An intelligent management method and device applied to the manufacturing industry

Abstract

The present application is applicable to the field of data processing technology, and provides an intelligent management method and device applied to the manufacturing industry, including: responding to a task request corresponding to a target user, determining the request type of the task request; calling a data interface associated with the request type to obtain task data associated with the task request; importing the task data into a neural network built based on a manufacturing database, and outputting the response data of the task request; and converting the response data into a natural language conversion model associated with the target user to obtain the task feedback result of the task request. The above method can automatically obtain the task data through the corresponding data interface when receiving the task request of the target user, without the need for the user to manually import the associated task data, and select the processing algorithm through the neural network, and perform associated processing on the task data, thereby improving the processing efficiency of the task data.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and in particular relates to an intelligent management method and equipment applied in the manufacturing field. Background Art

[0002] With the continuous development of the manufacturing sector, its level of automation is also constantly improving. During the production process, the manufacturing sector generates a large amount of task data, such as product testing data and production line operation data. How to efficiently manage this task data has become a key focus for users.

[0003] Existing manufacturing technologies, due to the large amount of task data involved, require users to adopt specific methods to process different types of task data. For example, when verifying product test data, it is necessary to obtain the corresponding inspection specifications to determine whether the product meets the requirements; while for production line operation data, it is necessary to compare it with the corresponding error range. Therefore, in the manufacturing field, task data is large in volume and has many different types, requiring users to manually match data algorithms, which greatly increases the difficulty of task data management and reduces the efficiency of production line management. Summary of the Invention

[0004] This application provides an intelligent management method and device for the manufacturing industry, which can solve the problem that in the manufacturing industry, the task data is large in volume and has many data types, requiring users to manually perform data algorithm matching, which greatly increases the difficulty of task data management and reduces the efficiency of production line management.

[0005] In a first aspect, an embodiment of the present application provides an intelligent management method applied in the manufacturing field, the method comprising:

[0006] In response to a task request corresponding to a target user, determining a request type of the task request; the task request includes a report push request, an early warning analysis request, and / or an intelligent question and answer request;

[0007] Calling a data interface associated with the request type to obtain task data associated with the task request; the data interface is determined based on a correspondence between the request type and the interface;

[0008] Importing the task data into a neural network constructed based on a manufacturing database, and outputting response data of the task request;

[0009] The response data is converted into a natural language model associated with the target user to obtain a task feedback result of the task request.

[0010] In a possible implementation of the first aspect, importing the task data into a neural network constructed based on a manufacturing database and outputting response data for the task request includes:

[0011] Acquire user information of the target user; the user information includes the first node of the target user on the manufacturing production line and the language type of the target user;

[0012] The first task data is converted into text using the language type to obtain the calibration task data corresponding to the first task data; the first task data is the task data uploaded by the target user.

[0013] Based on the manufacturing production line, the upstream operating nodes of the first node are determined, and the manufacturing database is filtered according to the first node and all the upstream operating nodes to obtain a sub-database associated with the target user; the data in the sub-database is the data associated with the first node and all the upstream operating nodes.

[0014] The calibration task data and the second task data are imported into the neural network to generate the SQL statement corresponding to the task request. The SQL statement is then used to search for data in the sub-database to obtain the response data.

[0015] In one possible implementation of the first aspect, the step of determining the upstream operating nodes of the first node based on the manufacturing production line, and filtering the manufacturing database based on the first node and all the upstream operating nodes to obtain a sub-database associated with the target user, includes:

[0016] According to the manufacturing field corresponding to the target user, the manufacturing production line in the manufacturing field is obtained; the manufacturing production line includes multiple operation nodes, and each operation node corresponds to an operation sequence;

[0017] Based on the operation path of the first node on the manufacturing production line, selecting other operation nodes whose operation sequence is earlier than that of the first node as the upstream operation nodes;

[0018] Based on the first working period corresponding to the target user and the operation sequence, a second working period associated with each upstream operation node is determined; the time interval between the second working period and the first working period is determined based on the operation sequence and the process duration corresponding to each operation.

[0019] Based on the manufacturing database, the production data of the upstream operation node during the second working period and the interaction data of the associated users of the upstream operation node during the second working period are selected.

[0020] The sub-database is obtained according to the production data and the interaction data of all the upstream operation nodes and the production data of the first node in the first working period.

[0021] In a possible implementation of the first aspect, the text conversion of the first task data according to the language type to obtain calibration task data corresponding to the first task data, wherein the first task data is task data uploaded by the target user, includes:

[0022] According to the manufacturing field corresponding to the manufacturing database, a keyword library corresponding to the manufacturing field is obtained; the keyword library contains at least one professional keyword related to the manufacturing field;

[0023] Determining abnormal characters in the historical conversation records of the target user according to the historical conversation records of the target user;

[0024] Obtaining abnormal feature data of the target user based on the abnormal character and the associated characters adjacent to the abnormal character, and performing text calibration on the first task data using the abnormal feature data to obtain calibrated text data;

[0025] Convert each of the professional keywords in the keyword library into expression keywords corresponding to the language type;

[0026] Each of the expression keywords is matched with the calibration text data in sequence, and the calibration task data is generated based on the matching results.

[0027] In a possible implementation of the first aspect, importing the calibration task data and the second task data into the neural network, generating an SQL statement corresponding to the task request, and performing a data search in the sub-database using the SQL statement to obtain the response data includes:

[0028] Importing the calibration task data and the second task data into the convolutional layer of the neural network to generate a text feature vector corresponding to the task data;

[0029] The matching degree between the candidate keywords in the manufacturing field and the text feature vector is calculated respectively; the matching degree is:

[0030]

[0031] Among them, MatchPiont iis the matching degree between the i-th candidate keyword and the text feature vector; Text(j) is the j-th text keyword in the text feature vector; M is the total number of text keywords in the text feature vector; Keywork i is the i-th candidate keyword; BaseDis is the preset benchmark word vector distance; Min[x] is the minimum value selection function;

[0032] Based on the matching degree, the target keyword associated with the text feature vector is selected from all the candidate keywords;

[0033] Generate the SQL statement based on all target keywords.

[0034] In one possible implementation of the first aspect, the step of calling the data interface associated with the request type to obtain the task data associated with the task request includes:

[0035] If the request type is a composite request type, determining the data types included in the composite request type; the data types include graphic data types and text data types;

[0036] Calling a graphics data interface to obtain graphics task data from multiple target objects; the target objects include distributed terminals deployed on the production line to obtain image data and office applications corresponding to the production line;

[0037] Calling a text data interface to import the graphic task data into a preset graphic understanding model to generate first text data corresponding to the graphic task data;

[0038] The second text data corresponding to the task request is obtained through the text data interface;

[0039] The task data is obtained based on the graphic task data, the first text data, and the second text data.

[0040] In a possible implementation of the first aspect, calling a graphics data interface to obtain graphics task data from multiple target objects includes:

[0041] Determining, by means of a dust sensor on the distributed terminal, a dust index corresponding to when the distributed terminal acquires raw image data;

[0042] Processing the original image data based on an image enhancement model corresponding to the dust index to obtain calibrated image data; the image enhancement model is trained based on the production equipment and products included in the production line;

[0043] Performing image segmentation on the calibration image data according to the shooting angle of the distributed terminal to obtain a target area image corresponding to the area where the production product is located;

[0044] The graphics task data is generated based on the target area images collected by all distributed terminals.

[0045] Secondly, embodiments of this application provide an intelligent management device applied in the manufacturing industry, the device comprising:

[0046] A task request response unit is configured to respond to a task request initiated by a target user and determine the request type of the task request; the task request includes a report push request, an early warning analysis request, and / or an intelligent question and answer request;

[0047] A task data acquisition unit, configured to call a data interface associated with the request type to acquire task data associated with the task request; the data interface is determined based on a correspondence between the request type and the interface;

[0048] a response data generating unit, configured to import the task data into a neural network constructed based on a manufacturing database, and output response data of the task request;

[0049] The task result feedback unit is used to convert the response data into a natural language conversion model associated with the target user to obtain a task feedback result of the task request.

[0050] In a third aspect, an embodiment of the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method described in any one of the first aspects above is implemented.

[0051] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in any one of the first aspects above is implemented.

[0052] In a fifth aspect, an embodiment of the present application provides a computer program product, which, when executed on an electronic device, enables the electronic device to execute any one of the methods described in the first aspect above.

[0053] Compared with the prior art, the embodiments of the present application have the following beneficial effects: when a task request initiated by a target user is received, the task data associated with the task request can be obtained through different data interfaces, and the task data can be automatically processed through a neural network constructed based on a manufacturing database, so that the task data can be analyzed and processed in the neural network using an algorithm corresponding to the task request to obtain corresponding response data, and then the response data is converted into natural language to obtain a task feedback result that is easy for the user to read and understand, thereby achieving the purpose of automating the management of task data in the manufacturing industry. Compared with the existing manufacturing technology, the embodiments of the present application can automatically obtain task data through a corresponding data interface when a task request is received from a target user, without the need for the user to manually import the associated task data, and select a processing algorithm through a neural network, and perform associated processing on the task data, thereby improving the processing efficiency of the task data, and processing the response data through a natural language conversion model, thereby improving the readability of the task feedback result, improving the management efficiency of task data in the manufacturing industry, and reducing the difficulty of management. Attached Figure Description

[0054] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the embodiments or descriptions of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0055] Figure 1 This is a schematic diagram of the structure of a manufacturing management system provided by an embodiment of the present application;

[0056] Figure 2 This is a schematic diagram of an implementation of an intelligent management method applied to the manufacturing industry provided by an embodiment of the present application;

[0057] Figure 3 This is a data flow diagram of a manufacturing management system provided by an embodiment of the present application;

[0058] Figure 4 This is a flowchart of a specific implementation of an intelligent management method applied to the manufacturing industry in S203 provided in the second embodiment of the present application;

[0059] Figure 5 This is a flowchart of a specific implementation of an intelligent management method applied to the manufacturing industry in step S2033 provided in the third embodiment of the present application;

[0060] Figure 6 is a schematic diagram of a manufacturing production line provided in one embodiment of the present application;

[0061] Figure 7 This is a flowchart illustrating the specific implementation of an intelligent management method applied in the manufacturing industry according to the fourth embodiment of this application in S2032.

[0062] Figure 8 This is a flowchart illustrating the specific implementation of an intelligent management method applied in the manufacturing industry according to the fifth embodiment of this application in S2034.

[0063] Figure 9 This is a flowchart illustrating the specific implementation of an intelligent management method applied in the manufacturing industry provided in the sixth embodiment of this application in S202;

[0064] Figure 10 This is a schematic diagram of the structure of an intelligent management device applied in the manufacturing industry, as provided in an embodiment of this application;

[0065] Figure 11 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. DETAILED DESCRIPTION

[0066] In the following description, specific details such as specific system structures and techniques are provided for purposes of illustration rather than limitation to facilitate a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to avoid obscuring the description of the present application with unnecessary detail.

[0067] It should be understood that, when used in this application specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or a collection thereof.

[0068] Furthermore, in the description of this application and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0069] The intelligent management method for manufacturing provided in the embodiment of the present application can be applied to manufacturing management systems. For example, Figure 1 A schematic diagram of the structure of a manufacturing management system provided in one embodiment of this application is shown. See also... Figure 1As shown, the manufacturing management system includes distributed terminals 11 for collecting relevant data during the manufacturing process, production lines 12 for producing products, and management equipment 13 for processing relevant data during the production process. The distributed terminals 11 can obtain relevant data from production lines 12 during the production process, such as product data, production line operation data, and relevant staff operation data.

[0070] The distributed terminals 11 may transmit the acquired relevant data to the management device 13 for data storage and processing. The management device 13 may be a server configured with a manufacturing database, and the data collected by the distributed terminals 11 may be stored in the manufacturing database. Of course, the manufacturing management system may also include an independent database server, and the relevant data collected by the distributed terminals 11 may be stored in the database server. The management device 13 may obtain task data from the database server when responding to task requests.

[0071] The production line 12 may include multiple production equipment, and workers can control the production equipment to perform automated production. During the operation of the production line 12, related products will be produced one after another. If the distributed terminal 11 is a camera terminal, it can capture the appearance of the products being produced and collect image data and video data during the production process. If the distributed terminal is a detection terminal, it can capture product parameters, such as dimensional parameters and product weight. The type of data collected can be selected based on actual conditions.

[0072] In this embodiment, the data acquired by the above-mentioned distributed terminal includes multiple data types. For example, if the above-mentioned distributed terminal is a camera device, the acquired data is image data and / or video data. By performing data analysis on the image data and / or video data, the corresponding image semantic text and audio text, that is, text data, can be obtained. The collected related data can be classified and stored according to the data type.

[0073] See also Figure 2 , Figure 2 The following is a schematic diagram illustrating an implementation of an intelligent management method for the manufacturing industry provided by an embodiment of the present application. The management method is applied to the manufacturing management system provided by an embodiment of the present application. The execution subject of the intelligent management method can be an electronic device, such as a computer, a laptop, or a server. In particular, the electronic device is the management device 13 described above. Specifically, the method includes the following steps:

[0074] In S201, in response to a task request corresponding to a target user, the request type of the task request is determined; the task request includes a report push request, an early warning analysis request, and / or an intelligent question answering request.

[0075] In this embodiment, after receiving a task request initiated by the target user, the electronic device can determine the task type corresponding to the request. This task request can be generated by the target user on their terminal and sent to the electronic device, or it can be automatically generated based on triggering conditions pre-set by the target user.

[0076] In this embodiment, after receiving the task request, the electronic device can determine the request type of the task request. The task request may carry a type identifier; by identifying this type identifier, the request type corresponding to the task request can be determined. The request types include: report push type, early warning analysis type, and intelligent question answering type.

[0077] In this embodiment, the aforementioned task request can be a report push request. In the manufacturing sector, the number of products produced is large, and the production line involves numerous operational nodes. To better manage the manufacturing process, the target user can send a report push request to the electronic device to generate relevant management reports during product production. This report push request can be initiated by the target user through a relevant application installed on their terminal, or it can be automatically generated based on corresponding automatic triggering conditions set by the target user.

[0078] In this embodiment, the aforementioned task request can be an early warning analysis request. In the manufacturing sector, automated production is often employed. To improve the stability of the production process and promptly detect anomalies during operation, the target user can initiate a corresponding early warning analysis request through a user terminal to monitor the operation of the production process, determine whether there are any abnormal events, and generate a corresponding anomaly early warning report through electronic devices when an anomaly is detected.

[0079] In this embodiment, the aforementioned task request can be an intelligent question-and-answer request. Since the production process involves multiple stages, and the procedures and specifications within each stage differ, the target user often cannot fully understand the operational details and precautions of each stage. Therefore, the target user can send a corresponding intelligent question-and-answer request through their user terminal. This request contains a description of at least one question that the target user needs to understand, so that the corresponding intelligent question-and-answer result can be generated through an electronic device.

[0080] For example, Figure 3 A schematic diagram of the data flow of a manufacturing management system according to an embodiment of this application is shown. See also Figure 3As shown, the manufacturing management system includes source data 31, a big data model 32, and published data 33. The source data 31 can be acquired through multiple channels, such as electronic documents, external web pages, user chat logs, and monitoring data acquired through distributed terminals on the production line. All of this source data 31 is cleaned and summarized to generate a corresponding manufacturing database 311. Electronic devices can use the manufacturing database 311 to train neural networks and optimize them during use to generate the corresponding big data model 32. Users can initiate task requests, for example, through a webpage, official account, API call, or related application. The big data model 32 can then construct a corresponding intelligent robot 321 to receive the task requests sent by the user and provide feedback to the big data model 32 for task processing. After the big data model 32 completes the task processing and obtains the task feedback results, the intelligent robot 321 can display the corresponding processing results to the user.

[0081] In S202, a data interface associated with the request type is called to obtain task data associated with the task request; the data interface is determined according to a corresponding relationship between the request type and the interface.

[0082] In this embodiment, the electronic device is configured with multiple data interfaces, and different data interfaces can be used to implement different calling functions. For example, the above-mentioned data interface can be used to retrieve relevant data from a database; for another example, the electronic device can be an interface for calling an office application, so as to open an office document through an office application deployed on the cloud, and then obtain relevant data from the above-mentioned office document, such as text data, graphic data, table data, and image data; for another example, the electronic device can be a video stream interface associated with a distributed camera device, and obtain video stream data fed back by the camera device through the video stream interface. It can be seen that the electronic device can obtain different types of task data by calling the interface, and the electronic device can store the correspondence between the request type and the interface, and by querying the above-mentioned correspondence, determine the data interface associated with the request type corresponding to the task request, and call the associated data interface to obtain the task data corresponding to the task request.

[0083] In S203, the task data is imported into a neural network built on a manufacturing database, and the response data for the task request is output.

[0084] In this embodiment, the electronic device can construct a corresponding manufacturing database based on acquired historical task data and results. The electronic device can construct corresponding neural networks based on different request types, and use data from the manufacturing database as training samples to train each request type's neural network, thereby obtaining the trained neural network. Importing the corresponding task data into the neural network yields response data for the task request. Since the task data can be used to determine the task content of the request, processing the task data through the neural network corresponding to the task request allows for the acquisition of response data corresponding to the task content.

[0085] It should be noted that the above response data is a data set of task results. For example, if the task request is to generate a corresponding early warning report, then the above response data consists of the relevant early warning indicators and early warning results in the early warning report. However, the corresponding monitoring item for each of the above early warning indicators and results may not be marked in the above response data, which means that its readability is poor. Therefore, it is necessary to optimize the readability of the above response data through step S204 so that the task feedback results can be more easily understood by the user.

[0086] In S204, the response data is converted into a natural language model associated with the target user to obtain a task feedback result of the task request.

[0087] In this embodiment, the electronic device is configured with a natural language conversion model. After generating response data corresponding to the task request through a neural network, the above response data can be imported into the above natural language conversion model to generate the corresponding task feedback result.

[0088] In some implementations, different request types correspond to different result templates. Based on this, the electronic device can obtain a natural language conversion model associated with the request type according to the request type. The natural language conversion model is generated through learning and training based on multiple training task results generated based on the above-mentioned result templates, thereby improving the accuracy of the generated task feedback results.

[0089] In this embodiment, the electronic device acquires the target user's user information and adjusts the preset language conversion model based on the user information to obtain a natural language conversion model associated with the target user. The aforementioned user information may include the target user's historical chat records, allowing the determination of the target user's language habits, and thus ensuring that the generated task feedback results match the target user's language habits.

[0090] In some implementations, the above user information may also include the target user's native language type and educational information, because the native language type will affect the target user's word usage habits, and the educational information can reflect the target user's language comprehension ability. In the manufacturing industry, there are large differences in the educational background of users, and there are certain differences in the ability to understand texts with different educational backgrounds. In addition, different users use different native languages, which will also affect their conventional description methods and reading comprehension methods. Therefore, when converting the response data through a natural language model, the original language conversion model can be calibrated according to the above-mentioned educational information and native language type, so as to generate a natural language conversion model corresponding to the target user.

[0091] From the above, it can be seen that the embodiment of the present application provides an intelligent management method for the manufacturing industry. When a task request initiated by a target user is received, the task data associated with the task request can be obtained through different data interfaces, and the task data can be automatically processed through a neural network built based on a manufacturing database. Thus, the task data can be analyzed and processed using an algorithm corresponding to the task request in the neural network to obtain corresponding response data, and then the response data is converted into natural language to obtain a task feedback result that is easy for the user to read and understand, thereby achieving the purpose of automated management of task data in the manufacturing industry. Compared with existing manufacturing technologies, the embodiment of the present application can automatically obtain task data through a corresponding data interface when a task request is received from a target user, without the need for the user to manually import the associated task data, and select a processing algorithm through a neural network, and perform associated processing on the task data, thereby improving the processing efficiency of the task data, and processing the response data through a natural language conversion model, thereby improving the readability of the task feedback result, improving the management efficiency of task data in the manufacturing industry, and reducing the difficulty of management.

[0092] Figure 4 This diagram illustrates a detailed implementation flowchart in step S203 of an intelligent management method applied in the manufacturing industry, as provided in the second embodiment of this application. See also... Figure 4 , relative to Figure 2 In the embodiment described herein, S203 of an intelligent management method applied in the manufacturing industry includes: S2031 to S2034, which are detailed below:

[0093] Furthermore, the step of importing the task data into a neural network built on a manufacturing database and outputting the response data for the task request includes:

[0094] In S2031, the user information of the target user is obtained; the user information includes the first node of the target user on the manufacturing production line and the language type of the target user.

[0095] In this embodiment, the electronic device can retrieve user information corresponding to the user identifier carried in the task request from a user database, i.e., the user information of the target user. This user information includes the operation node (i.e., the first node) of the operation performed by the target user on the manufacturing production line. For example, if the target user is responsible for product design, the corresponding first node is the product design node; for another example, if the target user is responsible for product packaging, the corresponding first node is the product packaging node.

[0096] In this embodiment, the user information also includes a language type. The electronic device can determine the language type of the target user based on information such as the target user's nationality and / or place of origin. The language type can include language and dialect types, and the specific language type division granularity can be set according to actual conditions. For example, if the product produced on the manufacturing production line is a product that requires collaborative manufacturing in multiple countries, the language division granularity can be language types, such as Chinese, English, and Japanese; if the product is specifically developed domestically, the language division granularity can be dialects, such as Mandarin, Shanghainese, and Cantonese.

[0097] In S2032 , text conversion is performed on the first task data according to the language type to obtain calibration task data corresponding to the first task data; the first task data is the task data uploaded by the target user.

[0098] In this embodiment, the electronic device can perform data cleaning on the task data in the manufacturing database. Since the first task data is the task data uploaded by the target user, for example, the task data can be the voice data of the target user, or text data written in the language of the target user, and its content will be affected by the language habits of the target user, in order to achieve the uniformity of the task data, the electronic device can perform text conversion on the first task data uploaded by the target user in the task data according to the language type of the target user, thereby obtaining corresponding calibration task data.

[0099] In some implementations, the first task data includes an audio signal from a target user. The electronic device may then, based on the target user's language type, obtain a speech recognition algorithm corresponding to the target user's language type and input the audio signal into the speech recognition algorithm to convert the audio signal into text data, thereby obtaining the calibration task data. Optionally, the electronic device may also perform keyword calibration on the converted text data. For example, based on specialized vocabulary in the manufacturing industry, the electronic device may calibrate similar vocabulary in the text data to improve the accuracy of the converted calibration task data.

[0100] In some implementations, the first task data includes chat records of the target user. The electronic device can determine the synonymous keywords corresponding to multiple professional keywords in the manufacturing field under the language type based on the language type of the target user, and convert the synonymous keywords contained in the chat records into corresponding professional keywords, thereby obtaining calibration task data for word replacement in the chat content to improve the uniformity of professional vocabulary in multiple languages.

[0101] In S2033, the upstream operation node of the first node is determined based on the manufacturing production line, and the manufacturing database is screened according to the first node and all the upstream operation nodes to obtain a sub-database associated with the target user; the data in the sub-database is data associated with the first node and all the upstream operation nodes.

[0102] In this embodiment, the manufacturing production line has a certain operation sequence, and different operation nodes in the same operation path need to be completed in sequence. For example, the production operation node needs to be executed after the product design node is completed. That is, when the user executes the corresponding task request, the preceding operation node will generate the task data corresponding to this production cycle, while the task data of the subsequent operation node is not generated, that is, the subsequent operation node only has the task data of the previous production cycle. Therefore, in order to improve the accuracy of the task request response, the electronic device can filter the data in the manufacturing database to obtain the corresponding sub-database to reduce the time required for data retrieval.

[0103] In this embodiment, the electronic device can obtain the operation process corresponding to the manufacturing production line, locate the above-mentioned first node within the above-mentioned operation process, and use other operation nodes whose operation timing is earlier than the above-mentioned first node as upstream operation nodes, and then construct a corresponding database, i.e., a sub-database, for the relevant data in the first node and the relevant data in each upstream operation node.

[0104] In S2034, the calibration task data and the second task data are imported into the neural network, an SQL statement corresponding to the task request is generated, and data search is performed in the sub-database through the SQL statement to obtain the response data; the second task data is other task data except the first task data.

[0105] In this embodiment, the electronic device can import the calibration task data and the second task data into the corresponding neural network, and convert them into Structured Query Language (SQL) statements for data retrieval through the neural network. For example, if the task data is the operation behavior data of the target user, it can be converted into an SQL statement associated with the operation behavior data through the neural network, so that other data whose data content has changed based on the operation behavior data can be searched in the sub-database, thereby obtaining the behavior logic chain corresponding to the target user and determining whether the behavior logic chain meets the corresponding behavior specification, or whether the behavior result obtained based on the behavior logic chain meets the product production specification, etc.

[0106] In this embodiment, after the electronic device converts the task calibration data and the second task data into corresponding SQL statements, it can search the sub-database to retrieve the response data associated with the task data. For example, if the task request is an intelligent question-and-answer request, the task data is the question content, the SQL statement is the query for searching for the question content, and the answer corresponding to the question content is retrieved from the manufacturing database. The answer to the question is the response data.

[0107] In this embodiment of the application, the task data is cleaned by the target user's language type, and the data in the manufacturing database is filtered according to the operation node where the target user is located. This can improve the accuracy of the task data and reduce the workload when performing data retrieval based on SQL statements.

[0108] Figure 5 This diagram illustrates a detailed implementation flowchart in S2033 of an intelligent management method for the manufacturing industry provided in the third embodiment of this application. See also... Figure 5 In contrast Figure 4 The embodiment provided in this embodiment, an intelligent management method applied to the manufacturing field, further includes S501 to S505 in S2033, which are detailed below:

[0109] Further, the step of determining the upstream operating nodes of the first node based on the manufacturing production line, and filtering the manufacturing database based on the first node and all the upstream operating nodes to obtain a sub-database associated with the target user, includes:

[0110] In S501, the manufacturing production line of the manufacturing field corresponding to the target user is obtained; the manufacturing production line includes multiple operation nodes, and each operation node corresponds to an operation sequence.

[0111] In this embodiment, the electronic device can determine the manufacturing production line corresponding to the manufacturing industry in which the target user is located. Different industries have different manufacturing production lines, and the electronic device can obtain a pre-set manufacturing production line based on the target user's industry (i.e., the aforementioned manufacturing industry).

[0112] In S502 , based on the operation path of the first node on the manufacturing production line, other operation nodes whose operation sequence is earlier than that of the first node are selected as the upstream operation nodes.

[0113] In this embodiment, a manufacturing production line may include multiple operation paths, and different operation paths can be executed in parallel. Based on this, if a first node is on a certain operation path, the operation nodes of other operation paths different from the operation path are unrelated to the first node. Therefore, by determining the operation path of the first node on the manufacturing production line, other operation nodes unrelated to the first node can be filtered out. Based on the operation timing of each operation node on the operation path, the electronic device can select other operation nodes with an operation timing earlier than the first node as the upstream operation nodes of the first node.

[0114] For example, Figure 6 Schematic diagram of a manufacturing production line provided by an embodiment of the present application is shown. Figure 6 As shown, the manufacturing production line includes multiple operation nodes, and different operation nodes can constitute multiple different operation paths, namely path 1 and path 2. Among them, the first node is specifically node 5, and the node 5 is on path 1, that is, path 1 includes (node ​​1→node 2→node 3→node 5→node 6), and path 2 includes (node ​​1→node 2→node 4→node 7→node 8), wherein node 1 and node 2 are common operation nodes. The operation timing of the above-mentioned node 4 is earlier than that of the above-mentioned node 5, but because the above-mentioned two nodes are on different operation paths, when the electronic device determines the upstream operation node of node 5, it does not need to consider other operation nodes on different operation paths from the first node, that is, the upstream operation nodes corresponding to node 5 are node 1, node 2 and node 3.

[0115] In S503, a second working period associated with each upstream operation node is determined based on the first working period corresponding to the target user and the operation sequence; the time interval between the second working period and the first working period is determined based on the operation sequence and the process duration corresponding to each operation.

[0116] In this embodiment, the electronic device can determine the first working period corresponding to the target user when performing the operation task corresponding to the first node, and determine the second working period corresponding to each upstream operation node based on the first working period, so that the task performed by the upstream operation node in the second working period can be regarded as the predecessor task of the production task performed by the target user, that is, through the relationship between different operation nodes in the working period, more accurate data screening can be achieved.

[0117] Exemplarily, the operation interval between each operation node is Time, and the timing difference between the first node and the upstream operation node is T, then the difference between the second operation period and the first operation period is: T*Time.

[0118] In S504, based on the manufacturing database, the production data of the upstream operation node in the second working period and the interaction data of the users associated with the upstream operation node in the second working period are selected.

[0119] In this embodiment, after determining the second operating period, the electronic device can extract production data generated by the upstream operating node during the second operating period from the manufacturing database. This production data can specifically be time-stamped production data from the second operating period. This production data can be a record of product changes. In addition to obtaining product-related production data, the electronic device can also obtain interaction data from associated users when performing production operations on the upstream operating node. This interaction data can include operational data between the associated user and production equipment on the production line, or interaction records between the associated user and other users during the production process.

[0120] In S505 , the sub-database is obtained according to the production data and the interaction data of all the upstream operation nodes and the production data of the first node in the first working period.

[0121] In this embodiment, the electronic device may construct a corresponding sub-database based on the corresponding production data and interaction data in the upstream operation node, and the production data of the target user in the first working period.

[0122] In an embodiment of the present application, by determining the first working period of the first node and determining the second working period of the upstream operation node, it is possible to extract data associated with the target user in the time dimension, thereby constructing a corresponding sub-database to improve the correlation between the data in the sub-database and the target user, and then improve the subsequent data search efficiency.

[0123] Figure 7 The fourth embodiment of the present application provides a smart management method for the manufacturing industry in S2032. Figure 7 , relative to Figure 4 In the embodiment, the intelligent management method applied to the manufacturing industry provided by this embodiment further includes, in S2032, S701 to S705, which are described in detail as follows:

[0124] In S701, according to the manufacturing field corresponding to the manufacturing database, a keyword library corresponding to the manufacturing field is obtained; the keyword library contains at least one professional keyword related to the manufacturing field.

[0125] In this embodiment, the electronic device can determine a keyword database corresponding to the manufacturing industry in which the target user is located. Since different industries use significantly different professional terminology, and in particular, for some English abbreviations, the same letter combination may have different meanings, the electronic device can determine a corresponding keyword database based on the manufacturing industry in which the target user is located to improve the accuracy of identifying the first task data uploaded by the target user.

[0126] In S702, abnormal characters of the target user in the historical conversation record are determined based on the historical conversation record of the target user.

[0127] In this embodiment, the electronic device stores a user database containing chat records between a target user and other users, i.e., the aforementioned historical conversation records. The electronic device can select, based on the target user's user ID, chat records containing the target user ID from all historical conversation records, thereby obtaining the aforementioned historical conversation records.

[0128] In this embodiment, the electronic device can import the above historical conversation records into the above anomaly detection algorithm to determine abnormal characters contained in the historical conversation records. The abnormal characters are specifically typos that are abnormally input by the target user during the conversation.

[0129] In S703 , abnormal feature data of the target user is obtained according to the abnormal character and the associated characters adjacent to the abnormal character, and text calibration is performed on the first task data using the abnormal feature data to obtain calibrated text data.

[0130] In this embodiment, the electronic device can mark the abnormal characters identified above in the historical dialogue records, determine other characters adjacent to the abnormal characters in the historical dialogue records, and use these other characters adjacent to the abnormal characters as the associated characters. Therefore, based on the combination of abnormal characters and associated characters, the abnormal input bias of the target user during input can be determined. Since the input method used by the target user when inputting relevant text is generally the same as the input method used during the chat process, the corresponding abnormal input preference can be obtained through the combination of abnormal characters and associated characters during the chat process, and corresponding abnormal feature data can be generated based on the abnormal input preference. Specifically, the aforementioned abnormal feature data can be the input content corresponding to the character combination used to input abnormal characters and associated characters, such as pinyin combinations or stroke combinations.

[0131] In this embodiment, after determining the above-mentioned abnormal feature data, the electronic device can detect whether the character combination in the above-mentioned first task data matches the above-mentioned abnormal feature data. For example, it can detect whether there is a character combination based on the same pinyin combination in the first task data. If so, the above-mentioned character combination may contain abnormal characters. At this time, any character in the character combination can be corrected, and so on, until all character combinations in the first task data are matched with abnormal features to obtain corresponding calibration text data.

[0132] In S704, each of the professional keywords in the keyword library is converted into an expression keyword corresponding to the language type.

[0133] In this embodiment, the electronic device can perform word conversion on the professional keywords in the keyword library according to the language type of the target user, thereby generating expression keywords expressed based on the language type of the target user.

[0134] For example, if the target user's language type is English and the keyword library is constructed in Chinese, the electronic device can convert the keywords expressed in Chinese into keywords expressed in English, namely the above-mentioned expression keywords.

[0135] In S705 , each of the expression keywords is matched with the calibration text data in sequence, and the calibration task data is generated based on the matching results.

[0136] In this embodiment, the electronic device can detect whether the above-mentioned calibration text data contains professional keywords in the manufacturing field. Therefore, the expression keywords corresponding to the language type can be matched with each character in the calibration text data in sequence, so that the professional keywords contained in the calibration text data can be determined according to the matching results, and the content tendency corresponding to the calibration text data can be determined, so that the content tendency can be combined with the above-mentioned calibration text data to obtain the above-mentioned calibration task data.

[0137] In the embodiment of the present application, the first task data is calibrated and professional words are recognized through language type and historical conversation records, which can improve the accuracy of the calibrated task data.

[0138] Figure 8 The fifth embodiment of the present application provides a specific implementation flow chart of an intelligent management method applied to the manufacturing industry in S2034. Figure 8 In contrast Figure 4 In the embodiment, an intelligent management method applied to the manufacturing industry provided in this embodiment includes: S801 to S804, which are described in detail as follows:

[0139] In S801, the calibration task data and the second task data are imported into the convolution layer of the neural network to generate a text feature vector corresponding to the task data.

[0140] In this embodiment, the electronic device can import the calibration task data corresponding to the target user and the second task data of other users into the convolutional layer within the corresponding neural network, and convert the text data into a text feature vector expressed in a vector, so that redundant characters in the text data can be filtered, thereby obtaining feature words that have a strong correlation with the manufacturing field, and constructing a corresponding text feature vector based on the feature words.

[0141] In S802, the matching degree between the candidate keywords in the manufacturing field and the text feature vector is calculated respectively; the matching degree is:

[0142]

[0143] Among them, MatchPiont i is the matching degree between the i-th candidate keyword and the text feature vector; Text(j) is the j-th text keyword in the text feature vector; M is the total number of text keywords in the text feature vector; Keywork i is the i-th candidate keyword; BaseDis is the preset benchmark word vector distance; Min[x] is the minimum value selection function.

[0144] In this embodiment, the electronic device may calculate a matching degree between multiple candidate keywords in the manufacturing field and a text feature vector, and determine whether the text feature vector contains associated characters that match the candidate keywords. A greater matching degree between the candidate keywords and the text feature vector indicates that the task calibration data and the second task data may contain the candidate keywords. Conversely, a lower matching degree indicates that the two sets of task data may not contain the candidate keywords.

[0145] In this embodiment, the electronic device can calculate the word vector distance between the candidate keyword and each keyword in the text feature vector, and determine the corresponding matching degree based on the minimum value of the word vector distance.

[0146] In S803 , based on the matching degree, a target keyword associated with the text feature vector is selected from all the candidate keywords.

[0147] In this embodiment, the electronic device may select, based on the matching degrees corresponding to the candidate keywords, a candidate keyword with a matching degree greater than a preset matching threshold as a target keyword included in the text feature vector.

[0148] In S804, the SQL statement is generated according to all target keywords.

[0149] In this embodiment, the electronic device can generate corresponding SQL statements based on all identified target keywords. For example, corresponding SQL statements can be generated for different target keywords respectively, or a combination of two or more target keywords can be used to generate corresponding SQL statements.

[0150] In an embodiment of the present application, by calculating the matching degree between the candidate keywords and the text feature vector, the target keywords contained in the task data can be determined, and then the corresponding SQL statements are generated to improve the accuracy of the response data obtained from subsequent searches.

[0151] Figure 9 The sixth embodiment of the present application provides a specific implementation flow chart of an intelligent management method applied to the manufacturing industry in S202. Figure 9 , relative to Figure 2-8 In any of the embodiments, the present embodiment provides an intelligent management method applied to the manufacturing industry, including, in S202, S2021 to S2025, which are described in detail as follows:

[0152] In S2021, if the request type is a composite request type, the data types included in the composite request type are determined; the data types include graphic data types and text data types.

[0153] In this embodiment, if the request type is a composite request type, it indicates that the task request requires the acquisition of multiple types of data. The data types required to be acquired include graphic data types and text data types. Optionally, the data types may also include audio data types and table data types, etc., which may be determined based on actual circumstances.

[0154] In S2022, a graphic data interface is called to obtain graphic task data from multiple target objects; the target objects include distributed terminals deployed on the production line to obtain image data and office applications corresponding to the production line.

[0155] In this embodiment, the graphic task data may include image data collected by distributed terminals and image data in office documents edited by the target user during the production process. Based on this, the electronic device can obtain the image task data associated with the target user from multiple different data interfaces.

[0156] Furthermore, as another embodiment of the present application, when the above-mentioned graphics task data is image data obtained through a distributed terminal, the electronic device may further process the image data, which may specifically include the following steps:

[0157] Step 1: Determine the dust index corresponding to when the distributed terminal acquires the original image data through the dust sensor on the distributed terminal.

[0158] In this embodiment, the manufacturing production line may produce a lot of dust during the production process, which may affect the image quality of the distributed terminals when acquiring image data. Based on this, when the electronic device obtains the raw image data fed back to the distributed terminals, it can use the dust sensor to determine the current dust density in the air and convert the dust density into a corresponding dust index.

[0159] Step 2: Process the original image data based on the image enhancement model corresponding to the dust index to obtain calibrated image data; the image enhancement model is trained based on the production equipment and products included in the production line.

[0160] In this embodiment, the electronic device can adjust the image enhancement level of the image enhancement model based on the dust index. A larger value of the dust index indicates a higher dust concentration, and the corresponding image enhancement level is higher. Conversely, a smaller value of the word segmentation index indicates a lower image enhancement level. For example, the image enhancement may include contrast enhancement, brightness enhancement, and saturation enhancement.

[0161] Step 3: Perform image segmentation on the calibration image data according to the shooting angle of the distributed terminal to obtain a target area image corresponding to the area where the production product is located.

[0162] Step 4: Generate the graphic task data based on the target area images collected by all distributed terminals.

[0163] In this embodiment, the electronic device can filter background graphics irrelevant to production, so as to obtain target area images relevant to production, thereby reducing the amount of image task data and increasing the amount of data processing required for subsequent processing.

[0164] In S2023, the text data interface is called to import the graphics task data into a preset graphics understanding model, and the first text data corresponding to the graphics task data is generated.

[0165] In this embodiment, the electronic device is configured with a graphic understanding model. The acquired image task data is imported into the above-mentioned graphic understanding model, and the graphic data can be converted into text data, so that the data content of the graphic task data can be determined through multiple dimensions, thereby improving the accuracy of subsequent task data.

[0166] In S2024, the second text data corresponding to the task request is obtained through the text data interface.

[0167] In S2025, the task data is obtained based on the graphic task data, the first text data, and the second text data.

[0168] In an embodiment of the present application, when corresponding to a task request containing multiple types of data types, the corresponding task data can be obtained through the corresponding data interface, and the non-text data can be converted through multiple dimensions to improve the accuracy of the data content of the task data.

[0169] In this embodiment, Figure 10 This application provides a structural block diagram of an intelligent management device for use in the manufacturing industry, according to an embodiment of the present application. The intelligent management device includes units for performing various tasks. Figure 2 The steps implemented by the first device in the corresponding embodiment are described in detail. Figure 2 and Figure 2 For the sake of convenience, only the parts related to this embodiment are shown.

[0170] See also Figure 10 Intelligent management devices applied in the manufacturing industry include:

[0171] The task request response unit 101 is configured to respond to a task request initiated by a target user and determine the request type of the task request; the task request includes a report push request, an early warning analysis request, and / or an intelligent question and answer request;

[0172] The task data acquisition unit 102 is configured to call a data interface associated with the request type to acquire task data associated with the task request; the data interface is determined based on a correspondence between the request type and the interface;

[0173] A response data generating unit 103 is used to import the task data into a neural network constructed based on a manufacturing database and output response data of the task request;

[0174] The task result feedback unit 104 is configured to convert the response data into a natural language conversion model associated with the target user to obtain a task feedback result of the task request.

[0175] It should be understood that Figure 10 In the structural block diagram of the device shown, each module is used to execute Figures 2 to 9 The steps in the corresponding embodiments, and Figures 2 to 9 Each step in the corresponding embodiment has been explained in detail in the above embodiment. Figures 2 to 9 as well as Figures 2 to 9 The relevant descriptions in the corresponding embodiments will not be repeated here.

[0176] Figure 11 This is a structural block diagram of an electronic device provided by another embodiment of the present application. Figure 11 As shown, the electronic device 1100 of this embodiment includes: a processor 1110, a memory 1120, and a computer program 1130 stored in the memory 1120 and executable by the processor 1110, such as a program for an intelligent management method applied to the manufacturing field. When the processor 1110 executes the computer program 1130, the steps in each embodiment of the intelligent management method applied to the manufacturing field are implemented, such as Figure 2 Alternatively, the processor 1110 executes the computer program 1130 to implement the above Figure 11 The functions of each module in the corresponding embodiment are, for example, Figure 10 For details on the functions of the units 101 to 104, please refer to Figure 11 Related description in the corresponding embodiment.

[0177] Exemplarily, the computer program 1130 may be divided into one or more modules, one or more of which are stored in the memory 1120 and executed by the processor 1110 to complete the present application. One or more modules may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program 1130 in the electronic device 1100. For example, the computer program 1130 may be divided into various unit modules, and the specific functions of each module are as described above.

[0178] The electronic device 1100 may include, but is not limited to, a processor 1110 and a memory 1120. Those skilled in the art will appreciate that Figure 11 This is merely an example of electronic device 1100 and does not constitute a limitation on electronic device 1100. It may include more or fewer components than shown, or combine certain components, or different components. For example, electronic device may also include input / output devices, network access devices, buses, etc.

[0179] The processor 1110 may be a central processing unit, or it may be other general-purpose processors, digital signal processors, application-specific integrated circuits, off-the-shelf programmable gate arrays or other programmable logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0180] The memory 1120 can be an internal storage unit of the electronic device 1100, such as a hard disk or memory of the electronic device 1100. The memory 1120 can also be an external storage device of the electronic device 1100, such as a plug-in hard disk, smart memory card, flash memory card, etc. equipped on the electronic device 1100. Furthermore, the memory 1120 can include both internal storage units and external storage devices of the electronic device 1100.

[0181] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. An intelligent management method applied in the manufacturing industry, characterized in that: include: In response to a task request corresponding to a target user, determining a request type of the task request; The task request includes a report push request, an early warning analysis request and / or an intelligent question and answer request; Calling a data interface associated with the request type to obtain task data associated with the task request; The data interface is determined according to the corresponding relationship between the request type and the interface; Importing the task data into a neural network constructed based on a manufacturing database, and outputting response data of the task request; A natural language conversion model that associates the response data with the target user to obtain a task feedback result of the task request; The step of importing the task data into a neural network constructed based on a manufacturing database and outputting response data of the task request includes: Acquire user information of the target user; the user information includes the first node of the target user on the manufacturing production line and the language type of the target user; Performing text conversion on the first task data according to the language type to obtain calibration task data corresponding to the first task data; the first task data is the task data uploaded by the target user; Determine an upstream operation node of the first node based on the manufacturing production line, and screen the manufacturing database according to the first node and all the upstream operation nodes to obtain a sub-database associated with the target user; the data in the sub-database is data associated with the first node and all the upstream operation nodes; Importing the calibration task data and the second task data into the neural network, generating an SQL statement corresponding to the task request, and performing a data search in the sub-database through the SQL statement to obtain the response data; the second task data is other task data except the first task data; The step of importing the calibration task data and the second task data into the neural network, generating an SQL statement corresponding to the task request, and performing a data search in the sub-database through the SQL statement to obtain the response data includes: Importing the calibration task data and the second task data into the convolution layer of the neural network to generate a text feature vector corresponding to the task data; The matching degree between the candidate keywords in the manufacturing field and the text feature vector is calculated respectively; the matching degree is: Among them, MatchPoint i is the matching degree between the i-th candidate keyword and the text feature vector; Text(j) is the j-th text keyword in the text feature vector; M is the total number of text keywords in the text feature vector; Keywork i is the i-th candidate keyword; BaseDis is the preset benchmark word vector distance; Min[x] is the minimum value selection function; Based on the matching degree, selecting a target keyword associated with the text feature vector from all the candidate keywords; The SQL statement is generated according to all target keywords.

2. The method according to claim 1, characterized in that The determining of the upstream operation node of the first node based on the manufacturing production line, and screening the manufacturing database according to the first node and all the upstream operation nodes to obtain a sub-database associated with the target user, includes: According to the manufacturing field corresponding to the target user, the manufacturing production line in the manufacturing field is acquired; the manufacturing production line includes a plurality of operation nodes, each operation node corresponds to an operation sequence; Based on the operation path of the first node on the manufacturing production line, selecting other operation nodes whose operation sequence is earlier than that of the first node as the upstream operation nodes; Determine the second working period associated with each upstream operation node according to the first working period corresponding to the target user and the operation sequence; the time interval between the second working period and the first working period is determined according to the operation sequence and the process duration corresponding to each operation; Based on the manufacturing database, selecting production data of the upstream operation node in the second working period, and interaction data of users associated with the upstream operation node in the second working period; The sub-database is obtained according to the production data and the interaction data of all the upstream operation nodes and the production data of the first node in the first working period.

3. The intelligent management method according to claim 1, characterized in that: performing text conversion on the first task data according to the language type to obtain calibration task data corresponding to the first task data; The first task data is the task data uploaded by the target user, including: According to the manufacturing field corresponding to the manufacturing database, a keyword library corresponding to the manufacturing field is obtained; the keyword library contains at least one professional keyword related to the manufacturing field; According to the historical conversation records of the target user, determining abnormal characters of the target user in the historical conversation records; Obtaining abnormal feature data of the target user according to the abnormal character and the associated characters adjacent to the abnormal character, and performing text calibration on the first task data according to the abnormal feature data to obtain calibrated text data; Convert each of the professional keywords in the keyword library into expression keywords corresponding to the language type; Each of the expression keywords is matched with the calibration text data in sequence, and the calibration task data is generated based on the matching results.

4. The intelligent management method according to any one of claims 1 to 3, characterized in that: The calling of the data interface associated with the request type to obtain the task data associated with the task request includes: If the request type is a composite request type, determining the data types included in the composite request type; the data types include graphic data types and text data types; Calling a graphic data interface to obtain graphic task data from a plurality of target objects; the target objects include a distributed terminal deployed on a production line to obtain image data and an office application corresponding to the production line; Calling a text data interface to import the graphic task data into a preset graphic understanding model, and generating first text data corresponding to the graphic task data; Acquire second text data corresponding to the task request through the text data interface; The task data is obtained according to the graphic task data, the first text data and the second text data.

5. The intelligent management method according to claim 4, characterized in that: The calling of the graphics data interface to obtain graphics task data from multiple target objects includes: Determining, by means of a dust sensor on the distributed terminal, a dust index corresponding to when the distributed terminal acquires raw image data; The original image data is processed based on the image enhancement model corresponding to the dust index to obtain calibrated image data; the image enhancement model is trained based on the production equipment and products included in the production line; According to the shooting angle of the distributed terminal, the calibration image data is segmented to obtain a target area image corresponding to the area where the production product is located; The graphic task data is generated based on the target area images collected by all distributed terminals.

6. An intelligent management device applied in the manufacturing field, characterized in that: include: A task request response unit, configured to respond to a task request initiated by a target user and determine a request type of the task request; The task request includes a report push request, an early warning analysis request and / or an intelligent question and answer request; A task data acquisition unit, used for calling a data interface associated with the request type to acquire task data associated with the task request; The data interface is determined according to the corresponding relationship between the request type and the interface; A response data generating unit, used for importing the task data into a neural network constructed based on a manufacturing database, and outputting response data of the task request; A task result feedback unit, used for converting the response data into a natural language conversion model associated with the target user to obtain a task feedback result of the task request; The response data generating unit is specifically used for: Acquire user information of the target user; the user information includes the first node of the target user on the manufacturing production line and the language type of the target user; Performing text conversion on the first task data according to the language type to obtain calibration task data corresponding to the first task data; the first task data is the task data uploaded by the target user; Determine an upstream operation node of the first node based on the manufacturing production line, and perform data screening on a manufacturing database according to the first node and all the upstream operation nodes to obtain a sub-database associated with the target user; The data in the sub-database is data associated with the first node and all the upstream operation nodes; Importing the calibration task data and the second task data into the neural network, generating an SQL statement corresponding to the task request, and performing a data search in the sub-database through the SQL statement to obtain the response data; The second task data is other task data except the first task data; The step of importing the calibration task data and the second task data into the neural network, generating an SQL statement corresponding to the task request, and performing a data search in the sub-database through the SQL statement to obtain the response data includes: Importing the calibration task data and the second task data into the convolution layer of the neural network to generate a text feature vector corresponding to the task data; The matching degree between the candidate keywords in the manufacturing field and the text feature vector is calculated respectively; the matching degree is: Among them, MatchPoint i is the matching degree between the i-th candidate keyword and the text feature vector; Text(j) is the j-th text keyword in the text feature vector; M is the total number of text keywords in the text feature vector; Keywork i is the i-th candidate keyword; BaseDis is the preset benchmark word vector distance; Min[x] is the minimum value selection function; Based on the matching degree, selecting a target keyword associated with the text feature vector from all the candidate keywords; The SQL statement is generated according to all target keywords.

7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images