JP7440220B2 - Sheet metal processing estimate creation support device and sheet metal processing estimate creation support method - Google Patents

Description

The present invention relates to a sheet metal processing estimate creation support device and a sheet metal processing estimate creation support method.

Generally, a company that performs sheet metal processing is given product drawings and asked to submit a quotation by the client. For example, Patent Document 1 discloses a technique for simulating product manufacturing and creating an estimate when an order for sheet metal processing product manufacturing is received. In the technique disclosed in Patent Document 1, when an orderer requests a product manufacturing estimate from a contractor, the contractor creates a three-dimensional view of the product from a three-dimensional view of the product. The three-dimensional figure is divided into multiple parts so that processing can be performed properly.

An exploded view of each part is created. Material costs, blank processing costs, etc. are calculated from the developed drawing. The process of making holes is determined from each developed view. The processing cost for drilling holes is calculated from the hole drilling process. The bending process is determined from each developed view. The bending cost is calculated from the bending process. Parts are assembled and products are manufactured. Welding costs, painting costs, assembly costs, etc. are calculated based on the product. An estimate is created from the costs calculated as described above.

Japanese Patent Application Publication No. 2002-203007

By the way, in the above technique, drawings are referred to for each step of making holes, bending, assembling parts, etc., and costs are calculated for each step. Therefore, there is a drawback that it takes time to create an estimate. It is also conceivable to create an estimate in accordance with an order for manufacturing a new sheet metal processing product by referring to data on estimates for sheet metal processing performed in the past. However, in response to an order for the manufacture of a new sheet metal processed product, it is difficult to search through the huge amount of data on sheet metal processing that has been performed in the past, and currently estimates are created based on human memory and experience. Therefore, there is a drawback that it takes a lot of effort to prepare estimates for sheet metal processing.

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to provide a sheet metal processing estimate creation support device and a sheet metal processing estimate creation support method that can reduce the labor required to create a sheet metal processing estimate.

The present invention provides a database in which existing drawing data regarding sheet metal processing and estimation data regarding estimation of sheet metal processing based on the existing drawing data are stored in association with each other, and a new drawing data input reception unit that receives input of new drawing data regarding sheet metal processing. and a search for existing drawing data that is similar to the new drawing data related to the input received by the new drawing data input reception unit from the existing drawing data stored in the database using a similarity judgment model generated in advance by machine learning. a selection input reception unit that receives input for selecting existing drawing data searched by the search unit; and an estimate stored in the database in association with the existing drawing data selected by the input received by the selection input reception unit. This is a sheet metal processing estimate creation support device that includes an estimate data output unit that outputs data.

According to this configuration, in the sheet metal processing estimate creation support device, existing drawing data and estimate data regarding sheet metal processing are stored in association with each other in the database. The new drawing data input reception unit accepts input of new drawing data regarding sheet metal processing. The search unit searches existing drawing data stored in the database for existing drawing data similar to the new drawing data using a similarity determination model generated in advance by machine learning. The selection input reception unit accepts an input for selecting existing drawing data searched by the search unit. The estimate data output unit outputs estimate data stored in the database in association with the selected existing drawing data. As a result, by inputting new drawing data, existing drawing data similar to the new drawing data is searched, and the estimate data associated with the existing drawing data selected from the searched existing drawing data is output. It is possible to more easily refer to the sheet metal processing estimate data that has been provided, and it is possible to reduce the labor required to create a sheet metal processing estimate.

In this case, similarity judgment is performed by training a similarity judgment model using the new drawing data related to the input received by the new drawing data input reception unit and the existing drawing data selected by the input received by the selection input reception unit as teachers. Preferably, the apparatus further includes a model training section.

According to this configuration, the similarity determination model training unit trains the new drawing data related to the input received by the new drawing data input reception unit and the existing drawing data selected by the input received by the selection input reception unit. Since a similarity determination model is trained as follows, the trained similarity determination model can improve the accuracy with which existing drawing data similar to new drawing data is searched.

In addition, the database stores attribute data related to attributes of sheet metal processing based on existing drawing data, existing drawing data, and estimate data in association with each other, and allows input of attribute data related to attributes of sheet metal processing based on new drawing data. The search unit further includes an attribute data input reception unit for receiving data, and the search unit searches for new drawing data using a similarity determination model from existing drawing data stored in the database in association with attribute data related to the input received by the attribute data input reception unit. It is preferable to search for existing drawing data similar to the new drawing data related to the input received by the input reception unit.

According to this configuration, attribute data regarding attributes of sheet metal processing based on existing drawing data, existing drawing data, and estimate data are stored in association with each other in the database. The attribute data input reception unit accepts input of attribute data regarding attributes of sheet metal processing based on new drawing data. The search unit searches existing drawing data stored in the database in association with attribute data related to the input received by the attribute data input reception unit, using a similarity determination model to search for data related to the input received by the new drawing data input reception unit. Search for existing drawing data similar to new drawing data. That is, the search unit searches for existing drawing data similar to the new drawing data from the existing drawing data limited by the attribute data. Therefore, the effort required by the search unit to search existing drawing data similar to the new drawing data from the existing drawing data can be reduced.

In this case, the search unit has a plurality of similarity determination models generated in advance for each attribute data, and the existing drawing data stored in the database in association with the attribute data related to the input received by the attribute data input reception unit. , search for existing drawing data that is similar to the new drawing data related to the input received by the new drawing data input reception unit using the similarity determination model generated for the attribute data related to the input received by the attribute data input reception unit. It is preferable that

According to this configuration, the search unit has a plurality of similarity determination models generated in advance for each attribute data. The search unit generates attribute data related to the input received by the attribute data input reception unit from existing drawing data stored in the database in association with the attribute data related to the input received by the attribute data input reception unit. Existing drawing data similar to new drawing data is searched for using a similarity judgment model, that is, a similarity judgment model specialized for each attribute data. Therefore, compared to the case where a search is performed using a general similarity determination model for many attribute data, the search unit can improve the accuracy with which existing drawing data is searched for existing drawing data that is similar to new drawing data. .

The present invention also provides a new drawing data input receiving step in which the new drawing data input receiving unit of the sheet metal processing estimate creation support device receives input of new drawing data regarding sheet metal processing, and the search unit of the sheet metal processing estimate creation support device includes: Existing drawing data related to sheet metal processing and estimate data related to sheet metal processing estimates based on existing drawing data are stored in association with each other. From existing drawing data stored in a database, a similarity judgment model generated in advance by machine learning is used to generate new A search process for searching for existing drawing data similar to the new drawing data input received in the drawing data input reception process, and a selection input reception unit of the sheet metal processing estimate creation support device searching for existing drawing data searched in the search process. A selection input reception step that accepts an input for selecting a selection input, and an estimate data output unit of the sheet metal processing quotation creation support device are stored in the database in association with the existing drawing data selected by the input received in the selection input reception step. This is a sheet metal processing estimate creation support method comprising an estimate data output step of outputting estimate data.

In this case, the similarity determination model training unit of the sheet metal processing estimate creation support device uses new drawing data related to the input received in the new drawing data input reception process and existing drawings selected based on the input received in the selection input reception process. It is preferable to include a similarity determination model training step of training a similarity determination model using the data as a teacher.

In addition, the database stores attribute data regarding sheet metal processing attributes based on existing drawing data, existing drawing data, and quotation data in association with each other, and the attribute data input reception unit of the sheet metal processing quotation creation support device The search unit further includes an attribute data input reception step that receives input of attribute data regarding attributes of sheet metal processing based on new drawing data, and in the search step, the search unit receives input of attribute data associated with the input received in the attribute data input reception step. It is preferable to use a similarity determination model to search existing drawing data stored in the database for existing drawing data similar to the new drawing data related to the input received in the new drawing data input receiving step.

In this case, the search unit has a plurality of similarity determination models generated in advance for each attribute data, and in the search step, the search unit has a plurality of similarity determination models that are associated with the attribute data related to the input received in the attribute data input reception step. From the existing drawing data stored in the database, new drawing data related to the input received by the new drawing data input reception unit is created using the similarity judgment model generated for the attribute data related to the input received in the attribute data input reception process. It is preferable to search for similar existing drawing data.

According to the sheet metal processing estimate creation support device and the sheet metal processing estimate creation support method of the present invention, it is possible to reduce the labor required to create a sheet metal processing estimate.

FIG. 1 is a block diagram showing a sheet metal processing estimate creation support device according to an embodiment. FIG. 3 is a diagram illustrating an example of existing drawing data, attribute data, estimate data, and a similarity determination model stored in association with a database. FIG. 2 is a diagram schematically showing a similarity determination model. It is a flowchart which shows the process of the sheet metal processing estimate preparation support method of embodiment. It is a figure which shows the example of the display screen of an attribute data input reception process. It is a figure which shows the example of the display screen of a new drawing data input reception process. It is a figure which shows the example of a display screen after a search process. It is a figure which shows the example of the display screen of a selection input reception process. It is a figure showing an example of a display screen of an estimate data output process.

DESCRIPTION OF THE PREFERRED EMBODIMENTS A sheet metal processing estimate creation support device and a sheet metal processing estimate creation support method according to embodiments of the present invention will be described below with reference to the drawings. As shown in FIG. 1, the sheet metal processing estimate creation support device 1 according to the present embodiment includes a database 11, an attribute data input reception section 12, a new drawing data input reception section 13, a search section 14, a selection input reception section 15, an estimate It includes a data output section 16 and a similarity determination model training section 17. For example, the sheet metal processing estimate creation support device 1 of the present embodiment is configured to handle sheet metal processing when there is a request to create an estimate when sheet metal processing is performed according to new drawing data related to sheet metal processing in sheet metal processing estimation work. support the creation of quotations to reduce the effort required to create quotations.

The sheet metal processing estimate creation support device 1 is, for example, a computer device such as a personal computer having a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), a hard disk, a liquid crystal display, and the like. In the sheet metal processing estimate creation support device 1, a program stored in the ROM is loaded into the RAM and executed by the CPU, thereby executing the operations of each unit such as the search unit 14 described above. Note that the sheet metal processing estimate creation support device 1 may include a plurality of computer devices and storage devices. For example, the database 11, the attribute data input reception unit 12, the new drawing data input reception unit 13, the search unit 14, the selection input reception unit 15, the estimate data output unit 16, and the similarity determination model training unit 17 are installed on separate hard disks and personal computers. It may be configured from a computer device such as a computer.

The database 11 is a storage device such as a hard disk. As shown in FIG. 2, the database 11 stores existing drawing data related to sheet metal processing, attribute data related to sheet metal processing attributes based on existing drawing data, and quotation data related to estimates for sheet metal processing based on existing drawing data in association with each other. has been done. In the example of FIG. 2, in addition to existing drawing data, attribute data, and estimate data, a similarity determination model, which will be described later, is stored in association with the existing drawing data, attribute data, and estimate data.

The existing drawing data is, for example, data of a drawing for which a request was made to create an estimate when sheet metal processing was executed in the past according to the existing drawing data. The attributes of the attribute data include, for example, the person in charge of sheet metal processing based on existing drawing data, the machine name for sheet metal processing, the type of 5/6-axis CAM (computer aided manufacturing), the type of bending machine, material, and sheet. plate thickness, material size, part size, update date (update date of existing drawing data, attribute data, and estimate data), creation date (creation date of existing drawing data, attribute data, and estimate data), and The presence or absence of a processing process (the presence or absence of punching/bending/welding). The estimate data is, for example, data of an estimate when a request was made to create an estimate when sheet metal processing was executed in the past according to the existing drawing data. In the example of FIG. 2, existing drawing data having the same attributes except for the update date and creation date is shown.

The attribute data input reception unit 12 receives input of attribute data regarding attributes of sheet metal processing based on new drawing data. Note that the attribute data received by the attribute data input reception unit 12 does not necessarily have to be data related to all attributes of sheet metal processing based on new drawing data, but may be data related to some attributes of sheet metal processing based on new drawing data, for example. .

The new drawing data input reception unit 13 receives input of new drawing data regarding sheet metal processing. The new drawing data is, for example, drawing data when a request is made to create an estimate when sheet metal processing is executed according to new drawing data related to sheet metal processing in a sheet metal processing estimation service.

The search unit 14 uses a similarity judgment model M1, a similarity judgment model M2, a similarity judgment model M3 to a similarity judgment model Mn (n is a natural number of 4 or more) generated in advance by machine learning from existing drawing data stored in the database 11. ), searches for existing drawing data similar to the new drawing data related to the input received by the new drawing data input reception unit 13. The search unit 14 searches for new drawing data using similarity determination models M1, M2, M3 to Mn from existing drawing data stored in the database 11 in association with the attribute data input received by the attribute data input receiving unit 12. Existing drawing data similar to the new drawing data related to the input received by the input receiving unit 13 is searched.

The search unit 14 has a plurality of similarity determination models M1, M2, M3 to Mn generated in advance for each attribute data, and is associated with the attribute data related to the input received by the attribute data input reception unit 12 in the database 11. The data is received by the new drawing data input receiving unit 13 based on the similarity determination models M1, M2, M3 to Mn generated for the attribute data related to the input received by the attribute data input receiving unit 12 from the existing drawing data stored in the . Existing drawing data similar to the new drawing data related to the input is searched.

The similarity determination models M1, M2, M3 to Mn are information used to classify images of existing drawing data stored in the database 11 in association with attribute data corresponding to the similarity determination models M1, M2, M3 to Mn. . For example, the similarity determination model M2 for attribute data regarding attributes of machine name: XXX, material: AL-N2 (aluminum nitride), plate thickness: 1.5 mm, material size: 1524 mm x 3048 mm, and part size: 442 mm x 228 mm is as follows. Classification (similarity determination) is performed, such as whether a part shown in an image of drawing data is a rectangular part or an L-shaped part. In this embodiment, the classification destination (in the above example, rectangular parts, L-shaped parts, etc.) is called a category. For example, category C1 is a classification for rectangular parts, and category C2 is a classification for L-shaped parts.

The similarity determination models M1, M2, M3 to Mn receive information based on images of drawing data as input and output information indicating the category into which the image is classified. For example, the similarity determination models M1, M2, M3 to Mn output scores (similarities) for each category. The higher the numerical value of the score, the higher the possibility that the image will be classified into the category.

The similarity determination models M1, M2, M3 to Mn are configured including neural networks. For example, the similarity determination models M1, M2, M3 to Mn are models using deep learning. As shown in FIG. 3, each of the similarity determination models M1, M2, M3 to Mn is composed of a plurality of neurons (nodes) included in any of the input layer (first layer) L1, intermediate layer L2, and output layer L3. ing. The intermediate layer L2 includes neurons of multiple layers.

Each of the plurality of neurons included in the input layer L1 receives an input based on the image P of the drawing data. The information input to the neurons of the input layer L1 is generated, for example, as follows. First, the image P is divided into a fixed number of regions A. For example, the image P is divided vertically and horizontally into a preset number at equal intervals, and divided into rectangular areas A of the same size generated by the divisions. From the pixel values of the pixels in each of the rectangular areas A, information (vectors) to be input to the neurons of the input layer L1 is generated. For example, the average value, maximum value, etc. of pixel values in the region A are assumed to be information input to the neurons of the input layer L1.

As the information input to the input layer L1 is transmitted to deeper layers, features are extracted at each layer. That is, the neurons in each layer form a filter of a certain size, and information is slid over to extract multiple features for each filter. Information (numerical values) output from each of the plurality of neurons included in the output layer L3 indicates the score of each category.

A plurality of learning images and label information, which is information indicating the characteristics of the learning images, are input to the search unit 14, and machine learning is performed using the input learning images and label information to determine similarity. Models M1, M2, M3 to Mn are generated. The label information is information indicating the category of the learning image. That is, the learning image and the label information of the learning image are teachers for machine learning.

The learning images and label information are prepared by the administrator of the sheet metal processing estimate creation support device 1 and input into the search unit 14 . For example, when an image showing an L-shaped part is prepared as a learning image, the label information of the learning image is information indicating the category C2 of the L-shaped part. The search unit 14 performs deep learning using the learning image and the label information corresponding to the learning image, and generates similarity determination models M1, M2, M3 to Mn for each attribute data. Note that the similarity determination models M1, M2, M3 to Mn generated by the search unit 14 are used for the search of existing drawing data by the search unit 14, so the similarity determination models M1, M2, M3 to Mn generated by the search unit 14 are The generation is performed prior to the search unit 14 searching for existing drawing data.

The search unit 14 performs the above-described processing on the image P of the new drawing data related to the input received by the new drawing data input receiving unit 13, and outputs a score for each category. For example, for the category with the highest score of the image P of the new drawing data, the search unit 14 outputs the order of the highest score of the category of the existing drawing data stored in the database 11 as the order of similarity.

The selection input receiving section 15 receives an input for selecting existing drawing data searched by the search section. The estimate data output unit 16 outputs estimate data stored in the database 11 in association with the existing drawing data selected by the input received by the selection input receiving unit 15.

The similarity determination model training unit 17 uses the new drawing data related to the input received by the new drawing data input reception unit 13 and the existing drawing data selected based on the input received by the selection input reception unit 15 as teachers. The decision models M1, M2, M3 to Mn are trained. The training of the similarity judgment models M1, M2, M3 to Mn by the similarity judgment model training unit 17 is performed using, for example, an image of new drawing data related to the input received by the new drawing data input reception unit 13 as a learning image, and a selected image. Similarity determination models M1, M2, M3 to Mn are generated by the search unit 14, except that the label information indicating the category of the existing drawing data selected by the input received by the input reception unit 15 is the machine learning teacher. It is done in the same way.

Hereinafter, a sheet metal processing estimate creation support method using the sheet metal processing estimate creation support device of this embodiment will be described. As shown in FIG. 4, an attribute data input reception step is executed in which the attribute data input reception unit 12 of the sheet metal processing estimate creation support device 1 receives input of attribute data regarding attributes of sheet metal processing based on new drawing data (S1). .

As shown in FIG. 5, in the attribute data input reception step, an attribute item input reception frame 101, an attribute content input reception frame 102, and A new drawing data input reception frame 103 is displayed. In the attribute item input reception frame 101, attributes of sheet metal processing such as material size and material as described above are specified. In the attribute content input acceptance frame 102, the content of the attribute of sheet metal processing specified in the attribute item input acceptance frame 101 is specified. On the display screen 100, the attribute item input acceptance frame 101 and the attribute content input acceptance frame 102 can be deleted or added. In the example of FIG. 5, inputs of material size: 1524 mm x 3048 mm and material: AL-N2 are accepted as attribute data. The new drawing data input reception frame 103 is a frame used for inputting new drawing data in the new drawing data input reception process.

As shown in FIG. 4, a new drawing data input reception step is executed in which the new drawing data input reception unit 13 of the sheet metal processing estimate creation support device 1 receives input of new drawing data regarding sheet metal processing (S2). As shown in FIG. 6, in the new drawing data input reception process, the new drawing data file is dragged and dropped from the new drawing data window 104 showing the new drawing data file to the new drawing data input reception frame 103. As a result, input of new drawing data is accepted. Note that acceptance of input of new drawing data may be performed by a method other than dragging and dropping into the new drawing data input reception frame 103, such as directly opening a file displayed in the new drawing data window 104. .

As shown in FIG. 4, the search unit 14 of the sheet metal processing estimate creation support device 1 searches the database 11 in which existing drawing data regarding sheet metal processing and estimate data regarding sheet metal processing estimates based on the existing drawing data are stored in association with each other. Existing drawing data similar to the new drawing data related to the input received in the new drawing data input reception step 13 is determined by similarity determination models M1, M2, M3 to Mn generated in advance by machine learning from the stored existing drawing data. A search step is performed to search for (S3).

In the search step, the search unit 14 searches for similarity determination models M1, M2, M3 from existing drawing data stored in the database 11 in association with the attribute data related to the input received in the attribute data input reception step. . . . Mn to search for existing drawing data similar to the new drawing data related to the input received in the new drawing data input reception step. In other words, the search unit 14 does not search all the existing drawing data stored in the database 11, but searches a part of the existing drawing data corresponding to the attribute data related to the input received in the attribute data input reception step. do.

In the search process, the search unit 14 searches existing drawing data stored in the database 11 in association with attribute data related to the input received in the attribute data input reception process. Using the similarity determination models M1, M2, M3 to Mn generated for the attribute data, existing drawing data similar to the new drawing data related to the input received by the new drawing data input receiving unit 13 is searched. For example, as in the example in Fig. 5, if the input of material size: 1524 mm x 3048 mm and material: AL-N2 is accepted as attribute data, the search is executed using the similarity determination model M2 generated for the attribute data. be done.

The search unit 14 performs the above-described processing on the image P of the new drawing data using the similarity determination model M2 generated for the attribute data related to the input received in the attribute data input reception step, and outputs a score for each category. do. For the category with the highest score of the image P of the new drawing data, the search unit 14 searches the category of the existing drawing data stored in the database 11 in association with the attribute data related to the input received in the attribute data input reception step. Output in order of similarity in order of high score.

As shown in FIG. 7, after the search process, the new drawing data display frame 105 of the display screen 100 displays a drawing thumbnail of the new drawing data and attribute data regarding attributes of sheet metal processing based on the new drawing data. In addition, the search result display frame 106 of the display screen 100 displays drawing thumbnails of the existing drawing data, attribute data regarding sheet metal processing attributes based on the existing drawing data, and estimate data stored in association with the existing drawing data. are displayed in order of similarity.

As shown in FIG. 4, a selection input reception step is executed in which the selection input reception unit 15 of the sheet metal processing estimate creation support device 1 receives an input for selecting the existing drawing data retrieved in the search step (S4). As shown in FIG. 8, in the selection input receiving step, for example, existing drawing data that the user considers to be most similar is selected from the existing drawing data displayed in the order of similarity in the search result display frame 106.

As shown in FIG. 4, the quotation data output unit 16 of the sheet metal processing quotation creation support device 1 outputs quotation data stored in the database 11 in association with the existing drawing data selected by the input received in the selection input reception step. An estimate data output step is executed to output (S5). As shown in FIG. 9, in the estimate data output step, the drawing thumbnail of the new drawing data displayed in the new drawing data display frame 105 and the attribute data regarding the attributes of sheet metal processing based on the new drawing data are displayed in the estimate data output frame 107. , a thumbnail of the drawing of the existing drawing data selected by the input received in the selection input reception step, attribute data regarding attributes of sheet metal processing based on the existing drawing data, and estimate data stored in association with the existing drawing data. is displayed.

The estimate data output in the estimate data output step is used, for example, to create an estimate for sheet metal processing based on new drawing data by an arbitrary estimate creation program installed on a personal computer or the like that constitutes the sheet metal processing estimate creation support device 1. It is possible to do so. After creating an estimate for sheet metal processing using new drawing data, the new drawing data is stored in the database 11 as existing drawing data, with attribute data related to attributes of sheet metal processing based on the existing drawing data, existing drawing data, and estimate data being associated with each other. be remembered.

As shown in FIG. 4, the similarity judgment model training unit of the sheet metal processing estimate creation support device 1 uses new drawing data related to the input received in the new drawing data input reception process and the input received in the selection input reception process. A similarity determination model training step for training a similarity determination model using the selected existing drawing data as a teacher is executed (S6).

In this embodiment, in the sheet metal processing estimate creation support device 1, existing drawing data and estimate data related to sheet metal processing are stored in association with each other in the database 11. The new drawing data input reception unit 13 accepts input of new drawing data regarding sheet metal processing. The search unit 14 searches existing drawing data stored in the database 11 for existing drawing data similar to the new drawing data using similarity determination models M1, M2, M3 to Mn generated in advance by machine learning. The selection input acceptance unit 15 accepts an input for selecting existing drawing data searched by the search unit 14. The estimate data output unit 16 outputs the estimate data stored in the database 11 in association with the selected existing drawing data. As a result, by inputting new drawing data, existing drawing data similar to the new drawing data is searched, and the estimate data associated with the existing drawing data selected from the searched existing drawing data is output. It is possible to more easily refer to the sheet metal processing estimate data that has been provided, and it is possible to reduce the labor required to create a sheet metal processing estimate.

In other words, in response to orders for the production of new sheet metal processing products, as in the past, estimates are created based on human memory and experience while referring to a huge amount of data on sheet metal processing performed in the past. However, there are disadvantages in that it requires a lot of labor and that the estimated amount varies. On the other hand, in this embodiment, when new drawing data is input, existing drawing data similar to the new drawing data is searched, and estimation data associated with the existing drawing data is output, so the effort required to create an estimate for sheet metal processing is is significantly reduced. Moreover, since existing drawing data similar to the new drawing data is searched for within a certain range of accuracy, it is possible to improve the accuracy of the estimated amount.

In the present embodiment, the similarity determination model training unit 17 generates new drawing data related to the input received by the new drawing data input reception unit 13 and existing drawings selected based on the input received by the selection input reception unit 15. Since the similarity judgment models M1, M2, M3 to Mn are trained using the data as teachers, existing drawing data similar to the new drawing data is searched by the trained similarity judgment models M1, M2, M3 to Mn. Accuracy can be improved.

Further, in the present embodiment, the database 11 stores attribute data regarding attributes of sheet metal processing based on existing drawing data, existing drawing data, and estimate data in association with each other. The attribute data input accepting unit 12 accepts input of attribute data regarding attributes of sheet metal processing based on new drawing data. The search unit 14 searches for new drawing data using similarity determination models M1, M2, M3 to Mn from existing drawing data stored in the database 11 in association with the attribute data input received by the attribute data input receiving unit 12. Existing drawing data similar to the new drawing data related to the input received by the input receiving unit 13 is searched. In other words, the search unit 14 searches for existing drawing data similar to the new drawing data from the existing drawing data limited by the attribute data. Therefore, the effort required by the search unit 14 to search existing drawing data similar to the new drawing data from the existing drawing data can be reduced.

In particular, when the search unit 14 searches for existing drawing data that is similar to new drawing data using the similarity judgment models M1, M2, M3 to Mn, the search unit 14 uses the similarity judgment models M1, M2, M3 to Mn to search for existing drawing data that is similar to new drawing data. However, it is difficult to determine similarity regarding the size and materials of objects. Therefore, by searching for existing drawing data similar to new drawing data from existing drawing data whose size, material, etc. are limited by attribute data, it is possible to improve the accuracy of the search.

Further, according to the present embodiment, the search unit 14 has a plurality of similarity determination models M1, M2, M3 to Mn generated in advance for each attribute data. The search unit 14 retrieves attribute data related to the input received by the attribute data input receiving unit 12 from existing drawing data stored in the database 11 in association with attribute data related to the input received by the attribute data input receiving unit 12. Existing drawing data similar to the new drawing data is searched using the similarity judgment models M1, M2, M3 to Mn generated for each attribute data, that is, the similarity judgment models M1, M2, M3 to Mn specialized for each attribute data. Therefore, compared to the case where a search is performed using a general similarity determination model for many attribute data, the search unit can improve the accuracy with which existing drawing data is searched for existing drawing data that is similar to new drawing data. .

It should be noted that the present invention is not limited to the embodiments described above, and it goes without saying that various changes can be made without departing from the gist of the embodiments of the present invention.

DESCRIPTION OF SYMBOLS 1... Sheet metal processing estimate creation support device, 11... Database, 12... Attribute data input reception unit, 13... New drawing data input reception unit, 14... Search unit, 15... Selection input reception unit, 16... Estimate data output unit, 17 ...Similarity determination model training section, 100...Display screen, 101...Attribute item input acceptance frame, 102...Attribute content input acceptance frame, 103...New drawing data input acceptance frame, 104...New drawing data window, 105...New drawing data display Frame, 106... Search result display frame, 107... Estimate data output frame, M1, M2, M3, Mn... Similarity determination model, P... Image, A... Area, L1... Input layer, L2... Middle layer, L3... Output layer , C1, C2...Category.

Claims (8)

a database in which existing drawing data regarding sheet metal processing and estimation data regarding estimation of sheet metal processing based on the existing drawing data are stored in association with each other;
a new drawing data input reception unit that accepts input of new drawing data related to sheet metal processing;
From the existing drawing data stored in the database, one or more of the existing drawing data that are similar to the new drawing data related to the input received by the new drawing data input reception unit are determined by a similarity determination model generated in advance by machine learning from the existing drawing data stored in the database. a search unit that searches for drawing data and displays the searched one or more existing drawing data on a display ;
a selection input reception unit that receives an input for selecting one of the existing drawing data from among the one or more existing drawing data searched by the search unit and displayed on the display ;
an estimate data output unit that outputs the estimate data stored in the database in association with the one existing drawing data selected by the input received by the selection input receiver;
A sheet metal processing estimate creation support device equipped with The similarity determination model is created using the new drawing data related to the input received by the new drawing data input reception unit and the one existing drawing data selected by the input received by the selection input reception unit as teachers. The sheet metal processing estimate creation support device according to claim 1, further comprising a similarity determination model training unit for training. In the database, attribute data regarding attributes of sheet metal processing based on the existing drawing data, the existing drawing data, and the estimate data are stored in association with each other,
further comprising an attribute data input reception unit that receives input of the attribute data regarding attributes of sheet metal processing based on the new drawing data,
The search unit searches the existing drawing data stored in the database in association with the attribute data related to the input received by the attribute data input reception unit, using the similarity determination model to search the new drawing data input reception unit. 3. The method according to claim 1 or 2, wherein the one or more existing drawing data similar to the new drawing data related to the input received by is searched for, and the searched one or more existing drawing data is displayed on the display. Sheet metal processing estimate creation support device. The search section includes:
having a plurality of the similarity determination models generated in advance for each of the attribute data,
Generating the attribute data related to the input received by the attribute data input reception unit from the existing drawing data stored in the database in association with the attribute data related to the input received by the attribute data input reception unit. The one or more existing drawing data that are similar to the new drawing data related to the input received by the new drawing data input reception unit are searched using the similarity determination model that has been searched, and the one or more existing drawing data that have been searched are The sheet metal processing estimate creation support device according to claim 3, wherein the sheet metal processing estimate creation support device displays on the display . a new drawing data input reception step in which a new drawing data input reception unit of the sheet metal processing estimate creation support device receives input of new drawing data regarding sheet metal processing;
The search unit of the sheet metal processing estimate creation support device searches the existing drawing data stored in a database in which existing drawing data regarding sheet metal processing and estimate data regarding sheet metal processing estimates based on the existing drawing data are stored in association with each other. , searching for one or more of the existing drawing data that is similar to the new drawing data related to the input received in the new drawing data input reception step using a similarity determination model generated in advance by machine learning, and A search process of displaying the above existing drawing data on a display ,
A selection input reception unit of the sheet metal processing estimate creation support device receives an input for selecting one of the existing drawing data from among the one or more existing drawing data searched in the search step and displayed on the display. process and
The estimate data output unit of the sheet metal processing estimate creation support device outputs the estimate data stored in the database in association with the one existing drawing data selected by the input received in the selection input receiving step. Estimate data output process,
A sheet metal processing estimate creation support method. The similarity determination model training unit of the sheet metal processing estimate creation support device uses the new drawing data related to the input received in the new drawing data input receiving step and the new drawing data selected based on the input received in the selection input receiving step. 6. The sheet metal processing estimate creation support method according to claim 5, further comprising a similarity determination model training step of training said similarity determination model using one existing drawing data as a teacher. In the database, attribute data regarding attributes of sheet metal processing based on the existing drawing data, the existing drawing data, and the estimate data are stored in association with each other,
The attribute data input reception unit of the sheet metal processing estimate creation support device further comprises an attribute data input reception step of receiving input of the attribute data regarding attributes of sheet metal processing based on the new drawing data,
In the search step, the search unit searches the existing drawing data stored in the database in association with the attribute data related to the input received in the attribute data input reception step, using the similarity determination model to determine the new 5. Searching for the one or more existing drawing data similar to the new drawing data related to the input received in the drawing data input receiving step , and displaying the searched one or more existing drawing data on the display. Or the sheet metal processing estimate creation support method described in 6. The search unit has a plurality of similarity determination models generated in advance for each of the attribute data,
In the search step, the search unit selects the data received in the attribute data input reception step from the existing drawing data stored in the database in association with the attribute data related to the input received in the attribute data input reception step. searching for the one or more existing drawing data that is similar to the new drawing data related to the input received by the new drawing data input reception unit, using the similarity determination model generated for the attribute data related to the received input ; 8. The sheet metal processing estimate creation support method according to claim 7, wherein the retrieved one or more existing drawing data is displayed on the display .

Images