Drawing processing device, inference device, machine learning device, drawing processing method, inference method, and machine learning method

Abstract

[Problem] To provide a drawing processing device capable of appropriately selecting a request recipient for processing or estimation of a processed product described in a drawing, without depending on the knowledge and experience of a user. [Solution] A drawing processing device 2A comprises: a drawing reception unit 202 that receives new drawing data D10 about a processed product; and a request recipient selection unit 203A that selects a candidate for a request recipient that is requested to process or estimate the processed product of the new drawing data D10, on the basis of the new drawing data D10 received by the drawing reception unit 202. The request recipient selection unit 203A acquires processed product feature data D11 with respect to the new drawing data D10 by inputting the new drawing data D10 into a first learning model 220A, acquires processed product request data D12 with respect to the new drawing data D10 by inputting the processed product feature data D11 into a second learning model 220B, and selects a candidate for the request recipient on the basis of the processed product request data D12.

Description

Drawing processing device, inference device, machine learning device, drawing processing method, inference method, and machine learning method

[0001] The present invention relates to a drawing processing device, an inference device, a machine learning device, a drawing processing method, an inference method, and a machine learning method.

[0002] Drawings used in various fields such as machinery, architecture, civil engineering, electricity, and apparel contain, for example, lines (external lines, dimension lines, etc.) that represent the shape and structure of processed products, as well as various types of information written in text. Since the various types of information written in drawings are important for understanding the details of processing, systems have been developed to utilize this information. For example, Patent Document 1 discloses a support system that acquires text that represents dimensional quantities written accompanying the dimension lines of a figure and creates a list showing the dimensional quantities of the figure.

[0003] Japanese Patent Application Publication No. 10-293777

[0004] The support system disclosed in Patent Document 1, as described above, acquires information about processed products from drawings, but it only goes so far as to acquire information from drawings. Therefore, a user of the support system must select a supplier to request processing or a supplier to request an estimate prior to processing from the information described in the drawings. The task of selecting an appropriate supplier requires advanced knowledge and experience in processing, and is a highly difficult task.

[0005] The present invention has been made in response to the above-mentioned problems, and aims to provide a drawing processing device, an inference device, a machine learning device, a drawing processing method, an inference method, and a machine learning method that can appropriately select a party to request processing or an estimate for a processed product described in a drawing, without relying on the user's knowledge or experience.

[0006] In order to achieve the above-mentioned object, a drawing processing device according to one embodiment of the present invention comprises a drawing reception unit that receives new drawing data relating to a processed product, and a requestee selection unit that selects candidate requestees to process or estimate the processed product in the new drawing data based on the new drawing data received by the drawing reception unit.

[0007] According to a drawing processing apparatus according to one aspect of the present invention, it is possible to appropriately select a party to request processing or quotation for a processed product shown in a drawing, without relying on the user's knowledge or experience.

[0008] Problems, configurations, and effects other than those described above will become apparent from the detailed description of the invention that follows.

[0009] 1 is an overall configuration diagram showing an example of a drawing management system 1. FIG. 2 is a diagram showing an example of drawing data D10 related to an assembly drawing. FIG. 3 is a diagram showing an example of drawing data D10 related to a part drawing. FIG. 4 is a diagram showing an example of a drawing database 210. FIG. 5 is a block diagram showing an example of a drawing processing apparatus 2A according to a first embodiment. FIG. 6 is a functional explanatory diagram showing an example of a learning model generation unit 201A according to the first embodiment. FIG. 7 is a functional explanatory diagram showing examples of a drawing reception unit 202 and a request destination selection unit 203A according to the first embodiment. FIG. 8 is a hardware configuration diagram showing an example of a computer 900. FIG. 9 is a flowchart showing an example of the operation (drawing processing method) of the drawing processing apparatus 2A according to the first embodiment. FIG. 10 is a block diagram showing an example of a drawing processing apparatus 2B according to a second embodiment. FIG. 11 is a functional explanatory diagram showing an example of a learning model generation unit 201B according to the second embodiment. FIG. 12 is a functional explanatory diagram showing examples of a drawing reception unit 202 and a request destination selection unit 203B according to the second embodiment. FIG. 13 is a flowchart showing an example of the operation (drawing processing method) of the drawing processing apparatus 2B according to the second embodiment. FIG. 14 is a block diagram showing an example of a drawing processing apparatus 2C according to a third embodiment. FIG. 15 is a functional explanatory diagram showing examples of a drawing reception unit 202 and a request destination selection unit 203C according to the third embodiment. 11 is a flowchart showing an example of the operation (drawing processing method) of a drawing processing apparatus 2C according to the third embodiment.

[0010] Hereinafter, an embodiment for carrying out the present invention will be described with reference to the drawings. The scope necessary for the explanation to achieve the object of the present invention will be schematically shown, and the scope necessary for explaining the relevant parts of the present invention will be mainly explained, and the parts that are omitted from the explanation will be based on publicly known techniques.

[0011] 1 is a diagram showing an overall configuration of an example of a drawing management system 1. The drawing management system 1 functions as a system for managing drawing data D10 relating to processed products.

[0012] The drawing data D10 handled by the drawing management system 1 is, for example, any drawing recorded as digital data, such as mechanical drawings such as assembly drawings and parts drawings, architectural drawings, electrical circuit drawings, pneumatic circuit drawings, hydraulic circuit drawings, and apparel drawings. In this case, the drawing data D10 may be vector-format data or raster-format data. For example, the drawing data D10 may be CAD data (an example of vector format) output by various CAD software, or image data (an example of raster format) output by scanning a drawing printed on paper using a scanner or the like. The drawing data D10 may be drawn using any projection method or may be a three-dimensional drawing. The drawing data D10 may also include tables, and may be, for example, documents such as slips recorded as digital data.

[0013] 1, the drawing management system 1 includes a drawing processing device 2A and a user terminal device 3. The drawing processing device 2A and the user terminal device 3 are connected to a wired or wireless network 4 and are configured to be able to send and receive various data to and from each other. Note that the number of drawing processing devices 2A and user terminal devices 3 and the connection configuration of the network 4 are not limited to the example in FIG. 1 and may be changed as appropriate.

[0014] The drawing processing device 2A is a server-type computer or a cloud-type computer, and is configured as a general-purpose or dedicated computer (see FIG. 7 described later). The drawing processing device 2A accepts new drawing data D10 (hereinafter referred to as new drawing data D10) from the user terminal device 3 and performs various processes on the new drawing data D10. The drawing processing device 2A is provided with a drawing database 210 that can associate the new drawing data D10 with processed product characteristic data D11 and processed product request data D12 obtained as a result of processing the new drawing data D10 and register them in the drawing database 210. The drawing processing device 2A also provides the user terminal device 3 with display information for referencing and editing existing drawing data D10 (hereinafter referred to as existing drawing data D10) registered in the drawing database 210, and the processed product characteristic data D11 and processed product request data D12 associated with the existing drawing data D10.

[0015] The user terminal device 3 is a client-type computer, and is configured by a general-purpose or dedicated computer (see FIG. 7 described later), etc. The user terminal device 3 accepts various input operations via a display screen of an application, a browser, etc., in order to register new drawing data D10, refer to and edit existing drawing data D10, etc., and outputs various information via the display screen or voice.

[0016] Fig. 2A is a diagram showing an example of drawing data D10 relating to an assembly drawing, and Fig. 2B is a diagram showing an example of drawing data D10 relating to a part drawing.

[0017] The drawing area 10 of the drawing data D10 includes a shape area 11, a title block 12, and a parts list 13. The drawing area 10 corresponds to the entire area of ​​the paper when the drawing data D10 is printed on a paper medium.

[0018] The shape area 11 is an area where the shape and dimensions of an assembly or part are described using graphics and text. Lines defining the shape and structure of an assembly or part, such as outline lines, dimension lines, hidden lines, center lines, and imaginary lines, are described in the shape area 11. Text indicating dimensions, tolerances, part numbers, and the like is also described in the shape area 11. Note that the shape area 11 is not limited to six-sided views based on orthographic projection, and may represent other types of drawings, such as cross-sectional views, perspective views, and exploded views.

[0019] The title block 12 and the parts list 13 are types of tables included in the drawing data D10. A table has vertical and horizontal ruled lines and boxes separated by the ruled lines. Characters are written in the boxes, and attributes of the characters written in the box are defined for each box. Boxes are also classified by properties that indicate the type of box. Examples of properties include attribute heading properties, serial number heading properties, attributeless field properties, attributed field properties, etc.

[0020] As shown in FIGS. 2A and 2B , the attributes of the title block 12 include, for example, product name, drawing number, scale, creation date, designer, approver, etc. Each box in the title block 12 is classified into an attributed field property, an attribute header property, and an attributeless field property. The boxes classified as attributed field properties contain field characters (in the example of FIG. 2A , “1:2,” “2024 / 1 / 11,” “ASSY A,” and “F1-222-33-A”) as the contents of the title block 12, and attribute characters (in the example of FIG. 2A , “scale,” “creation date,” “product name,” and “drawing number”) that define the attributes of the field characters. The boxes classified as attribute header properties contain attribute characters (in the example of FIG. 2A , “designer” and “approver”) that define the attributes of the field characters. The boxes classified as attributeless field properties contain field characters (in the example of FIG. 2A , “AAA” and “BBB”) as the contents of the title block 12.

[0021] As shown in FIG. 2A , attributes of the BOM 13 include, for example, number, item name, material, and quantity. The boxes in the first row from the top of the BOM 13 are classified as attribute header properties, and attribute header characters defining the attributes of each column of the BOM 13 as headers (in the example of FIG. 2A , "Number," "Item Name," "Material," and "Quantity") are written therein. The boxes in the first column on the left side of the BOM 13 are classified as serial number header properties, and serial number header characters defining the serial numbers of each row of the BOM 13 as headers (in the example of FIG. 2A , "1," "2," and "3") are written therein. The other boxes in the BOM 13 are classified as attributeless field properties, which do not contain attribute characters, and field characters representing the contents of the BOM 13 (in the example of FIG. 2A , "COVER," "PPP," "1," "BODY," "QQQ," "1," "LEG," "RRR," and "4") are written therein.

[0022] The types of tables included in the drawing data D10 are not limited to the title block 12 and the parts list 13, but may be other types of tables. The arrangement and number of boxes constituting a table may be changed as appropriate depending on the type of table, and the attributes defined for each box are not limited to the above examples.

[0023] 3 is a diagram showing an example of the drawing database 210. In the drawing database 210, existing drawing data D10, processed product characteristic data D11 related to the processed product in the existing drawing data D10, and processed product request data D12 are registered in association with each existing drawing data D10.

[0024] The processed product characteristic data D11 is data indicating the characteristics of the processed product in the existing drawing data D10. The processed product characteristic data D11 is obtained from lines, characters, symbols, etc. included in the shape region 11 and characters (mainly field characters) included in the title block 12 and parts list 13. Attributes related to the characteristics of the processed product include, but are not limited to, the creation date, product name, drawing number, scale, shape, material, maximum dimension, quantity, and hole processing, as shown in FIG. 3 . Attributes related to hole processing include, but are not limited to, multiple attributes combining the hole diameter and the processing method (drilling, countersinking, countersinking, etc.).

[0025] For each attribute included in the processed product characteristic data D11, an attribute value in a data format corresponding to the attribute is registered. For attributes related to the product name and drawing number, a character string is registered as the attribute value. For attributes related to scale, shape, material, surface treatment, and welding, a value selected from a plurality of options prepared in advance is registered as the attribute value. In the example of FIG. 3, for example, "plate" is selected and registered for the shape from options such as plate, bar, and pipe. For attributes related to maximum dimension, quantity, and hole drilling, a numerical value is registered as the attribute value. In the example of FIG. 3, for example, the numerical value "200" is registered for the maximum dimension.

[0026] The processed product request data D12 is data indicating a requestee to process or estimate the processed product in the existing drawing data D10. The processed product request data D12 includes attributes related to the requestee, such as the name of the requestee to process or estimate, the date and time of the request for processing or estimate, the processing price or estimated price, etc.

[0027] For each attribute included in the processed product request data D12, an attribute value in a data format corresponding to the attribute is registered. Attributes relating to the destination of processing or quotation are selected from a destination list in which destinations are registered in advance, and for example, an identifier uniquely assigned to each destination is registered. It is also possible to add new destinations to the destination list or delete existing destinations.

[0028] In addition, the drawing database 210 may be configured to allow data other than those described above to be registered, and the data formats of the attributes and attribute values ​​contained in the processed product characteristic data D11 and the processed product request data D12 are not limited to the examples described above.

[0029] Fig. 4 is a block diagram showing an example of a drawing processing device 2A according to the first embodiment. Fig. 5 is a functional explanatory diagram showing an example of a learning model generation unit 201A according to the first embodiment. Fig. 6 is a functional explanatory diagram showing an example of a drawing reception unit 202 and a request destination selection unit 203A according to the first embodiment.

[0030] The drawing processing device 2A includes a control unit 20, a data storage unit 21, a trained model storage unit 22, a communication unit 23, an input unit 24, and an output unit 25.

[0031] The communication unit 23 is connected to an external device (e.g., the user terminal device 3, etc.) via the network 4 and functions as a communication interface for transmitting and receiving various types of data. The input unit 24 accepts various input operations, and the output unit 25 functions as a user interface by outputting various types of information via a display screen or voice. Note that the input unit 24 and the output unit 25 may be omitted.

[0032] The data storage unit 21 stores a drawing database 210 and a drawing processing program 211A. As shown in Fig. 3, existing drawing data D10 and various information obtained from the existing drawing data D10 are registered in the drawing database 210 in association with each other. The specific configuration of the drawing database 210 is not limited to the example shown in Fig. 3 and may be designed as appropriate.

[0033] The trained model storage unit 22 stores a trained first training model 220A and a trained second training model 220B. The training models 220A and 220B stored in the trained model storage unit 22 may be provided to other devices via the network 4, a recording medium, or the like. The number of training models 220A and 220B stored in the trained model storage unit 22 is not limited to one each, and multiple training models 220A and 220B with different conditions, such as differences in machine learning methods or data, may be stored and used selectively or in parallel.

[0034] 3, the data storage unit 21 and the trained model storage unit 22 are shown as two storage units, but they may be configured as a single storage unit or three or more storage units. Furthermore, at least one of the data storage unit 21 and the trained model storage unit 22 may be configured as a storage unit of an external computer (e.g., a server-type computer or a cloud-type computer).

[0035] The control unit 20 functions as a transmission / reception control unit 200, a learning model generation unit 201A, a drawing reception unit 202, a request destination selection unit 203A, and a database management unit 204 by executing the drawing processing program 211A recorded in the data storage unit 21.

[0036] (Transmission / reception control unit 200) The transmission / reception control unit 200 transmits and receives various types of data to and from an external device (for example, the user terminal device 3, etc.). For example, the transmission / reception control unit 200 transmits display information to the user terminal device 3 for outputting various display screens on the user terminal device 3, and receives operation information from the user terminal device 3 for accepting input operations performed on the display screen of the user terminal device 3. In doing so, the transmission / reception control unit 200 cooperates with the respective units 201 to 204 to transmit display information to the user terminal device 3 and receive operation information from the user terminal device 3.

[0037] (Learning Model Generation Unit 201A) As shown in FIGS. 4 and 5, the learning model generation unit 201A includes a learning data acquisition unit 2010A and a machine learning unit 2011A.

[0038] The learning data acquisition unit 2010A references the drawing database 210 and acquires first learning data D13A and second learning data D13B, each of which is composed of input data and output data. The learning data D13A and D13B are used as teacher data (training data), verification data, and test data in supervised learning. The output data constituting the learning data D13A and D13B is used as a correct answer label in supervised learning.

[0039] The input data constituting the first learning data D13A includes the existing drawing data D10. The existing drawing data D10 included in the input data of the first learning data D13A may be data corresponding to the entire drawing area 10, or may be data corresponding to a partial area (e.g., shape area 11) cut out from the drawing area 10.

[0040] The output data constituting the first learning data D13A includes processed product characteristic data D11 that indicates the characteristics of the processed product in the existing drawing data D10. In this case, the processed product characteristic data D11 included in the output data of the first learning data D13A may relate to some of the multiple attributes related to the characteristics of the processed product shown in FIG.

[0041] In addition, when a selection from a plurality of pre-prepared options, such as shape, material, surface treatment, or welding, is specified, the processed product characteristic data D11 included in the output data of the first learning data D13A is defined as data in which only the flag for the selected option is assigned "1." For example, when "plate" is selected from options such as plate, bar, and tube for shape, the processed product characteristic data D11 included in the output data of the first learning data D13A is defined as data in which only the flag for "plate" is assigned "1," such as "plate = 1, bar = 0, tube = 0."

[0042] Furthermore, when a value such as a maximum dimension, quantity, or hole processing is specified as a numerical value, the processed product characteristic data D11 included in the output data of the first learning data D13A is defined as data into which the specified numerical value (which may be a normalized numerical value) is substituted. For example, when "200" is specified for the maximum dimension, the processed product characteristic data D11 included in the output data of the first learning data D13A is defined as data into which "200" is substituted for the maximum dimension parameter, such as "maximum dimension = 200." Note that when the upper limit is normalized to "1000," the processed product characteristic data D11 may be defined as data into which "0.2" is substituted for the maximum dimension parameter, such as "maximum dimension = 0.2."

[0043] The input data constituting the second learning data D13B includes processed product characteristic data D11 indicating the characteristics of the processed product in the existing drawing data D10, and is defined in the same manner as the processed product characteristic data D11 included in the output data of the first learning data D13A. Note that, like the processed product characteristic data D11 included in the output data of the first learning data D13A, the processed product characteristic data D11 included in the input data of the second learning data D13B may relate to some of the multiple attributes related to the characteristics of the processed product shown in Figure 3.

[0044] The output data constituting the second learning data D13B includes processed product request data D12 indicating a requestee to process or estimate the processed product in the existing drawing data D10. The processed product characteristic data D11 included in the output data of the second learning data D13B relates to at least the requestee to process or estimate, among the multiple attributes related to the requestee shown in FIG.

[0045] In this case, when multiple request destinations are registered in advance as a request destination list for the request destination, such as "S001, S002, ..., S011, S012, ..., S020," and "S011" is specified as the request destination, the processed product characteristic data D11 included in the output data of the second learning data D13B is defined as data in which "1" is assigned to only the flag for "S011," such as "S001 = 0, S002 = 0, ..., S011 = 1, S012 = 0, ..., S020 = 0."

[0046] The learning data acquisition unit 2010A acquires first learning data D13A by referring to existing drawing data D10 registered in the drawing database 210 and processed product characteristic data D11 associated with the existing drawing data D10. The learning data acquisition unit 2010A also acquires second learning data D13B by referring to the processed product characteristic data D11 registered in the drawing database 210 and processed product request data D12 associated with the processed product characteristic data D11.

[0047] For example, the learning data acquisition unit 2010A displays on a display screen a list of existing drawing data D10 registered in the drawing database 210 in the drawing processing device 2A or the user terminal device 3, and acquires the first learning data D13A and the second learning data D13B by reading the existing drawing data D10 specified on the display screen and the processed product characteristic data D11 and processed product request data D12 associated with the existing drawing data D10 from the drawing database 210. The learning data acquisition unit 2010A may acquire the first learning data D13A and the second learning data D13B in cooperation with an external device connected via the network 4.

[0048] The machine learning unit 2011A performs machine learning to make the first learning model 220A learn the correlation between the input data (existing drawing data D10) and the output data (processed product characteristic data D11) using a plurality of sets of first learning data D13A acquired by the learning data acquisition unit 2010A. The machine learning unit 2011A also performs machine learning to make the second learning model 220B learn the correlation between the input data (processed product characteristic data D11) and the output data (processed product request data D12) using a plurality of sets of second learning data D13B acquired by the learning data acquisition unit 2010A.

[0049] Examples of learning models 220A and 220B that can be used include neural network types (including deep learning) such as convolutional neural networks, recurrent neural networks, and vision transformers, tree types such as decision trees and regression trees, ensemble learning such as bagging and boosting, clustering types such as hierarchical clustering, non-hierarchical clustering, k-nearest neighbors, and k-means, multivariate analyses such as principal component analysis, factor analysis, and logistic regression, and support vector machines.

[0050] The timing at which the machine learning unit 2011A performs machine learning may be when the number of existing drawing data D10 registered in the drawing database 210 exceeds a predetermined number, or when instructions are received from the user, but is not limited to these.

[0051] (Drawing Receiving Unit 202) The drawing receiving unit 202 receives new drawing data D10, for example, by receiving the drawing data D10 from the user terminal device 3 via the transmission / reception control unit 200. At that time, the drawing receiving unit 202 may receive a designation as to whether to select a party to request processing or a party to request an estimate.

[0052] (Requestee Selection Unit 203A) The requestee selection unit 203A selects candidate requestees to process or estimate the processed product in the new drawing data D10, based on the new drawing data D10 received by the drawing receiving unit 202. For example, if the drawing receiving unit 202 receives a designation of a requestee to process or estimate, the requestee selection unit 203A selects either a requestee to process or a requestee to estimate, based on the designation. Note that there may be one or more candidate requestees.

[0053] Specifically, the request destination selection unit 203A performs a feature acquisition process to acquire processed product feature data D11 indicating the features of the processed product in the new drawing data D10 from the new drawing data D10 accepted by the drawing acceptance unit 202, and a candidate selection process to acquire processed product request data D12 indicating the destination to be requested for the processed product in the new drawing data D10 based on the processed product feature data D11 acquired in the feature acquisition process, and selects candidate destinations based on the processed product request data D12.

[0054] In this embodiment, the request destination selection unit 203A inputs new drawing data D10 accepted by the drawing acceptance unit 202 into the first learning model 220A as a feature acquisition process, thereby acquiring processed product feature data D11 for the processed product in the new drawing data D10.

[0055] At this time, for attributes defined by options, the first learning model 220A outputs a score for each attribute value, for example, for shape, such as "plate = 0.9, rod = 0.04, pipe = 0.06," so that "plate," which has the highest score, is acquired as the shape for the new drawing data D10. Also, for attributes defined by numerical values, for example, for maximum dimension, it outputs "maximum dimension = 250" (or "maximum dimension = 0.2" when normalized), so that "250" (or "0.2" when normalized) is acquired as the maximum dimension for the new drawing data D10.

[0056] In addition, as a candidate selection process, the request destination selection unit 203A inputs the processed product feature data D11 acquired in the feature acquisition process into the second learning model 220B to acquire processed product request data D12 for the new drawing data D10 accepted by the drawing acceptance unit 202, and selects candidate request destinations based on the processed product request data D12.

[0057] At this time, the second learning model 220B outputs a score for each of the multiple request recipients included in the request recipient list, such as "S001 = 0.12, S002 = 0.08, ..., S011 = 0, S012 = 0, ..., S020 = 0.8." Therefore, the request recipient assigned the identifier "S020" with the highest score is selected as the candidate request recipient. When multiple candidate request recipients are selected, the request recipients may be selected in descending order of score up to a predetermined rank, or a request recipient with a score equal to or greater than a predetermined reference value may be selected.

[0058] (Database management unit 204) The database management unit 204 associates the processed product characteristic data D11 and processed product request data D12 acquired as a result of the feature acquisition process and complementary selection process for the new drawing data D10 with the new drawing data D10 and registers them in the drawing database 210. The database management unit 204 then transmits display information for displaying the processing results of the requestee selection unit 203A to the user terminal device 3 via the transmission / reception control unit 200. When candidate requestees (or multiple candidate requestees) based on the processed product request data D12 are displayed on the user terminal device 3, the database management unit 204 may accept an input operation from the user to instruct the final decision on the requestee. In this case, the database management unit 204 may transmit request information related to the processing request or estimate request to the requestee corresponding to the input operation via the transmission / reception control unit 200.

[0059] Furthermore, the database management unit 204 reads out the existing drawing data D10 registered in the drawing database 210 and transmits display information for displaying the existing drawing data D10 to the user terminal device 3 via the transmission / reception control unit 200. Then, when the database management unit 204 receives operation information for editing the existing drawing data D10 from the user terminal device 3 via the transmission / reception control unit 200, it modifies the existing drawing data D10 registered in the drawing database 210.

[0060] 7 is a hardware configuration diagram showing an example of a computer 900. The drawing processing device 2A and the user terminal device 3 in the drawing management system 1 are configured by a general-purpose or dedicated computer 900.

[0061] 7, the computer 900 includes, as its main components, a bus 910, a processor 912, a memory 914, an input device 916, an output device 917, a display device 918, a storage device 920, a communication I / F (interface) unit 922, an external device I / F unit 924, an I / O (input / output) device I / F unit 926, and a media input / output unit 928. Note that the above components may be omitted as appropriate depending on the application of the computer 900.

[0062] The processor 912 is composed of one or more arithmetic processing devices (such as a CPU (Central Processing Unit), an MPU (Micro-processing unit), a DSP (Digital Signal Processor), or a GPU (Graphics Processing Unit)), and operates as a control unit that controls the entire computer 900. The memory 914 stores various data and programs 930, and is composed of, for example, a volatile memory (DRAM, SRAM, etc.) that functions as a main memory, a non-volatile memory (ROM), a flash memory, etc.

[0063] The input device 916 is composed of, for example, a keyboard, a mouse, a numeric keypad, an electronic pen, a microphone, etc., and functions as an input unit. The output device 917 is composed of, for example, a sound (audio) output device, a vibration device, etc., and functions as an output unit. The display device 918 is composed of, for example, a liquid crystal display, an organic EL display, electronic paper, a projector, etc., and functions as an output unit. The input device 916 and the display device 918 may be integrated, such as a touch panel display. The storage device 920 is composed of, for example, an HDD, an SSD, etc., and functions as a storage unit. The storage device 920 stores various data necessary for executing the operating system and the program 930.

[0064] The communication I / F unit 922 is connected to a network 940 such as the Internet or an intranet (which may be the same as network 4 in FIG. 1) via a wired or wireless connection and functions as a communication unit that transmits and receives data to and from other computers in accordance with a predetermined communication standard. The external device I / F unit 924 is connected to an external device 950 such as a camera, printer, scanner, or reader / writer via a wired or wireless connection and functions as a communication unit that transmits and receives data to and from the external device 950 in accordance with a predetermined communication standard. The I / O device I / F unit 926 is connected to I / O devices 960 such as various sensors and actuators and functions as a communication unit that transmits and receives various signals and data, such as detection signals from sensors and control signals to actuators, to and from the I / O devices 960. The media input / output unit 928 is composed of a drive device such as a DVD drive or a CD drive and reads and writes data from and to media (non-transitory storage media) 970 such as DVDs and CDs.

[0065] In the computer 900 having the above configuration, the processor 912 loads a program 930 stored in the storage device 920 into the memory 914, executes the program, and controls each unit of the computer 900 via the bus 910. The program 930 may be stored in the memory 914 instead of the storage device 920. The program 930 may be recorded on the medium 970 in an installable file format or an executable file format and provided to the computer 900 via the media input / output unit 928. The program 930 may be provided to the computer 900 by downloading it via the communication I / F unit 922 over the network 940. Furthermore, the computer 900 may implement various functions realized by the processor 912 executing the program 930 using hardware such as an FPGA or an ASIC.

[0066] The computer 900 is an electronic device of any type, such as a desktop computer or a portable computer. The computer 900 may be a client computer, a server computer, a cloud computer, or an embedded computer such as a control panel or a controller (including a microcomputer, a programmable logic controller, or a sequencer).

[0067] (Operation of drawing processing device 2A) Figure 8 is a flowchart showing an example of the operation (drawing processing method) of the drawing processing device 2A according to the first embodiment. Note that the series of drawing processing methods by the drawing processing device 2A shown in Figure 8 will be described as being executed when the drawing processing device 2A receives new drawing data D10 from a user terminal device 3 operated by a user. Also, the description will be given assuming that the learned model storage unit 22 stores a trained first learning model 220A and a second learning model 220B as a result of the learning model generation unit 201A performing a learning data acquisition step, a machine learning step, and a trained model storage step as shown in Figure 5.

[0068] First, in step S100 (drawing reception step), the drawing reception unit 202 receives new drawing data D10 from the user terminal device 3, thereby accepting the new drawing data D10. At this time, the drawing reception unit 202 may accept a plurality of pieces of new drawing data D10.

[0069] Next, in step S110 (request destination selection process), the request destination selection unit 203A inputs the new drawing data D10 accepted in step S100 into the first learning model 220A as a feature acquisition process, thereby acquiring processed product feature data D11 for the processed product in the new drawing data D10. Note that if the drawing acceptance unit 202 accepts multiple pieces of new drawing data D10, it acquires processed product feature data D11 for each of the new drawing data D10.

[0070] Next, in step S111 (request destination selection process), the request destination selection unit 203A inputs the processed product characteristic data D11 acquired in step S110 into the second learning model 220B as a candidate selection process to acquire processed product request data D12 for the processed product in the new drawing data D10 accepted in step S100. The request destination selection unit 203A then selects the candidate request destinations indicated by the processed product request data D12 as candidate request destinations to process or estimate the processed product in the new drawing data D10 accepted in step S100. Note that if the drawing acceptance unit 202 accepts multiple pieces of new drawing data D10, it acquires processed product request data D12 for each piece of new drawing data D10 and selects candidate request destinations.

[0071] Next, in step S120 (database management process), the database management unit 204 associates the processed product characteristic data D11 and processed product request data D12 obtained as the processing results of steps S110 and S111 with the new drawing data D10 accepted in step S100 and registers them in the drawing database 210.

[0072] Then, in step S130, the database management unit 204 transmits display information for displaying the processing results of steps S110 and S111 to the user terminal device 3. As a result, the processed product characteristic data D11 and processed product request data D12 registered for the new drawing data D10 are displayed on the display screen of the user terminal device 3.

[0073] As described above, according to the drawing processing device 2A and drawing processing method of this embodiment, by using the first learning model 220A and the second learning model 220B that have been trained on past performance when selecting a contractor for existing drawing data D10, it is possible to appropriately select a contractor to process or estimate the processed product described in the drawing without relying on the user's knowledge or experience.

[0074] Second Embodiment Fig. 9 is a block diagram showing an example of a drawing processing device 2B according to a second embodiment. Fig. 10 is a functional explanatory diagram showing an example of a learning model generation unit 201B according to the second embodiment. Fig. 11 is a functional explanatory diagram showing an example of a drawing reception unit 202 and a request destination selection unit 203B according to the second embodiment.

[0075] The second embodiment differs from the first embodiment in that the drawing processing device 2B uses one learning model 220 instead of two learning models 220A and 220B to select candidate contractors for the processed product in the new drawing data D10. The following describes the drawing processing device 2B according to the second embodiment, focusing on the differences from the first embodiment.

[0076] The trained model storage unit 22 stores a trained learning model 220.

[0077] The control unit 20 functions as a transmission / reception control unit 200, a learning model generation unit 201B, a drawing reception unit 202, a request destination selection unit 203B, and a database management unit 204 by executing the drawing processing program 211B recorded in the data storage unit 21.

[0078] (Learning Model Generation Unit 201B) As shown in FIGS. 9 and 10, the learning model generation unit 201B includes a learning data acquisition unit 2010B and a machine learning unit 2011B.

[0079] The learning data acquisition unit 2010B refers to the drawing database 210 and acquires the learning data D13.

[0080] The input data constituting the learning data D13 includes existing drawing data D10. The existing drawing data D10 included in the input data of the learning data D13 is the same data as the existing drawing data D10 included in the input data of the first learning data D13A in the first embodiment, and therefore a detailed description thereof will be omitted.

[0081] The output data constituting the learning data D13 includes processed product request data D12 indicating a request destination for processing or estimating the processed product in the existing drawing data D10. The processed product request data D12 included in the output data of the learning data D13 is the same data as the processed product request data D12 included in the output data of the second learning data D13B in the first embodiment, and therefore a detailed description thereof will be omitted.

[0082] The learning data acquisition unit 2010B acquires learning data D13 by referring to the existing drawing data D10 registered in the drawing database 210 and the processed product request data D12 associated with the existing drawing data D10. For example, as in the first embodiment, the learning data acquisition unit 2010A acquires learning data D13 by displaying a list of the existing drawing data D10 registered in the drawing database 210 on a display screen of the drawing processing device 2B or the user terminal device 3, and reading out the existing drawing data D10 specified on the display screen and the processed product request data D12 associated with the existing drawing data D10 from the drawing database 210.

[0083] The machine learning unit 2011B performs machine learning using multiple sets of learning data D13 acquired by the learning data acquisition unit 2010B to have the learning model 220 learn the correlation between the input data (existing drawing data D10) and the output data (processed product request data D12).

[0084] (Request destination selection unit 203B) The request destination selection unit 203B inputs the new drawing data D10 accepted by the drawing acceptance unit 202 into the first learning model 220A, thereby obtaining processed product request data D12 for the processed product in the new drawing data D10.

[0085] (Operation of drawing processing device 2B) Fig. 12 is a flowchart showing an example of the operation (drawing processing method) of drawing processing device 2B according to the second embodiment. Note that, since the steps having the same step numbers as those in the flowchart shown in Fig. 8 are the same as those in the first embodiment, the following description will mainly focus on the parts that are different from the first embodiment.

[0086] First, in step S100 (drawing reception step), the drawing reception unit 202 receives new drawing data D10.

[0087] Next, in step S112 (request destination selection process), the request destination selection unit 203B acquires processed product request data D12 for the processed product in the new drawing data D10 by inputting the new drawing data D10 accepted in step S100 into the learning model 220. Then, the request destination selection unit 203A selects candidate request destinations indicated by the processed product request data D12 as candidate request destinations to process or estimate the processed product in the new drawing data D10 accepted in step S100.

[0088] Next, in step S121 (database management step), the database management unit 204 associates the processed product request data D12 acquired as a result of the processing in step S112 with the new drawing data D10 accepted in step S100 and registers the data in the drawing database 210. Then, in step S131, the database management unit 204 transmits display information for displaying the processing result of step S112 to the user terminal device 3.

[0089] As described above, according to the drawing processing device 2B and drawing processing method of this embodiment, by using the learning model 220 that has learned past performance when selecting a contractor for existing drawing data D10, it is possible to appropriately select a contractor to process or estimate the processed product described in the drawing, without relying on the user's knowledge or experience.

[0090] Third Embodiment Fig. 13 is a block diagram showing an example of a drawing processing apparatus 2C according to a third embodiment. Fig. 14 is a functional explanatory diagram showing an example of a drawing receiving unit 202 and a request destination selecting unit 203C according to the third embodiment.

[0091] The third embodiment differs from the first embodiment in that the drawing processing device 2C selects candidate destinations for the processed product in the new drawing data D10 by performing the feature acquisition process and the candidate selection process without using the first learning model 220A and the second learning model 220B. The following describes the drawing processing device 2C according to the third embodiment, focusing on the differences from the first embodiment.

[0092] The control unit 20 executes the drawing processing program 211C recorded in the data storage unit 21, thereby functioning as a transmission / reception control unit 200, a drawing reception unit 202, a request destination selection unit 203C, and a database management unit 204.

[0093] (Request destination selection unit 203C) As a feature acquisition process, the request destination selection unit 203C performs character recognition processing on the new drawing data D10 accepted by the drawing acceptance unit 202 to acquire processed product feature data D11 indicating the features of the processed product in the new drawing data D10.

[0094] The destination selection unit 203C performs a known optical character recognition process (OCR) as the character recognition process, reading characters included in the drawing data D10 as text data to obtain the processed product characteristic data D11. Furthermore, if text data is embedded in the drawing data D10, the destination selection unit 203C reads the text data to obtain the characters included in the drawing data D10 as the processed product characteristic data D11. The results of the character recognition process include text data indicating the characters and position data indicating the position where the characters are written.

[0095] For example, if the character "welding" is read, the workpiece characteristic data D11 is defined as data in which only the flag for "welding" is set to "1," such as "yes" = "1, no" = "0." Furthermore, if the character "4.5 drill" is read in eight locations, the workpiece characteristic data D11 is defined as data in which "8" is assigned to the parameter for 4.5 drilling, such as "4.5 drilling = 8," for the attribute of drilling. In this case, if the characters read by the character recognition process contain a mixture of characters that can be treated as synonyms, such as half-width and full-width characters or uppercase and lowercase characters, they may be treated as a single character. The characters read by the character recognition process may be displayed on the display screen of the user terminal device 3, where they may be edited by the user.

[0096] In addition, as a candidate selection process, the request destination selection unit 203C refers to the drawing database 210 and, based on the processed product characteristic data D11 for the new drawing data D10 acquired in the characteristic acquisition process, extracts existing drawing data D10 that is identical to or similar to the characteristics indicated by the processed product characteristic data D11 from the existing drawing data D10 registered in the drawing database 210 as similar drawing data D10.

[0097] For example, for attributes selected by options, existing drawing data D10 in which the options in the processed product characteristic data D11 for the new drawing data D10 are the same as or similar to the options in the processed product characteristic data D11 associated with the existing drawing data D10 is extracted from the drawing database 210. For attributes selected by numerical values, existing drawing data D10 in which the numerical values ​​in the processed product characteristic data D11 for the new drawing data D10 are the same as or fall within a predetermined numerical range is extracted from the drawing database 210. In this case, multiple attributes may be combined to extract existing drawing data D10 that is the same as or similar to the processed product characteristic data D11 for the new drawing data D10 from the drawing database 210.

[0098] Then, the request destination selection unit 203C acquires the processed product request data D12 indicating the destination to be requested to process or estimate the processed product in the new drawing data D10 accepted by the drawing acceptance unit 202 based on the processed product request data D12 associated with the extracted similar drawing data D10, and selects candidate destinations based on the processed product request data D12.

[0099] For example, when the request destination selection unit 203C extracts one similar drawing data D10, it selects the request destination indicated by the processed product request data D12 associated with the similar drawing data D10 as a candidate request destination. Also, when the request destination selection unit 203C extracts a plurality of similar drawing data D10 as shown in FIG. 14 , it selects the request destination as a candidate request destination by statistically processing the plurality of processed product request data D12 associated with the plurality of similar drawing data D10. For example, the plurality of processed product request data D12 may be sorted in descending order of frequency as a request destination, and the most frequently requested destination may be selected as a candidate request destination, or the request destinations from the top to a predetermined rank in descending order of frequency may be selected as candidate request destinations.

[0100] (Operation of drawing processing device 2C) Fig. 15 is a flowchart showing an example of the operation (drawing processing method) of drawing processing device 2C according to the third embodiment. Note that, since the processes having the same step numbers as those in the flowchart shown in Fig. 8 are the same as those in the first embodiment, the following description will mainly focus on the parts that are different from the first embodiment.

[0101] First, in step S100 (drawing reception step), the drawing reception unit 202 receives new drawing data D10.

[0102] Next, in step S113 (request destination selection process), the request destination selection unit 203C performs character recognition processing on the new drawing data D10 received in step S100 as a feature acquisition process, thereby acquiring processed product feature data D11 for the new drawing data D10.

[0103] Next, in step S114 (request destination selection process), the request destination selection unit 203C, as a candidate selection process, extracts, from the existing drawing data D10 registered in the drawing database 210, existing drawing data D10 that is identical to or similar to the characteristics indicated by the processed product characteristic data D11 acquired in step S113, as similar drawing data D10. Next, the request destination selection unit 203C acquires processing request data D12 for the processed product in the new drawing data D10 accepted in step S100, based on processing request data D12 associated with the extracted similar drawing data D10. Then, the request destination selection unit 203C selects the candidate request destinations indicated by the processing request data D12 as candidate request destinations to process or estimate the processed product in the new drawing data D10 accepted in step S100.

[0104] Next, in step S122 (database management step), the database management unit 204 associates the processed product characteristic data D11 and processed product request data D12 obtained as the processing results of steps S113 and S114 with the new drawing data D10 accepted in step S100 and registers them in the drawing database 210. Then, in step S132, the database management unit 204 transmits display information for displaying the processing results of steps S113 and S114 to the user terminal device 3.

[0105] As described above, according to the drawing processing device 2C and drawing processing method of this embodiment, by using the drawing database 210 in which past performance data when selecting a contractor for existing drawing data D10 is registered, it is possible to appropriately select a contractor to process or estimate the processed product described in the drawing without relying on the user's knowledge or experience.

[0106] (Other Embodiments) The present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit and scope of the present invention, all of which are included in the technical concept of the present invention.

[0107] In the above embodiment, the drawing processing devices 2A to 2C are described as being configured as a single device, but they may be configured as multiple devices. For example, by distributing the units 200 to 204 of the drawing processing devices 2A to 2C across multiple devices, the drawing management system 1 may be configured as a machine learning device that includes learning model generation units 201A and 201B and performs the machine learning process, and a requestee selection device that includes requestee selection units 203A to 203C and performs the requestee selection process. In this case, each unit (each process) of each of the above devices may be realized by a program executable on the computer 900.

[0108] In the above embodiment, the drawing processing devices 2A to 2C receive new drawing data D10 from the user terminal device 3 and select a supplier to process the drawing data D10. Alternatively, the user terminal device 3 may function as the drawing processing devices 2A to 2C.

[0109] The above embodiments can be combined as appropriate. For example, the request destination selection unit (request destination selection step) included in the drawing processing devices 2A to 2C may perform the feature acquisition process (step S110) in the first embodiment and the candidate selection process (step S114) in the third embodiment, or may perform the feature acquisition process (step S113) in the third embodiment and the candidate selection process (step S111) in the first embodiment.

[0110] In the above embodiment, the request destination selection units 203A to 203C of the drawing processing devices 2A to 2C acquire the processed product request data D12 for the new drawing data D10. However, useful information for the processing request or quotation request may also be acquired. As useful information, the request destination selection units 203A to 203C may acquire detailed processing information, such as the type of processing equipment used in the processing process, the processing time required for the processing process, and whether manual labor is required. The detailed processing information may be included in the output data of either the first learning model 220A or the second learning model 220B, or may be included in the output data of the learning model 220. Furthermore, the detailed processing information may be registered in the drawing database 210 and acquired by referencing the drawing database 210.

[0111] (Inference Device, Inference Method, and Inference Program) The present invention can be provided not only in the form of the drawing processing devices 2A to 2C (drawing processing method or drawing processing program) according to the above-described embodiments, but also in the form of an inference device (inference method or inference program) used to infer candidate destinations for a processed product in new drawing data D10. In this case, the inference device (inference method or inference program) can include a memory and a processor, and the processor executes a series of processes. The series of processes includes a data acquisition process (data acquisition step) for acquiring new drawing data D10 related to the processed product, and an inference process (inference step) for inferring candidate destinations for processing or estimating the processed product in the new drawing data D10 once the new drawing data D10 has been acquired by the data acquisition process.

[0112] 1...Drawing management system, 2A to 2C...Drawing processing device, 3...User terminal device, 10...Drawing area, 11...Shape area, 12...Title block, 13...Bill of materials, 20...Control unit, 21...Data storage unit, 22...Learned model storage unit, 23...Communication unit, 24...Input unit, 25...Output unit, 200...Transmission and reception control unit, 201A, 201B...Learning model generation unit, 202...Drawing reception unit, 203A to 203C...Request destination selection unit, 204...Database management unit, 210...Drawing database, 211A to 211C...Drawing processing program, 220...Learning model, 220A...First learning model, 220B...Second learning model, 2010A, 2010B...Learning data acquisition unit, 2011A, 2011B...Machine learning unit

Claims

1. A drawing processing device comprising: a drawing receiving unit that receives new drawing data relating to a processed product; and a requestee selection unit that selects candidate requestees to process or estimate the processed product in the new drawing data based on the new drawing data received by the drawing receiving unit.

2. The drawing processing device described in claim 1, wherein the request destination selection unit inputs the new drawing data received by the drawing receiving unit into a learning model to obtain processed product request data indicating the destination to which the processed product in the new drawing data should be submitted, and selects candidate destinations based on the processed product request data, and the learning model is trained by machine learning to determine the correlation between existing drawing data regarding the processed product and processed product request data indicating the destination to which the processing or estimate of the processed product in the existing drawing data should be submitted.

3. The drawing processing device of claim 1, wherein the request destination selection unit performs a feature acquisition process to acquire processed product feature data indicating the features of the processed product in the new drawing data from the new drawing data accepted by the drawing acceptance unit, and a candidate selection process to acquire processed product request data indicating the request destination for the processed product in the new drawing data based on the processed product feature data acquired in the feature acquisition process, and select candidate request destinations based on the processed product request data.

4. The drawing processing device of claim 3, wherein the request destination selection unit, as the feature acquisition process, inputs the new drawing data accepted by the drawing acceptance unit into a first learning model to acquire the processed product feature data for the processed product in the new drawing data, and the first learning model is trained by machine learning to determine the correlation between existing drawing data regarding the processed product and the processed product feature data for the processed product in the existing drawing data.

5. The drawing processing device according to claim 3 or claim 4, wherein the request destination selection unit, as the candidate selection process, inputs the processed product feature data acquired in the feature acquisition process into a second learning model to acquire the processed product request data for the processed product in the new drawing data accepted by the drawing acceptance unit, and selects the candidate request destination based on the processed product request data, and the second learning model is trained by machine learning to determine the correlation between the processed product feature data and the processed product request data for the processed product having the feature indicated by the processed product feature data.

6. The drawing processing device described in claim 3, wherein the request destination selection unit, as the feature acquisition process, performs character recognition processing on the new drawing data accepted by the drawing acceptance unit to acquire the processed product feature data for the processed product in the new drawing data.

7. The drawing processing device according to claim 3 or claim 6, wherein the request destination selection unit, as the candidate selection process, refers to a drawing database in which existing drawing data relating to processed products, the processed product feature data for the processed product in the existing drawing data, and the processed product request data for the processed product in the existing drawing data are registered in association with each other, extracts, from the existing drawing data registered in the drawing database, the existing drawing data that is identical to or similar to the features indicated by the processed product feature data based on the processed product feature data acquired in the feature acquisition process, as similar drawing data, and selects the candidate request destination for the processed product in the new drawing data accepted by the drawing acceptance unit based on the processed product request data associated with the similar drawing data.

8. An inference device comprising a memory and a processor, wherein the processor executes a data acquisition process for acquiring new drawing data relating to a processed product, and an inference process for inferring candidate contractors to which the new drawing data is to be requested to process or estimate the processed product in the new drawing data, upon acquiring the new drawing data through the data acquisition process.

9. A machine learning device comprising: a learning data acquisition unit that acquires multiple sets of learning data consisting of input data and output data; a machine learning unit that uses the multiple sets of learning data acquired by the learning data acquisition unit to train a learning model by machine learning to learn the correlation between the input data and the output data; and a trained model storage unit that stores the learning model that has learned the correlation by the machine learning unit, wherein the input data includes existing drawing data related to a processed product, and the output data includes processed product request data that indicates a destination to request processing or an estimate for the processed product in the existing drawing data.

10. A drawing processing method executed by a computer, comprising: a drawing receiving step of receiving new drawing data relating to a processed product; and a client selection step of selecting candidate clients to be requested to process or estimate the processed product in the new drawing data, based on the new drawing data received by the drawing receiving step.

11. An inference method executed by an inference device having a memory and a processor, wherein the processor executes a data acquisition process to acquire new drawing data related to a processed product, and an inference process to infer, upon acquiring the new drawing data through the data acquisition process, candidate contractors to be requested to process or estimate the processed product in the new drawing data.

12. A machine learning method executed by a computer, comprising: a learning data acquisition step of acquiring multiple sets of learning data consisting of input data and output data; a machine learning step of using the multiple sets of learning data acquired by the learning data acquisition step to have a learning model learn the correlation between the input data and the output data by machine learning; and a learned model storage step of storing the learning model that has learned the correlation by the machine learning step in a learned model storage unit, wherein the input data includes existing drawing data related to a processed product, and the output data includes processed product request data indicating a destination to request processing or an estimate for the processed product in the existing drawing data.

Images