JP2022073144A - Computer program, article related information search device, and article related information search method - Google Patents

Abstract

To provide a computer program, an article related information search device and an article related information search method, with which it is possible to provide related information regarding articles.SOLUTION: A computer program causes a computer to execute the processes of acquiring drawing data in which a first article is specified, extracting a feature quantity regarding the first article on the basis of the acquired drawing data, calculating the degree of similarity between the extracted feature quantity regarding the first article and a preliminarily stored feature quantity regarding each of a plurality of second article, specifying one or a plurality of second articles from among the plurality of stored second articles on the basis of the calculated degree of similarity, and outputting related information regarding the specified second article(s).SELECTED DRAWING: Figure 1

Description

The present invention is a computer program, an article-related information search device, and an article-related information search method.

In recent years, many products and articles such as parts used in products are designed using a CAD (Computer Aided Design) system. Manufacturing of goods requires various costs such as fixed costs and variable costs, and it is common to make an estimate at the design stage of goods. In addition, if the estimated cost is high, it may be necessary to change the design.

Patent Document 1 discloses an apparatus for calculating an estimated cost of a product based on physical quantity data and processing type data of members included in the design drawing when creating a design drawing of the product.

Japanese Patent Application Laid-Open No. 2003-196344

However, in the device of Patent Document 1, the designer needs to operate the data related to the product or member included in the design drawing when the drawing of the design drawing is completed, and the estimated cost is calculated each time the product is designed. Work is required and complicated. Further, although the device of Patent Document 1 can provide an estimate, it cannot provide information on an article required at a stage such as a production process or a sales time.

The present invention has been made in view of such circumstances, and an object of the present invention is to provide a computer program capable of providing related information about an article, an article-related information search device, and an article-related information search method.

The computer program according to the embodiment of the present invention acquires drawing data in which the first article is specified on a computer, extracts features related to the first article based on the acquired drawing data, and extracts the extracted features. The degree of similarity between the feature amount of the first article and the feature amount of each of the plurality of second articles stored in advance is calculated, and based on the calculated similarity, one or more of the plurality of stored second articles are calculated. The process of identifying the second article and outputting the related information regarding the identified second article is executed.

The article-related information search device according to the embodiment of the present invention is characterized by an acquisition unit that acquires drawing data in which the first article is specified and a feature that extracts a feature amount related to the first article based on the acquired drawing data. The degree of similarity between the quantity extraction unit, the storage unit that stores the feature amount for each of the plurality of second articles, the extracted feature amount for the first article, and the feature amount for each of the stored second articles is calculated. A calculation unit, a specific unit that identifies one or a plurality of second articles from a plurality of stored second articles based on the calculated similarity, and an output unit that outputs related information regarding the specified second article. To prepare for.

In the article-related information search method according to the embodiment of the present invention, the drawing data in which the first article is specified is acquired, and the feature amount related to the first article is extracted and extracted based on the acquired drawing data. The degree of similarity between the feature amount of the first article and the feature amount of each of the plurality of second articles stored in advance is calculated, and based on the calculated similarity, one or more of the plurality of stored second articles are calculated. The second article is specified, and the related information about the specified second article is output.

According to the present invention, it is possible to provide related information regarding an article.

It is a schematic diagram of an example of the structure of the article-related information search system of this embodiment. It is a block diagram which shows an example of the structure of the feature amount extraction part. It is a schematic diagram which shows an example of the calculation method of a normal vector. It is a schematic diagram which shows an example of the coordinate system used by the designer of a product. It is a schematic diagram which shows an example of a coordinate conversion. It is a schematic diagram which shows an example of the cluster generation method of a normal vector. It is a schematic diagram which shows an example of the cluster of a normal vector. It is a schematic diagram which shows an example of the extraction method of a feature vector. It is a schematic diagram which shows an example of the structure of the feature amount DB. It is a schematic diagram which shows an example of the similarity calculation of a feature vector. It is a schematic diagram which shows an example of the structure of the related information DB. It is a schematic diagram which shows an example of a search screen. It is a schematic diagram which shows an example of a search result. It is a flowchart which shows an example of the procedure of the related information search process by the article-related information search device.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a schematic diagram of an example of the configuration of the article-related information search system of the present embodiment. The article-related information search system includes an article-related information search device 50. The terminal device 10 is connected to the article-related information search device 50 via the communication network 1. The drawing data DB 61, the feature amount DB 62, the cluster DB 63, and the related information DB 64 are connected to the article-related information search device 50. The article-related information search device 50 may be configured by a server, a personal computer (PC), or the like. The article-related information search device 50 may be distributed to a plurality of devices (for example, a server, a PC). The terminal device 10 is a terminal device used by a user such as a designer of a product (also referred to as an article), and a PC, a smartphone, or a tablet terminal can be used. In this specification, the product also includes the parts constituting the product.

When the user uploads the drawing data of the target product (first article) using the terminal device 10, the article-related information search device 50 acquires the drawing data. The article-related information search device 50 can output related information about a similar product (second article) similar to the target product to the terminal device 10 based on the acquired drawing data.

The article-related information search device 50 includes a control unit 51 that controls the entire device, a data acquisition unit 52, a feature amount extraction unit 53, a storage unit 54, an output unit 55, a similarity calculation unit 56, and a specific unit 57. The control unit 51 can be composed of a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), and the like.

The data acquisition unit 52 includes a communication module and the like, and can acquire drawing data from the terminal device 10. The drawing data can be, for example, three-dimensional CAD (Computer Aided Design) data, but is not limited to three-dimensional CAD data, and is, for example, two-dimensional CAD data, PDF (Portable Document Format) data, or the like. There may be. The control unit 51 can store the drawing data acquired via the data acquisition unit 52 in the drawing data DB 61.

The feature amount extraction unit 53 can extract the feature amount related to the target product based on the drawing data acquired by the data acquisition unit 52. The feature amount may be any feature amount that can be used in the process of searching for similar products from a large number of products. The control unit 51 can store the feature amount extracted by the feature amount extraction unit 53 in the feature amount DB 62 in association with the product from which the feature amount has been extracted. Details of feature extraction will be described later.

The storage unit 54 can be configured as a semiconductor memory, a hard disk, or the like, and can store necessary information such as processing results in each unit of the article-related information search device 50.

The similarity calculation unit 56 calculates the similarity between the feature amount of the target product and the feature amount of the product to be searched stored in the feature amount DB 62. In the feature amount DB 62, for example, a product that has already been designed and a feature amount related to the product are stored in association with each other. Details of the feature amount DB 62 will be described later.

The specific unit 57 identifies similar products similar to the target product based on the similarity calculated by the similarity calculation unit 56. When specifying similar products similar to the target product, for example, products having a similarity equal to or higher than a threshold value may be specified, or a predetermined number of products may be specified from those having a higher degree of similarity.

By accessing the related information DB 64, the control unit 51 can specify the related information about the similar product specified by the specific unit 57. The related information DB 64 stores the product and the related information about the product in association with each other. The related information is information necessary for the designed product when the user designs a new product, and can be information that the user needs to newly generate using drawing data or other data. .. Details of the related information DB 64 will be described later.

The output unit 55 includes a communication module and the like, and can output the related information specified by the control unit 51 to the terminal device 10. The terminal device 10 can acquire related information and display the acquired related information on the display panel.

With the above configuration, a user such as a designer simply prepares drawing data of the designed target product and uploads it to the article-related information search device 50, for example, a large number (for example, thousands of types) designed in the past. It is possible to identify similar products similar to the target product from among the products of the above, and to confirm related information about the identified similar products.

That is, the user can acquire related information of a similar product similar to the designed target product. The acquired related information can be used as an alternative as related information regarding the target product. In addition, the user does not need to generate the related information about the product from the beginning every time the product is designed, and the user can use the related information of the similar product (for example, make some corrections to the provided related information). (Just add), related information about the target product can be generated, which improves user convenience.

Next, the details of the feature amount extraction method will be described.

FIG. 2 is a block diagram showing an example of the configuration of the feature amount extraction unit 53. The feature amount extraction unit 53 includes a point cloud data generation unit 531, a normal vector calculation unit 532, a principal component analysis unit 533, a coordinate conversion unit 534, and a feature vector extraction unit 535.

The point cloud data generation unit 531 generates point cloud data based on the acquired three-dimensional CAD data. The point cloud data is, for example, set data of a large number of points on the product surface in XYZ coordinates, and can represent the position and shape of the product surface.

The normal vector calculation unit 532 calculates the normal vector related to the target product based on the point cloud data generated by the point cloud data generation unit 531.

FIG. 3 is a schematic diagram showing an example of a method for calculating a normal vector. FIG. 3A shows an example of point cloud data of the target product. In the figure, the set of points indicated by black circles is point cloud data, which is also referred to as point data. In FIG. 3, for convenience, the target product is a hollow annular body having a square cross section. As shown in FIG. 3B, the normal vector can be obtained for each polygon (triangle) specified by the point group data, and the unit vector in the direction perpendicular to the surface of each polygon is used as the normal vector. be able to. As shown in FIG. 3C, the normal vector of the target product is a set of unit vectors (direction from the origin is specified) starting from the origin in XYZ coordinates. The XYZ coordinate system shown in FIG. 3C can be determined by performing principal component analysis of the point cloud data of the target product, as will be described later.

FIG. 4 is a schematic diagram showing an example of a coordinate system used by a product designer. FIG. 4A is assumed to be a target product designed using CAD. Since the coordinate system can be determined by the designer when designing the same product, for example, a coordinate system as shown in FIGS. 4 (B) to 5 (E) can be used. Coordinate systems other than those shown in FIGS. 4 (B) to 4 (E) can also be used. As described above, in general, the coordinate system when the designer designs the product can be set independently by the designer. Even for the same product, if the coordinate system is different, the point cloud data will be different and the normal vector will be different.

The principal component analysis unit 533 performs principal component analysis on the point cloud data generated by the point cloud data generation unit 531 to specify the coordinate system of the target product. Principal component analysis identifies the variability of 3D point cloud data, that is, the three directions of the eigenvectors in the orientation that maximizes the variance. The coordinate system (XYZ) of the target product can be specified by the principal component analysis.

In this way, by principal component analysis, the product-specific coordinate system can be used as the reference coordinate system instead of the coordinate system adopted by the designer, and the influence of the difference in the coordinate system used at the time of design is eliminated. be able to.

The coordinate conversion unit 534 performs predetermined coordinate conversion on the coordinate system of the product specified by the principal component analysis unit 533 to generate a plurality of coordinate systems.

FIG. 5 is a schematic diagram showing an example of coordinate conversion. FIG. 5 illustrates six coordinate systems (XYZ) of the coordinate systems 1 to 6. The coordinate system 1 is a coordinate system (XYZ) specified by principal component analysis. The coordinate system 2 is obtained by performing an XYZ → YZX conversion on the coordinate system 1. The coordinate system 3 is obtained by performing an XYZ → ZXY conversion on the coordinate system 1. The coordinate system 4 is an inverted version of the XYZ of the coordinate system 1. The coordinate system 5 is an inverted version of the XYZ of the coordinate system 2. The coordinate system 6 is an inverted version of the XYZ of the coordinate system 3.

The normal vector calculation unit 532 can calculate the normal vectors {N1} to {N6} in each of the coordinate systems 1 to 6, including the coordinate system specified by the principal component analysis unit 533. The normal vector in each coordinate system can be calculated by projecting the normal vector calculated based on the point cloud data onto the coordinate axes in each coordinate system. In the example of FIG. 5, six coordinate systems are illustrated, but the number of coordinate systems is not limited to six. For example, three may be used as in the coordinate systems 1 to 3, and other numbers may be used. good. By increasing the number of coordinate systems, the accuracy of calculating the similarity can be improved, as will be described later.

Next, a feature vector extraction method will be described as an example of the feature quantity.

The feature vector extraction unit 535 can extract a feature vector as a feature quantity related to the target product based on the calculated normal vector. As shown in FIG. 5, when the coordinate system of the target product is coordinate system 1 to coordinate system 6, the normal vector is calculated for each coordinate system, so that the feature vector can also be extracted for each coordinate system. .. Hereinafter, a specific description will be given.

FIG. 6 is a schematic diagram showing an example of a method for generating a cluster of normal vectors. As shown in FIG. 6, for drawing data (three-dimensional CAD data) of a plurality of products appropriately selected in advance for generating a cluster, the same point cloud data generation and normal vector calculation as described above are performed. , And by performing preprocessing such as principal component analysis, the normal vector in the unique coordinate system of each of the plurality of products is calculated. The number of a plurality of products can be appropriately determined, and may be, for example, 10, 20, 30 or the like. The normal vectors of each of the plurality of products are collected as one set data, and the normal vectors are grouped into a plurality of clusters based on the similarity between the normal vectors. Machine learning can be used for clustering. The number of clusters can be about 5 to 10, but is not limited to this. The generated cluster can be stored (registered) in the cluster DB 63 together with the information of the normal vector constituting each cluster.

FIG. 7 is a schematic diagram showing an example of a cluster of normal vectors. In the example of FIG. 7, it is shown that five clusters C1, C2, C3, C4, and C5 are generated as a whole by grouping normal vectors similar to each other into one cluster.

FIG. 8 is a schematic diagram showing an example of a feature vector extraction method. The feature vector extraction unit 535 can classify the normal vector related to the target product into a plurality of clusters and extract the feature vector related to the target product based on the classification result. For example, as shown in FIG. 8, a plurality of clusters are designated as C1, C2, C3, C4, and C5. Let {N} be the set of normal vectors of the target binding, and let the number (frequency) of the normal vectors classified into each of the clusters C1 to C5 be n1, n2, n3, n4, and n5. Here, {N} = n1 + n2 + n3 + n4 + n5. The feature vector FV can be represented by FV = (n1, n2, n3, n4, n5). After obtaining the normal vector, the normal vector is normalized (the magnitude of the vector is set to 1) to obtain a feature vector.

The control unit 51 can store (register) the extracted feature vector in the feature amount DB 62 each time the feature vector is extracted. The feature vector registered in the feature amount DB 62 is not limited to the one extracted by the article-related information search device 50, and the feature vector extracted by another device may be registered.

FIG. 9 is a schematic diagram showing an example of the configuration of the feature amount DB 62. The feature amount DB 62 can be configured by associating, for example, a product number, a product name, drawing data (CAD data), a feature vector, design information, and the like. The feature vector includes the feature vector in each coordinate system exemplified in FIG. The design information includes information handled by the designer when designing the product (for example, design conditions, specifications, standards, characteristic data, test data, etc.). With the configuration as shown in FIG. 9, the product number P0001, the product name XXX, the drawing data A00100, and the design information D00200 associated with the feature vector can be specified based on the feature vector (for example, FV01101).

The similarity calculation unit 56 calculates the similarity between each feature vector in each coordinate system of the target product and each feature vector in each coordinate system of the plurality of searched products stored in the feature amount DB 62. Similar to the target product, the feature vector in each coordinate system of the product to be searched stored in the feature amount DB 62 is also generated by performing a predetermined coordinate conversion on the coordinate system specified by the principal component analysis and the specified coordinate system. It can be obtained by calculating the normal vector in the coordinate system.

FIG. 10 is a schematic diagram showing an example of calculating the similarity of feature vectors. The feature vectors in the coordinate systems 1 to 6 of the target product are FV11, FV12, FV13, FV14, FV15, and FV16. The feature vectors in the coordinate systems 1 to 6 of the product to be searched are FV21, FV22, FV23, FV24, FV25, and FV26. The similarity calculation unit 56 calculates the similarity for, for example, a combination (6 ways) of the feature vectors FV11 to FV16 and the feature vector FV21. As the similarity, for example, cosine similarity may be used. As the similarity between the feature vector of the target product and the feature vector of the searched product, the most similar value among the six similarities obtained can be adopted. When the same coordinate transformation is performed for the target product and the searched product, the similarity calculation is the same as when the coordinate transformation is not performed. Therefore, the number of coordinate systems generated is 6 as a result. You can use the street combinations.

By extracting the feature vectors in each coordinate system and calculating the similarity between the feature vectors of the target product and the searched product in each coordinate system, the similarity in each of the multiple coordinate systems will be calculated. , The degree of similarity can be calculated accurately.

Since the normal vector specifies the orientation of polygons on the product surface, the feature vector is a feature quantity that characterizes the orientation of the polygon surface and the number of polygons in the same orientation when the product surface is represented by polygons (triangles). Can be. This makes it possible to calculate the degree of similarity between products based on the shape and position of the surface of the product.

As illustrated in FIG. 9, when the feature vector of the searched product similar to the feature vector of the target product is specified, the similar product similar to the target product can be specified. By accessing the related information DB 64, the control unit 51 can specify the related information about the similar product specified by the specific unit 57.

FIG. 11 is a schematic diagram showing an example of the configuration of the related information DB 64. The related information can be classified into, for example, quotation information and detailed information in association with the product name. The quotation information includes, for example, information such as an estimated amount of money and an estimated condition. The detailed information includes, for example, the quotation date, the creator of the quotation, the negotiation result, and the like. If the product name is specified, the quotation information and detailed information of the product can also be specified. The related information is not limited to that illustrated in FIG.

Next, an example of screen display on the terminal device 10 will be described.

FIG. 12 is a schematic diagram showing an example of a search screen. The search screen is a screen for outputting a search request to the article-related information search device 50 when the user searches for a similar product similar to the target product. The search screen displays a display area for displaying the file name of the drawing data of the selected target product, a selection area for selecting the drawing data file of the target product, and a display area for displaying the image of the selected target product. File. Further, when specifying a similar product similar to the target product, it is possible to select how many of the higher-level similarities are specified, or how many or more similarities are specified. By operating the "search" icon, a search request is transmitted to the article-related information search device 50 together with the drawing data of the target product. Although not shown, only the drawing data of the target product may be uploaded to the article-related information search device 50 first, and the search request may be output later.

FIG. 13 is a schematic diagram showing an example of the search result. Search results include relevant information about similar products that are similar to the target product. On the search result screen, the ranking and drawings of similar products are displayed in descending order of similarity. In the example of the figure, product 1, product 2, product 3, and product 4 similar to the target product in the order of 1st, 2nd, 3rd, and 4th are shown. As for the number of similar products to be displayed, a predetermined number of similar products may be displayed from the one having the highest similarity, or the similar products having a similarity equal to or higher than the threshold value may be displayed. As shown in FIG. 13, the file name of the drawing data of the similar product, the estimated price of the similar product, and the degree of similarity can be displayed together with the drawing of the similar product.

As described above, the degree of similarity between the target product and similar products (strictly speaking, the degree of similarity between feature quantities or feature vectors) is displayed together with the order of similarity, so that the user can use similar products similar to the target product. It can be confirmed in order of degree.

Related information includes drawings of similar products and estimated information (eg, estimated amount). As a result, the user can easily search and confirm the drawings and estimation information of similar products similar to the target product. In addition, since the drawings of similar products can be confirmed, for example, it is possible to immediately grasp whether the quotation information of the similar product can be used as the quotation information of the target product or whether some correction work is required, which is convenient for the user. improves.

In addition, the related information includes drawing data (file) in which similar products are identified. As a result, the user does not need to generate drawing data of a similar product similar to the target product from the beginning, and can utilize the drawing data of the similar product (for example, CAD data). For example, when it becomes necessary to change the design of the target product, a part of the drawing data of a similar product can be utilized.

As shown in FIG. 13, detailed information of similar products can be displayed by operating the "details" icon. The detailed information is information as illustrated in FIG. 11, but is not limited to the example of FIG. 11, and may be set as appropriate. When the related information is the quotation information, the detailed information may include, for example, the quotation date, the quotation creator, the negotiation result after the quotation, and the like. This allows the user to easily see detailed information about similar products.

When the drawing data DB 61 is, for example, a database using an external cloud service, when the control unit 51 extracts the feature amount of the target product and stores it in the feature amount DB 62, the drawing of the target product stored in the drawing data DB 61. The data may be discarded. Since the drawing data may contain confidential information, by discarding the drawing data, it is possible to prevent the drawing data from being stored in an external cloud service for a long period of time and improve security.

When the drawing data of the target product is acquired, the control unit 51 extracts the feature amount of the target product based on the acquired drawing data, and stores the extracted feature amount in the feature amount DB 62. The control unit 51 determines whether or not an output request for related information about a similar product similar to the target product has been accepted, and if the output request is received, outputs the related information about the similar product using the stored feature amount. Can be done. That is, if the drawing data is acquired before the output request for the related information is received, the feature amount is first extracted and stored. As a result, the feature amount extraction, which requires a relatively long processing time, can be completed at an early stage as compared with the case where the feature amount is extracted by using the reception of the output request of the related information as a trigger. The time from receiving the output request to outputting the related information can be shortened.

FIG. 14 is a flowchart showing an example of the procedure of the related information search process by the article-related information search device 50. Hereinafter, for convenience, the main body of the process will be described as the control unit 51. The control unit 51 acquires drawing data of the target product (S11) and generates point cloud data based on the acquired drawing data (S12). The control unit 51 calculates a normal vector for each polygon (S13), performs principal component analysis on the group data (S14), and specifies the coordinate system of the target product.

The control unit 51 performs predetermined coordinate conversion on the specified coordinate system (S15), and calculates a normal vector for each generated coordinate system (S16). The control unit 51 classifies the normal vector into a plurality of predetermined clusters for each coordinate system, extracts the feature vector based on the classification result, and registers the normal vector in the feature amount DB 62 (S17).

The control unit 51 determines whether or not there is a search request from the terminal device 10 (S18), and if there is no search request (NO in S18), the process of step S18 is continued. When there is a search request (YES in S18), the control unit 51 calculates the degree of similarity between the extracted feature vector (feature vector of the target product) and the feature vector of the registered product (search target product) (YES). S19).

The control unit 51 identifies a product (similar product) similar to the target product based on the calculated similarity (S20), accesses the related information DB 64, and outputs the related information of the specified product (S21). End the process.

The article-related information search device 50 of the present embodiment can also be realized by using a general-purpose computer including a CPU (processor), a ROM, a RAM (memory), and the like. That is, as shown in FIG. 14, a computer program that defines the procedure of each process is loaded into a RAM (memory) provided in the computer, and the computer program is executed by the CPU (processor), thereby relating to articles on the computer. The information search device 50 can be realized. The computer program may be recorded on a recording medium and distributed, or may be installed on the computer via a network.

In the present embodiment, as the related information, the estimation information about similar products has been described as an example, but the related information is not limited to the estimation information. For example, it may include information necessary for the production process such as production planning, process management, inspection / inspection for similar products, mold design for similar products, mold management, jig design, etc., or product description, It may be information required at the time of sale or at a stage after sale such as instruction manual and maintenance management.

In the present embodiment, the feature vector has been described as an example of the feature amount, but the feature amount is not limited to the feature vector. For example, a feature moment that can calculate the area occupied by the area surrounded by the line, the center of gravity of the area, the perimeter of the contour, and the like may be used from the line diagram described in the drawing. Further, a large number of drawing data may be learned using a neural network, and the features extracted by the neural network may be used. In this case, two-dimensional CAD data, PDF data, or the like can be used as the drawing data.

In the computer program of the present embodiment, the drawing data in which the first article is specified is acquired by a computer, and the feature amount relating to the first article is extracted based on the acquired drawing data, and the extracted first article is extracted. The degree of similarity between the feature amount related to and the feature amount related to each of the plurality of second articles stored in advance is calculated, and based on the calculated similarity, one or a plurality of second articles are stored from the plurality of stored second articles. Is specified, and the process of outputting the related information regarding the specified second article is executed.

The article-related information search device of the present embodiment has an acquisition unit that acquires drawing data in which the first article is specified, and a feature amount extraction unit that extracts feature amounts related to the first article based on the acquired drawing data. And a storage unit that stores the feature amount of each of the plurality of second articles, and a calculation unit that calculates the similarity between the extracted feature amount of the first article and the feature amount of each of the stored second articles. , A specific unit that identifies one or a plurality of the second articles from the plurality of stored second articles based on the calculated similarity, and an output unit that outputs related information about the specified second article.

In the article-related information search method of the present embodiment, the drawing data in which the first article is specified is acquired, and the feature amount related to the first article is extracted based on the acquired drawing data, and the extracted first article is extracted. The degree of similarity between the feature amount related to and the feature amount related to each of the plurality of second articles stored in advance is calculated, and based on the calculated similarity, one or a plurality of second articles are stored from the plurality of stored second articles. Is specified, and related information regarding the specified second article is output.

The computer program acquires the drawing data in which the first article is identified. The first article is, for example, a newly designed article and an article to be searched. The computer program extracts the feature amount related to the first article based on the acquired drawing data. The feature amount may be any feature amount that can search for similar articles from a large number of articles. The computer program stores the features for each of the plurality of second articles. The second article may be, for example, an article that has already been designed, and an article whose feature amount has already been extracted. The second article is an article to be searched.

The computer program calculates the similarity between the feature amount of the first article and the feature amount of each of the stored second articles, and identifies one or more second articles based on the calculated similarity. When specifying the second article, for example, the second article having a similarity degree equal to or higher than the threshold value may be specified, or a predetermined number of second articles may be specified from those having a higher similarity degree. The computer program outputs relevant information about the identified second article. The related information is information necessary for the first article when the first article is newly designed, and information that the designer (user) needs to newly generate using drawing data and other data. can do.

With the above configuration, the designer simply prepares the drawing data in which the first article is specified, and for example, identifies a second article similar to the first article from a plurality of second articles designed in the past. It can and provides relevant information about the identified second article. That is, the designer can acquire the related information of the second article similar to the first article. The acquired related information can be used as an alternative as related information regarding the first article. Also, each time the designer designs the article, the work of generating the relevant information about the article from the beginning is no longer necessary, and the provided relevant information is used (for example, a slight modification to the provided related information). Since the related information about the first article can be generated (just by adding work), the convenience of the designer is improved.

In the computer program of the present embodiment, the CAD data in which the first article is specified is acquired by a computer, point group data is generated based on the acquired CAD data, and the first item is generated based on the generated point group data. A normal vector for one article is calculated, a feature vector as a feature amount for the first article is extracted based on the calculated normal vector, the extracted feature vector and a feature vector for a plurality of second articles stored in advance. The process of calculating the similarity with and is executed.

The computer program acquires the CAD data in which the first article is specified, and generates point cloud data based on the acquired CAD data. The point cloud data is, for example, set data of a large number of points on the surface of the article in XYZ coordinates, and can represent the position and shape of the surface of the article.

The computer program calculates the normal vector for the first article based on the generated point cloud data. For example, the unit vector in the direction perpendicular to the surface of each polygon (triangle) specified by the point cloud data can be a normal vector. The normal vector of the first article is a set of unit vectors (the direction from the origin is specified) starting from the origin in XYZ coordinates.

The computer program extracts a feature vector as a feature quantity for the first article based on the calculated normal vector. The feature vector can represent a distribution state specified by the direction, number, or the like of each normal vector of the first article. The computer program can calculate the degree of similarity between the feature vector of the first article and the feature vector of the second article. For example, cosine similarity can be used to calculate the similarity.

The feature vector can be a feature quantity that characterizes the orientation of the polygon surface or the number of polygons having the same orientation when the surface of the article is represented by polygons (triangles). With the above configuration, the degree of similarity between articles can be calculated based on the shape and position of the surface of the articles.

In the computer program of the present embodiment, the computer classifies the normal vector related to the first article based on a plurality of clusters in which the normal vectors related to the plurality of articles are clustered, and the first article is based on the classification result. Extract the feature vector related to, and execute the process.

A plurality of clusters obtained by clustering normal vectors for a plurality of articles are prepared in advance. Machine learning can be used for clustering, and the normal vectors can be grouped into a plurality of clusters based on the similarity between the normal vectors of a plurality of articles. A plurality of articles may be appropriately selected.

The computer program can classify the normal vector related to the first article into a plurality of clusters and extract the feature vector related to the first article based on the classification result. For example, let a plurality of clusters be C1, C2, C3, C4, and C5. Let {N} be the set of normal vectors of the first article, and let n1, n2, n3, n4, and n5 be the number of normal vectors classified into each of the clusters C1 to C5. Here, {N} = n1 + n2 + n3 + n4 + n5. The feature vector FV can be represented by FV = (n1, n2, n3, n4, n5).

In the computer program of the present embodiment, the computer performs principal component analysis on the generated point cloud data to specify the coordinate system of the first article, and the normal of the first article in the specified coordinate system. Calculate the vector and execute the process.

The computer program performs principal component analysis on the generated point cloud data to specify the coordinate system of the first article, and calculates the normal vector of the first article in the specified coordinate system. Principal component analysis identifies the variability of 3D point cloud data, that is, the three directions of the eigenvectors in the orientation that maximizes the variance. The coordinate system (XYZ) of the first article is specified by principal component analysis. In general, the coordinate system when the designer designs the article can be set by the designer. Even for the same article, if the coordinate system is different, the point cloud data will be different and the normal vector will be different. Therefore, by principal component analysis, instead of the coordinate system adopted by the designer, the unique coordinate system of the article can be used as the reference coordinate system, and the difference in the coordinate system at the time of design can be ignored.

The computer program of the present embodiment generates a plurality of coordinate systems by performing predetermined coordinate transformations on the specified coordinate system of the first article on the computer, and the first article in each generated coordinate system. The normal vector of the above is calculated, the feature vector in each coordinate system is extracted based on the calculated normal vector, and the extracted feature vector and a plurality of second articles stored in advance are generated by the coordinate transformation. The process of calculating the similarity with the feature vector in each coordinate system is executed.

The computer program performs a predetermined coordinate transformation on the coordinate system of the first article specified by the principal component analysis to generate a plurality of coordinate systems. The predetermined coordinate transformation includes, for example, the XYZ → YZX transformation, the XYZ → ZXY transformation, and the coordinate transformation that inverts each of the above-mentioned three coordinate systems with respect to the specified coordinate system XYZ. The generated coordinate system can be six kinds of coordinate systems.

The computer program calculates the normal vector of the first article in each generated coordinate system, and extracts the feature vector in each coordinate system based on the calculated normal vector. The normal vector in each coordinate system can be calculated by projecting the normal vector calculated based on the point cloud data onto the coordinate axes in each coordinate system. The computer program calculates the similarity between the feature vector in each coordinate system of the first article and the feature vector in each coordinate system of the plurality of second articles stored in advance.

Similar to the first article, the feature vector in each coordinate system of the second article is also the coordinate system specified by the principal component analysis and the coordinate system generated by performing a predetermined coordinate conversion on the specified coordinate system. It can be obtained by calculating the normal vector of the second article.

By extracting the feature vectors in each coordinate system and calculating the similarity between the feature vectors of the first article and the second article in each coordinate system, the similarity in each of the plurality of coordinate systems is calculated. Therefore, the similarity can be calculated accurately.

In the computer program of the present embodiment, the related information includes a drawing of the second article and estimation information.

Related information includes drawings of the second article and estimation information. Thereby, the designer can easily search and confirm the drawing and the estimation information of the second article similar to the first article. Further, since the drawing of the second article can be confirmed, for example, it is possible to immediately grasp whether the estimation information of the second article can be used as the estimation information of the first article or whether some correction work is required, and the designer can check it. Convenience is improved.

In the computer program of this embodiment, the relevant information includes drawing data in which the second article is identified.

The relevant information includes drawing data in which the second article is identified. As a result, the designer does not need to generate drawing data of the second article similar to the first article from the beginning, and can utilize the drawing data of the second article (for example, CAD data).

The computer program of the present embodiment causes a computer to execute a process of outputting each degree of similarity between the feature amount of the first article and the feature amount of the specified plurality of second articles together with the similarity order.

The computer program outputs each degree of similarity between the feature amount of the first article and the feature amount of the specified second article, together with the similarity order. Thereby, the designer can confirm the second article similar to the first article in the order of similarity.

The computer program of the present embodiment causes a computer to perform a process of receiving a request for output of detailed information about the specified second article and outputting detailed information about the second article based on the received output request.

When the computer program receives the request for outputting the detailed information about the specified second article, the computer program outputs the detailed information about the second article. The detailed information may be set as appropriate, and may be, for example, detailed information about the related information of the second article. When the related information is the quotation information, the detailed information may include, for example, the quotation date, the quotation creator, the negotiation result after the quotation, and the like. This allows the designer to easily confirm detailed information about the second article similar to the first article.

The computer program of the present embodiment causes a computer to execute a process of discarding drawing data including CAD data in which the first article is specified when the feature amount relating to the first article is extracted and stored.

When the computer program extracts and stores the feature amount related to the first article, the computer program discards the drawing data including the CAD data in which the first article is specified. Drawing data may also contain sensitive information. Therefore, when the drawing data including the acquired CAD data is stored by using an external cloud service, the stored drawing data is discarded after the feature amount is extracted and stored. As a result, it is possible to prevent drawing data from being stored in an external cloud service for a long period of time and improve security.

When the computer program of the present embodiment acquires drawing data including CAD data in which the first article is specified before receiving a request for output of related information regarding the second article to the computer, the acquired drawing data. When the feature amount of the first article is extracted and stored based on the above and the output request of the related information about the second article is received, the related information about the second article is output by using the stored feature amount. To execute the process.

When the first article acquires the drawing data including the specified CAD data, the computer program extracts and stores the feature amount of the first article based on the acquired drawing data. It is determined whether or not the output request for the related information regarding the second article has been accepted, and if the output request is accepted, the related information regarding the second article is output using the stored feature amount. As a result, if the drawing data is acquired before the output request for the related information is accepted, the feature amount is first extracted and stored. As a result, it is possible to complete the feature amount extraction, which requires a relatively long processing time, at an earlier stage than in the case of extracting the feature amount by using the reception of the output request of the related information as a trigger. It is possible to shorten the time from receiving the output request of to outputting the related information.

1 Communication network 10 Terminal equipment 50 Article-related information search device 51 Control unit 52 Data acquisition unit 53 Feature quantity extraction unit 54 Storage unit 55 Output unit 56 Similarity calculation unit 57 Specific unit 61 Drawing data DB
62 Feature DB
63 Cluster DB
64 Related information DB

Claims (13)

On the computer
Obtain the drawing data in which the first article is specified,
Based on the acquired drawing data, the feature amount related to the first article is extracted.
The degree of similarity between the extracted feature amount of the first article and the feature amount of each of the plurality of second articles stored in advance was calculated.
Based on the calculated similarity, one or more second articles are identified from the plurality of stored second articles.
Output related information about the specified second article,
A computer program that executes processing. On the computer
The CAD data in which the first article is specified is acquired, and the CAD data is obtained.
Generate point cloud data based on the acquired CAD data,
A normal vector for the first article is calculated based on the generated point cloud data, and the normal vector is calculated.
Based on the calculated normal vector, a feature vector as a feature quantity related to the first article is extracted.
The degree of similarity between the extracted feature vector and the feature vector for a plurality of second articles stored in advance is calculated.
The computer program according to claim 1, wherein the processing is executed. On the computer
A normal vector related to the first article is classified based on a plurality of clusters in which normal vectors related to a plurality of articles are clustered, and a feature vector related to the first article is extracted based on the classification result.
The computer program according to claim 2, wherein the processing is executed. On the computer
Principal component analysis was performed on the generated point cloud data to identify the coordinate system of the first article.
Calculate the normal vector of the first article in the specified coordinate system,
The computer program according to claim 2 or 3, wherein the process is executed. On the computer
A plurality of coordinate systems are generated by performing predetermined coordinate transformations on the specified coordinate system of the first article.
Calculate the normal vector of the first article in each generated coordinate system,
Based on the calculated normal vector, the feature vector in each coordinate system is extracted and
The degree of similarity between the extracted feature vector and the feature vector in each coordinate system generated by the coordinate transformation for the plurality of second articles stored in advance is calculated.
The computer program according to claim 4, wherein the processing is executed. The relevant information includes drawings and estimation information of the second article.
The computer program according to any one of claims 1 to 5. The relevant information includes drawing data in which the second article is identified.
The computer program according to any one of claims 1 to 6. On the computer
Each degree of similarity between the feature amount of the first article and the feature amount of the specified second article is output together with the similarity order.
The computer program according to any one of claims 1 to 7, wherein the processing is executed. On the computer
Accepting a request to output detailed information about the specified second item,
Outputs detailed information about the second article based on the received output request.
The computer program according to any one of claims 1 to 8, wherein the processing is executed. On the computer
When the feature amount relating to the first article is extracted and stored, the drawing data including the CAD data specified by the first article is discarded.
The computer program according to any one of claims 1 to 9, wherein the processing is executed. On the computer
When the first article acquires drawing data including the specified CAD data before accepting the output request of the related information regarding the second article, the feature amount of the first article is determined based on the acquired drawing data. Extract and memorize,
When a request for output of related information regarding the second article is received, the stored feature amount is used to output the related information regarding the second article.
The computer program according to any one of claims 1 to 10, wherein the processing is executed. The acquisition unit that acquires the drawing data in which the first article is specified,
A feature amount extraction unit that extracts a feature amount related to the first article based on the acquired drawing data, and a feature amount extraction unit.
A storage unit that stores features for each of the plurality of second articles,
A calculation unit that calculates the degree of similarity between the extracted feature amount of the first article and the feature amount of each of the plurality of stored second articles.
A specific unit that identifies one or a plurality of second articles from a plurality of stored second articles based on the calculated similarity.
It is provided with an output unit that outputs related information regarding the specified second article.
Goods-related information search device. Obtain the drawing data in which the first article is specified,
Based on the acquired drawing data, the feature amount related to the first article is extracted.
The degree of similarity between the extracted feature amount of the first article and the feature amount of each of the plurality of second articles stored in advance was calculated.
Based on the calculated similarity, one or more second articles are identified from the plurality of stored second articles.
Output related information about the specified second article,
Article-related information search method.

Images