dimensional CAD model is characterized by being realized by the following modes:
storing an identification corresponding to a process design file corresponding to the three-dimensional CAD model in a database;
obtaining a three-dimensional CAD model similar to the target three-dimensional CAD model and a similarity ranking thereof according to the method of claim 6 or 7;
combining a specific three-dimensional CAD model structure according to the similarity sorting result, selecting a three-dimensional CAD model from the similarity sorting and obtaining a corresponding process design file identifier;
and pushing the corresponding process design file to serve as a basic file of the process design of the target three-dimensional CAD model according to the acquired process design file identification, and then combining a specific target three-dimensional CAD model on the basis of the basic file to further complete the process design file of the target three-dimensional CAD model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810551210.8A CN110555219B (en) | 2018-05-31 | 2018-05-31 | Three-dimensional CAD model similarity retrieval system and method based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810551210.8A CN110555219B (en) | 2018-05-31 | 2018-05-31 | Three-dimensional CAD model similarity retrieval system and method based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110555219A CN110555219A (en) | 2019-12-10 |
| CN110555219B true CN110555219B (en) | 2022-12-27 |
Family
ID=68733947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810551210.8A Active CN110555219B (en) | 2018-05-31 | 2018-05-31 | Three-dimensional CAD model similarity retrieval system and method based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110555219B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112001044B (en) * | 2020-07-31 | 2024-04-02 | 宁波智讯联科科技有限公司 | Method and system for estimating similarity of CAD model of part |
| CN112883443B (en) * | 2021-01-12 | 2022-10-14 | 南京维拓科技股份有限公司 | Method for judging similarity of part models based on geometry |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1614593A (en) * | 2004-10-14 | 2005-05-11 | 上海交通大学 | Similarity estimating method for three-dimensional CAD model based on shape |
| JP2006285627A (en) * | 2005-03-31 | 2006-10-19 | Hokkaido Univ | Device and method for retrieving similarity of three-dimensional model |
| CN101196930A (en) * | 2008-01-04 | 2008-06-11 | 覃征 | Three-dimensional model searching system |
| CN101408931A (en) * | 2007-10-11 | 2009-04-15 | Mv科技软件有限责任公司 | System and method for 3D object recognition |
| CN101916284A (en) * | 2010-08-19 | 2010-12-15 | 西北工业大学 | Three-dimensional model searching method based on shape orientation multi-resolution analysis |
| CN102568038A (en) * | 2011-12-28 | 2012-07-11 | 西北工业大学 | Method for translating dimensions on two-dimensional engineering drawing to three-dimensional CAD (computer aided design) model |
| CN104021255A (en) * | 2014-06-20 | 2014-09-03 | 上海交通大学 | Multi-resolution hierarchical presenting and hierarchical matching weighted comparison method for CAD (computer aided design) model |
| CN106844739A (en) * | 2017-02-14 | 2017-06-13 | 中国科学院遥感与数字地球研究所 | A kind of Remote Sensing Imagery Change information retrieval method based on neutral net coorinated training |
| CN106951460A (en) * | 2017-02-24 | 2017-07-14 | 北京宇航系统工程研究所 | A kind of MBD model retrieval methods based on figure matching |
| CN107066547A (en) * | 2017-03-20 | 2017-08-18 | 天津大学 | A kind of method for searching three-dimension model based on LIRE |
| CN108021683A (en) * | 2017-12-11 | 2018-05-11 | 首都航天机械公司 | A kind of scale model retrieval implementation method based on three-dimensional labeling |
| CN108052942A (en) * | 2017-12-28 | 2018-05-18 | 南京理工大学 | A kind of visual pattern recognition methods of aircraft flight attitude |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101350016B (en) * | 2007-07-20 | 2010-11-24 | 富士通株式会社 | Device and method for searching three-dimensional model |
| JP5654438B2 (en) * | 2011-11-29 | 2015-01-14 | 株式会社日立製作所 | Similar Design Structure Search Device and Similar Design Structure Search Method |
-
2018
- 2018-05-31 CN CN201810551210.8A patent/CN110555219B/en active Active
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1614593A (en) * | 2004-10-14 | 2005-05-11 | 上海交通大学 | Similarity estimating method for three-dimensional CAD model based on shape |
| JP2006285627A (en) * | 2005-03-31 | 2006-10-19 | Hokkaido Univ | Device and method for retrieving similarity of three-dimensional model |
| CN101408931A (en) * | 2007-10-11 | 2009-04-15 | Mv科技软件有限责任公司 | System and method for 3D object recognition |
| CN101196930A (en) * | 2008-01-04 | 2008-06-11 | 覃征 | Three-dimensional model searching system |
| CN101916284A (en) * | 2010-08-19 | 2010-12-15 | 西北工业大学 | Three-dimensional model searching method based on shape orientation multi-resolution analysis |
| CN102568038A (en) * | 2011-12-28 | 2012-07-11 | 西北工业大学 | Method for translating dimensions on two-dimensional engineering drawing to three-dimensional CAD (computer aided design) model |
| CN104021255A (en) * | 2014-06-20 | 2014-09-03 | 上海交通大学 | Multi-resolution hierarchical presenting and hierarchical matching weighted comparison method for CAD (computer aided design) model |
| CN106844739A (en) * | 2017-02-14 | 2017-06-13 | 中国科学院遥感与数字地球研究所 | A kind of Remote Sensing Imagery Change information retrieval method based on neutral net coorinated training |
| CN106951460A (en) * | 2017-02-24 | 2017-07-14 | 北京宇航系统工程研究所 | A kind of MBD model retrieval methods based on figure matching |
| CN107066547A (en) * | 2017-03-20 | 2017-08-18 | 天津大学 | A kind of method for searching three-dimension model based on LIRE |
| CN108021683A (en) * | 2017-12-11 | 2018-05-11 | 首都航天机械公司 | A kind of scale model retrieval implementation method based on three-dimensional labeling |
| CN108052942A (en) * | 2017-12-28 | 2018-05-18 | 南京理工大学 | A kind of visual pattern recognition methods of aircraft flight attitude |
Non-Patent Citations (4)
| Title |
|---|
| 基于视觉特征的农业机械三维模型相似性评价;赵秀艳 等;《山东农业大学学报(自然科学版)》;20160630;第47卷(第6期);第879-884页 * |
| 基于面属性相似的CAD模型检索方法;陶松桥 等;《武汉理工大学学报( 信息与管理工程版)》;20151030;第37卷(第5期);第399-403页 * |
| 面向制造领域的三视图模型组件快速检索方法研究;周燕 等;《中山大学学报( 自然科学版)》;20140731;第53卷(第4期);第62-68页 * |
| 面向设计重用的三维CAD模型局部结构检索方法;张开兴 等;《农业机械学报》;20170731;第48卷(第7期);第405-412页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110555219A (en) | 2019-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106126585B (en) | The unmanned plane image search method combined based on quality grading with perceived hash characteristics | |
| US7778462B2 (en) | Method of and apparatus for classifying an image | |
| CN101853299A (en) | Image searching result ordering method based on perceptual cognition | |
| CN109145964B (en) | Method and system for realizing image color clustering | |
| US20220375188A1 (en) | Semi-automatic image data labeling method, electronic apparatus, and storage medium | |
| CN110555219B (en) | Three-dimensional CAD model similarity retrieval system and method based on image recognition | |
| US20220414892A1 (en) | High-precision semi-automatic image data labeling method, electronic apparatus, and storage medium | |
| CN112580647A (en) | Stacked object oriented identification method and system | |
| CN108764302A (en) | A kind of bill images sorting technique based on color characteristic and bag of words feature | |
| CN115272204A (en) | Bearing surface scratch detection method based on machine vision | |
| WO2008142675A1 (en) | A method and a system for organizing an image database | |
| CN114863464A (en) | Second-order identification method for PID drawing picture information | |
| CN108052918A (en) | A kind of person's handwriting Compare System and method | |
| CN105404682A (en) | Digital image content based book retrieval method | |
| CN109062921A (en) | A kind of method and system for extracting ship pallet control information | |
| CN110019901A (en) | Three-dimensional model search device, searching system, search method and computer readable storage medium | |
| CN115546824B (en) | Taboo picture identification method, apparatus and storage medium | |
| CN103020630A (en) | Processing method and device of image features | |
| CN110580299A (en) | Method, system, device and storage medium for generating matching of recommendation language of object | |
| CN112818735B (en) | Article identification-based vessel spare part identification method | |
| CN112966727A (en) | Plant species identification system based on flower characteristics | |
| CN112016515A (en) | File cabinet vacancy detection method and device | |
| CN107273389A (en) | The querying method and device of trial video | |
| CN114241249B (en) | Image classification method and system based on target detection algorithm and convolutional neural network | |
| EP4189584A1 (en) | Automated annotation of visual data through computer vision template matching |
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 |