A data acquisition method for MES production management system

Abstract

The present invention discloses a data acquisition method for an MES production management system. The method comprises acquiring MES system production management data for management, automatically classifying all equipment production management data types in the MES system production management data using a single-layer feedforward neural network to obtain a data set of equipment usage mean values for management; performing data mining based on the field of equipment production management data type identification using a single-layer feedforward neural network; setting algorithm parameters and production management alarm signals for abnormal equipment production management data types through a user setting interface based on an MES server; feeding each frame of equipment production management monitoring image acquired by equipment production management monitoring into an equipment production management data type identification model to obtain an identification result of the equipment production management data type; feeding the identification result of the equipment production management data type into a Deep‑sort algorithm to track the production management data types. The present invention is used to accurately and efficiently identify equipment production management data type information.

Description

Technical Field

[0001] The present invention relates to the field of production management and supervision, and in particular to a data acquisition method for an MES production management system. Background Art

[0002] The MES production management system is a production information management system for the execution layer of manufacturing enterprises' workshops. It provides enterprises with management modules including manufacturing data management, planning and scheduling management, production dispatch management, inventory management, quality management, human resources management, work center / equipment management, tool and fixture management, procurement management, cost management, project kanban management, production process control, bottom-level data integration and analysis, and top-level data integration and decomposition, creating a solid, reliable, comprehensive, and feasible manufacturing collaborative management platform for enterprises.

[0003] Real-time and accurate production data collection is the foundation of the MES production management system business. The accuracy of existing data collection is not enough and needs further optimization. Summary of the Invention

[0004] In order to overcome the shortcomings and deficiencies of the prior art, the present invention provides a data acquisition method for an MES production management system.

[0005] The technical solution adopted by the present invention is a data acquisition method for an MES production management system provided by the present invention, comprising the following steps:

[0006] Step S1: multiple data acquisition units are set to collect data, obtain MES system production management data for management, and use a single-layer feedforward neural network to automatically classify all equipment production management data types in the MES system production management data to obtain a data set of equipment usage mean values for management;

[0007] Step S2, determining whether each data acquisition unit performs time synchronization; if synchronized, setting the synchronization time, starting to cycle the data acquisition units, and using the device usage mean data set to perform data mining based on the field of equipment production management data type identification using a single-layer feedforward neural network;

[0008] Step S3: determine whether the current data is valid. If it is valid, read the configuration parameters of the data acquisition unit, and set the algorithm parameters and the abnormal production management alarm signal of the equipment production management data type through the user setting interface based on the MES server;

[0009] Step S4, obtaining equipment production management monitoring images and video data, and feeding each frame of equipment production management monitoring images obtained by the equipment production management monitoring into the equipment production management data type recognition model to obtain equipment production management data type recognition results;

[0010] Step S5: determine whether the collected data needs to be cleaned. If so, call the corresponding cleaning method, feed the equipment production management data type identification result into the Deep-sort algorithm, and track the identified equipment production management data types;

[0011] Step S6: Identify and determine whether the tracked equipment production management data type is located in the equipment production management data type abnormal production management alarm signal type library, and determine whether the equipment production management data type is missing, save all collected business logic data, and upload it to the MES server through a standardized data interface.

[0012] Preferably, after using a single-layer feedforward neural network to perform data mining in the field of equipment production management data type identification, the parameters of the single-layer feedforward neural network algorithm are dynamically updated.

[0013] Preferably, determining whether the equipment production management data type is in the equipment production management data type abnormal production management alarm signal type library includes the following steps:

[0014] Step D1, obtaining the time and frequency of an abnormal production management alarm signal of a certain equipment production management data type in the current equipment production management monitoring image frame and the coordinates of the abnormal production management alarm signal location;

[0015] Step D2: If the time and frequency of the equipment production management data type abnormal production management alarm signal type library are both within the equipment production management data type abnormal production management alarm signal type library, it is determined that the equipment production management data type is missing;

[0016] Step D3: If only one side point of the equipment production management data type abnormal production management alarm signal is within the equipment production management data type abnormal production management alarm signal type library, proceed to step D4;

[0017] Step D4, determining whether the equipment production management data type location is within the equipment production management data type abnormal production management alarm signal type library; if the equipment production management data type location is within the equipment production management data type abnormal production management alarm signal type library, determining that the equipment production management data type is missing, otherwise proceed to the next step;

[0018] Step D5 , continuing to obtain other identification equipment production management data types in the current equipment production management monitoring image frame and determining whether there are any missing equipment production management data types.

[0019] The present application also includes a method for determining whether equipment production management data type is abnormal data in an equipment production management data type abnormal production management alarm signal type library, comprising the following steps:

[0020] Step T1, calculating the pixel distance between the center coordinates of the equipment production management data type in the previous frame of the equipment production management monitoring image and the center coordinates of the equipment production management data type in the current frame of the equipment production management monitoring image. If the pixel distance exceeds a threshold, it is determined that the equipment production management data type is abnormal;

[0021] Step T2: if it is determined that the equipment production management data type is in an abnormal state, the missing time period in the equipment production management data type information is cleared and equipment production management monitoring images of other equipment production management data types are processed;

[0022] Step T3, if the equipment production management data type is not moving, obtain the current time and calculate the cumulative time of the equipment production management data type stopping, compare the cumulative time of the equipment production management data type stopping with the set missing time threshold, if the cumulative time of the equipment production management data type stopping exceeds the missing time threshold, it is determined that the equipment production management data type is missing, if the cumulative time of the equipment production management data type stopping does not exceed the missing time, then the current equipment production management data type processing is completed, and continue to process other identification equipment production management monitoring images.

[0023] Preferably, the threshold adopts a dynamic form threshold of D / U to judge abnormal data of equipment production management data type, wherein the letter D represents the response speed of the abnormal production management alarm signal type library, and the letter U is the fluctuation function of the threshold.

[0024] Preferably, if a certain equipment production management data type is tracked in the previous frame of the equipment production management monitoring image in the tracking information, but the equipment production management data type is not tracked in the current frame, the platform sets a maximum number of missing frames. Before the maximum number of missing frames is reached, the platform does not determine that the equipment production management data type is lost. Then, the unscented Kalman filter in the Deep-sort algorithm is used to predict the equipment production management data type area of the current frame based on the position of the equipment production management data type area in the previous frame, and the predicted result is used as the equipment production management data type area of the current frame;

[0025] If the position of the equipment production management data category area in the next frame of equipment production management monitoring image matches the equipment production management data category area in the current frame, it is determined that the equipment production management data category disappears due to an error in the recognition algorithm;

[0026] If the maximum number of missing frames is reached, the device production management data type is directly deemed to have disappeared, and the platform deletes the tracking information of this device production management data type;

[0027] If the equipment production management data type reappears in a time period that does not reach the maximum number of missing frames, it is determined that the equipment production management data type is missing for a short period of time.

[0028] Preferably, the time for acquiring the equipment production management monitoring image is set according to the monitoring requirements, and the number of frames acquired per second is set according to actual conditions.

[0029] The method of the present application is implemented by an equipment production management data type model recognition module, an algorithm equipment production management monitoring image calculation unit, and a data missing judgment unit, wherein:

[0030] The equipment production management data type model identification module is used to identify the equipment production management data type of the equipment production management monitoring image and video data obtained by the algorithm equipment production management monitoring image calculation unit and obtain the abnormal production management alarm signal and abnormal production management alarm signal information of the equipment production management data type;

[0031] The algorithm equipment production management monitoring image calculation unit is used to obtain equipment production management monitoring images and video data, set the supervision area location, and match the equipment production management data type abnormal production management alarm signal and abnormal production management alarm signal information obtained by the equipment production management data type model recognition module with a specific identification code and then transmit it to the data missing judgment unit;

[0032] After receiving the abnormal production management alarm signal and abnormal production management alarm signal information of the equipment production management data type that matches the specific identification code, the data missing judgment unit outputs the tracker information, searches for each of the tracker information, and determines whether the equipment production management data type is located within the supervision area according to the location of the supervision area, and updates the tracker information again, and determines whether the equipment production management data type should be alarmed according to the updated tracker information.

[0033] Preferably, the tracker information includes: equipment production management data type identification code, equipment production management data type abnormal production management alarm signal, whether the equipment production management data type enters the equipment production management data type abnormal production management alarm signal, the time when the equipment production management data type enters the equipment production management data type abnormal production management alarm signal, and whether the equipment production management data type has been issued an alarm.

[0034] Preferably, the device can increase the amount of management data by using the mean value data set in a real-time updating manner.

[0035] The present invention applies the equipment production management data type recognition technology of a single-layer feedforward neural network to equipment production management data type recognition. This method can accurately identify equipment production management data types from equipment production management monitoring images and track the types of equipment production management data. Then, through a series of logical judgments on the types of equipment production management data missing, it accurately and efficiently identifies the types of equipment production management data and generates an alarm, thereby realizing intelligent supervision of equipment production management data type missing. This method significantly improves staff efficiency and also saves a lot of manpower and material resources.

[0036] Equipment Production Management Data Type Model Identification Module: This module utilizes a dataset of average equipment usage to perform data mining on a single-layer feedforward neural network based on the equipment production management data type identification domain, ensuring accurate identification of equipment production management data types. This module also dynamically updates the parameters of the single-layer feedforward neural network algorithm to ensure optimal performance within a limited budget.

[0037] Algorithmic equipment production management monitoring image calculation unit: The present invention tracks the identified equipment production management data types based on the tracking algorithm and maintains its individual structure to judge the status of the equipment production management data types and set a tracking failure threshold to avoid false alarms due to missing data;

[0038] Data missing judgment unit: The present invention first determines whether the tracked equipment production management data type is within the supervision area, and proposes multiple methods for judging abnormal data in equipment production management data types. By judging abnormal data types of equipment production management data types, the problem of false reporting of abnormal data equipment production management data types in traditional methods is avoided;

[0039] The present invention can conveniently set the missing identification algorithm from the MES server page, and the alarm message can also be displayed in real time on the MES server page for staff to view. Together with the core identification algorithm, it forms an intelligent identification platform for equipment production management data types.

[0040] The method proposed in the present invention can effectively control the production management in the MES system, thereby protecting the effectiveness of user data management and assisting in building a safe and effective MES system. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a flow chart of a data acquisition method for an MES production management system of the present invention;

[0042] Figure 2 To determine whether the equipment production management data type is abnormal or not, a production management alarm signal flow chart is provided;

[0043] Figure 3 A flow chart of a method for determining whether equipment production management data type is abnormal data within an equipment production management data type abnormal production management alarm signal according to the present invention;

[0044] Figure 4 This is a functional structure diagram of the platform of the present invention. DETAILED DESCRIPTION

[0045] It should be noted that, unless there is a conflict, the embodiments and features in the embodiments of this application can be combined with each other. The present application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0046] like Figure 1 As shown, a data acquisition method for an MES production management system includes the following steps:

[0047] Step S1: multiple data acquisition units are set to collect data, obtain MES system production management data for management, and use a single-layer feedforward neural network to automatically classify all equipment production management data types in the MES system production management data to obtain a data set of equipment usage mean values for management;

[0048] Step S2, determining whether each data acquisition unit performs time synchronization; if synchronized, setting the synchronization time, starting to cycle the data acquisition units, and using the device usage mean data set to perform data mining based on the field of equipment production management data type identification using a single-layer feedforward neural network;

[0049] Step S3: determine whether the current data is valid. If it is valid, read the configuration parameters of the data acquisition unit, and set the algorithm parameters and the abnormal production management alarm signal of the equipment production management data type through the user setting interface based on the MES server;

[0050] Step S4, obtaining equipment production management monitoring images and video data, and feeding each frame of equipment production management monitoring images obtained by the equipment production management monitoring into the equipment production management data type recognition model to obtain equipment production management data type recognition results;

[0051] Step S5: determine whether the collected data needs to be cleaned. If so, call the corresponding cleaning method, feed the equipment production management data type identification result into the Deep-sort algorithm, and track the identified equipment production management data types;

[0052] Step S6: Identify and determine whether the tracked equipment production management data type is located in the equipment production management data type abnormal production management alarm signal type library, and determine whether the equipment production management data type is missing, save all collected business logic data, and upload it to the MES server through a standardized data interface.

[0053] After using a single-layer feedforward neural network to conduct data mining in the field of equipment production management data type identification, the parameters of the single-layer feedforward neural network algorithm are dynamically updated.

[0054] like Figure 2 As shown, judging whether the equipment production management data type is in the equipment production management data type abnormal production management alarm signal type library includes the following steps:

[0055] Step D1, obtaining the time and frequency of an abnormal production management alarm signal type library of a certain equipment production management data type in the current equipment production management monitoring image frame and the coordinates of the abnormal production management alarm signal location;

[0056] Step D2: If the time and frequency of the equipment production management data type abnormal production management alarm signal type library are both within the equipment production management data type abnormal production management alarm signal type library, it is determined that the equipment production management data type is missing;

[0057] Step D3: If only one side point is in the equipment production management data type abnormal production management alarm signal type library, then proceed to step D4;

[0058] Step D4, determining whether the equipment production management data type location is within the equipment production management data type abnormal production management alarm signal type library; if the equipment production management data type location is within the equipment production management data type abnormal production management alarm signal type library, determining that the equipment production management data type is missing, otherwise proceeding to the next step;

[0059] Step D5 , continuing to obtain other identification equipment production management data types in the current equipment production management monitoring image frame and determining whether there are any missing equipment production management data types.

[0060] like Figure 3 As shown, the present application also includes a method for determining whether the equipment production management data type is abnormal data in the equipment production management data type abnormal production management alarm signal type library, including the following steps:

[0061] Step T1, calculating the pixel distance between the center coordinates of the equipment production management data type in the previous frame of the equipment production management monitoring image and the center coordinates of the equipment production management data type in the current frame of the equipment production management monitoring image. If the pixel distance exceeds a threshold, it is determined that the equipment production management data type is abnormal;

[0062] Step T2: if it is determined that the equipment production management data type is in an abnormal state, the missing time period in the equipment production management data type information is cleared and equipment production management monitoring images of other equipment production management data types are processed;

[0063] Step T3, if the equipment production management data type is not moving, obtain the current time and calculate the cumulative time of the equipment production management data type stopping, compare the cumulative time of the equipment production management data type stopping with the set missing time threshold, if the cumulative time of the equipment production management data type stopping exceeds the missing time threshold, it is determined that the equipment production management data type is missing, if the cumulative time of the equipment production management data type stopping does not exceed the missing time, then the current equipment production management data type processing is completed, and continue to process other identification equipment production management monitoring images.

[0064] The threshold uses the dynamic form of D / U to judge the abnormal data of the equipment production management data type, where the letter D represents the response speed of the abnormal production management alarm signal, and the letter U is the fluctuation function of the threshold.

[0065] If a certain equipment production management data type is tracked in the previous frame of the equipment production management monitoring image in the tracking information, but the equipment production management data type is not tracked in the current frame, the platform sets a maximum number of missing frames. Before the maximum number of missing frames is reached, the platform does not determine that the equipment production management data type is lost. Then, the unscented Kalman filter in the Deep-sort algorithm is used to predict the equipment production management data type area of the current frame based on the position of the equipment production management data type area in the previous frame. The predicted result is used as the equipment production management data type area of the current frame;

[0066] If the position of the equipment production management data category area in the next frame of equipment production management monitoring image matches the equipment production management data category area in the current frame, it is determined that the equipment production management data category disappears due to an error in the recognition algorithm;

[0067] If the maximum number of missing frames is reached, the device production management data type is directly deemed to have disappeared, and the platform deletes the tracking information of this device production management data type;

[0068] If the equipment production management data type reappears in a time period that does not reach the maximum number of missing frames, it is determined that the equipment production management data type is missing for a short period of time.

[0069] The time for acquiring the equipment production management monitoring image is set according to the monitoring requirements, and the number of frames acquired per second is set according to the actual situation.

[0070] like Figure 4 As shown, a data acquisition method of an MES production management system of the present application includes:

[0071] The equipment production management data type model identification module first obtains the data used for management based on the actual scenario. After obtaining the data, it uses a single-layer feedforward neural network to automatically classify all equipment production management data types in the MES system production management data to obtain the equipment usage mean data set used for management.

[0072] In order to better ensure the management effect, the present invention can greatly increase the amount of management data by using the mean data set of the management device and adopting data enhancement methods such as translation, flipping, and scaling.

[0073] Then, a single-layer feedforward neural network is used to perform data mining based on the equipment production management data type identification field by using the mean data set of the management equipment to ensure the accuracy of the model.

[0074] After obtaining the managed model, the traditional method generally directly deploys and utilizes the managed original model in combination with the business program. This method not only has a slow inference speed but also consumes a lot of hardware resources.

[0075] In order to ensure the real-time identification of equipment production management data types in the business, the original model in the present invention uses the particle swarm algorithm to perform operations such as speed update, position update, weight quantization, etc. to optimize the model reasoning throughput, perform forward reasoning, and accelerate reasoning.

[0076] Algorithm equipment production management monitoring image calculation unit, first, the intelligent recognition platform of the present invention sets the algorithm parameters through the user setting interface based on the MES server, and the MES server page displays the screen for setting equipment production management monitoring in real time. The user can use the mouse to draw the equipment production management data type abnormal production management alarm signal in the screen. When the user finishes drawing, the algorithm settings will be sent to the algorithm server through the network.

[0077] After receiving the settings, the algorithm will obtain the real-time data of the set equipment production management monitoring according to the specified video stream address, and send each frame of the equipment production management monitoring image obtained into the managed equipment production management data type recognition model to obtain the equipment production management data type recognition result.

[0078] The equipment production management data type identification results are then fed into the Deep-sort algorithm to track the identified equipment production management data types.

[0079] The benefit of tracking is that it matches the types of production management data of the same equipment in the time series and assigns the same identification code, avoiding the problem of continuous alarms caused by continuous identification of the same equipment.

[0080] The data missing judgment unit, based on the tracker tracking information of the algorithm equipment production management monitoring image calculation unit, traverses all tracked equipment production management data types, and initializes the information of newly tracked equipment production management data types;

[0081] If the platform has tracked this equipment production management data type before, and the current frame tracks this equipment production management data type again, it is first determined whether it is in the equipment production management data type abnormal production management alarm signal type library, and the equipment production management data type abnormal production management alarm signal time and frequency and abnormal production management alarm signal location coordinates are used to determine whether the equipment production management data type is in the equipment production management data type abnormal production management alarm signal type library.

[0082] The specific method is as follows: first, obtain the time and frequency of the abnormal production management alarm signal and the coordinates of the abnormal production management alarm signal location;

[0083] If both the time and frequency are in the equipment production management data type abnormal production management alarm signal type library, it is determined that the equipment production management data type is missing;

[0084] If only one side point is located in the equipment production management data type abnormal production management alarm signal type library, then determine whether the equipment production management data type location is located in the equipment production management data type abnormal production management alarm signal type library. If it is, then determine that the equipment production management data type is missing; the rest are deemed to be non-existent or missing.

[0085] If the equipment production management data type is not located in the equipment production management data type abnormal production management alarm signal type library, the current equipment production management data type processing is completed, and other identification equipment production management monitoring images are processed. When all equipment production management data types are traversed, the current frame processing is completed and the next frame is obtained.

[0086] If the equipment production management data type is located in the equipment production management data type abnormal production management alarm signal type library, the present invention adopts a method to determine whether the equipment production management data type is abnormal data to avoid the problem of equipment production management data type being mistakenly considered to be missing and an alarm being issued because it is in the equipment production management data type abnormal production management alarm signal type library.

[0087] The embodiments are:

[0088] (1) Calculate the pixel distance between the center coordinates of the equipment production management data type in the previous frame of the equipment production management monitoring image and the center coordinates of the equipment production management data type in the current frame. If this distance exceeds a certain threshold, it is determined that the equipment production management data type is abnormal;

[0089] (2) Since the types of equipment production management data are at different distances from the equipment production management monitoring, the Uox sizes identified by the types of equipment production management data at different distances from the equipment production management monitoring in the equipment production management monitoring image will be very different; for example, the actual distance between devices corresponding to a distance of 10 pixels far away is different from that corresponding to a distance of 10 pixels near.

[0090] Therefore, if the method in step (1) uses a single threshold as a judgment condition, it will result in different standards for judging abnormal data for different types of production management data of remote and nearby equipment.

[0091] The present invention proposes to use the dynamic threshold = D / U form to determine whether the type of equipment production management data is abnormal data;

[0092] Where D is the time distance of the abnormal production management alarm signal, and U is the fluctuation function of the threshold;

[0093] During the judgment process, U is a fixed value and can be adjusted by the algorithm personnel.

[0094] The effect of setting up a dynamic threshold is that when the library of abnormal production management alarm signal types is larger, the corresponding distance threshold is larger, and when the library of abnormal production management alarm signal types is smaller, the corresponding distance threshold is smaller, thereby making it more accurate to make abnormal judgments on the types of equipment production management data at different distances from the equipment production management monitoring.

[0095] If the equipment production management data type is abnormal, the missing time in the equipment production management data type information is cleared, and other identification equipment production management monitoring images are processed;

[0096] If the equipment production management data type is not abnormal, the current time is obtained and the cumulative time of the equipment production management data type is calculated, and the cumulative time of the equipment production management data type is compared with the missing time threshold. If the missing time threshold is exceeded, it is determined that the equipment production management data type is missing, and the equipment production management data type information is sent to the MES server through the network. After the MES server receives the alarm message, the alarm information will be displayed on the page; if it does not exceed the missing time, the current equipment production management data type processing is completed, and other identification equipment production management monitoring images will continue to be processed.

[0097] If a device was tracked in the previous frame but not in the current frame, the following three situations may occur:

[0098] 1. The recognition algorithm failed to detect the type of equipment production management data;

[0099] 2. The types of equipment production management data exceed the scope of the screen;

[0100] 3. There are omissions. Other objects will be missing the type of equipment production management data, resulting in the recognition algorithm being unable to recognize the type of equipment production management data.

[0101] The present invention addresses this situation by setting a maximum number of lost frames based on the three possible scenarios. The device production management data category is not considered lost until the maximum number of lost frames is reached. The unscented Kalman filter within the Deep-sort algorithm is used to predict the device production management data category area for the current frame based on the location of the device production management data category area in the previous frame. The predicted result is used as the device production management data category area for the current frame, but the tracking status remains "not tracked."

[0102] In the first case, if the recognition algorithm fails to detect the equipment production management data type, when the equipment production management data type area is recognized in the next frame, the equipment production management data type area obtained by the tracking algorithm will match the equipment production management data type area in the current frame, avoiding the problem of repeated alarms caused by unstable recognition algorithms;

[0103] For the second case, when the maximum number of missing frames is reached, the algorithm considers that the current device production management data type has disappeared and deletes the tracking information of this device production management data type;

[0104] For the third case, if it is a short-term loss, after the equipment production management data type reappears, the tracking algorithm can still match the previous equipment production management data type area with the missing equipment production management data type area, and identify it as the same equipment production management data type, avoiding the problem of repeated alarms.

[0105] It is recommended to set the maximum number of missing frames to the time for obtaining equipment production management monitoring images according to monitoring requirements, and set the number of frames obtained per second based on actual conditions.

[0106] The specific equipment production management data type model recognition module, the algorithm equipment production management monitoring image calculation unit and the data missing judgment unit communicate with each other through the following content:

[0107] The equipment production management data type model identification module generates a model file after management and acceleration;

[0108] The algorithm equipment production management monitoring image calculation unit first obtains the MES server settings through network communication; the MES server settings specifically include: equipment production management monitoring rtsU stream address, supervision area location;

[0109] After the setting is completed, the algorithm equipment production management monitoring image calculation unit loads the model file of the equipment production management data type model recognition module to identify the equipment production management data type on the equipment production management monitoring image and video data and obtains the abnormal production management alarm signal and abnormal production management alarm signal information of the equipment production management data type;

[0110] Then, the abnormal production management alarm signal is sent to the Deep-sort algorithm, and a specific identification code is matched for each abnormal production management alarm signal, and then the tracker information is transmitted to the data missing judgment unit to obtain the tracker information, wherein the tracker information includes the equipment production management data type identification code, the equipment production management data type abnormal production management alarm signal, whether the equipment production management data type enters the equipment production management data type abnormal production management alarm signal, the time when the equipment production management data type enters the equipment production management data type abnormal production management alarm signal, and whether the equipment production management data type has been issued an alarm;

[0111] The data missing judgment unit traverses each tracker information, determines whether the equipment production management data type is located in the supervision area according to the supervision area location, updates the tracker information, and determines whether an alarm should be issued for the equipment production management data type according to the tracker information.

[0112] It is worth noting that in the above system embodiment, the various units included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0113] In addition, those skilled in the art will appreciate that all or part of the steps in the above-mentioned embodiments can be accomplished by instructing related hardware through a program, and the corresponding program can be stored in a computer-readable storage medium.

[0114] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to the specific embodiments described. Obviously, many modifications and variations are possible based on the content of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. A data acquisition method for an MES production management system, characterized by: The following steps are involved: Step S1: multiple data acquisition units are set to collect data, obtain MES system production management data for management, and use a single-layer feedforward neural network to automatically classify all equipment production management data types in the MES system production management data to obtain a data set of equipment usage mean values for management; Step S2, determining whether each data acquisition unit performs time synchronization; If synchronized, the synchronization time is set, and a cyclic data acquisition unit is started, and the device is used to use the mean value data set to perform data mining based on the field of equipment production management data type identification using a single-layer feedforward neural network; Step S3: determine whether the current data is valid. If it is valid, read the configuration parameters of the data acquisition unit, and set the algorithm parameters and the abnormal production management alarm signal of the equipment production management data type through the user setting interface based on the MES server; Step S4, obtaining equipment production management monitoring image and video data, and feeding each frame of equipment production management monitoring image and video data obtained by the equipment production management monitoring into the equipment production management data type recognition model to obtain the equipment production management data type recognition result; Step S5: determine whether the collected data needs to be cleaned. If so, call the corresponding cleaning method, feed the equipment production management data type identification result into the Deep-sort algorithm, and track the identified equipment production management data types; Step S6: Identify and determine whether the equipment production management data type being tracked is in the equipment production management data type abnormal production management alarm signal type library, and determine whether the equipment production management data type is missing, save all collected equipment production management data, and upload it to the MES server through a standardized data interface.

2. The data acquisition method of the MES production management system according to claim 1, characterized in that: After using a single-layer feedforward neural network to conduct data mining in the field of equipment production management data type identification, the parameters of the single-layer feedforward neural network algorithm are dynamically updated.

3. The data acquisition method of the MES production management system according to claim 1, characterized in that: Determining whether the equipment production management data type is in the equipment production management data type abnormal production management alarm signal type library includes the following steps: Step D1, obtaining the time and frequency of an abnormal production management alarm signal of a certain equipment production management data type in the current equipment production management monitoring image frame and the coordinates of the abnormal production management alarm signal location; Step D2: If the time and frequency of the equipment production management data type abnormal production management alarm signal are both within the equipment production management data type abnormal production management alarm signal, it is determined that the equipment production management data type is missing; Step D3: If only one side point of the equipment production management data type abnormal production management alarm signal is within the equipment production management data type abnormal production management alarm signal type library, proceed to step D4; Step D4, determining whether the equipment production management data type location is within the equipment production management data type abnormal production management alarm signal type library; if the equipment production management data type location is within the equipment production management data type abnormal production management alarm signal type library, determining that the equipment production management data type is missing, otherwise proceeding to the next step; Step D5 , continuing to obtain other identification equipment production management data types in the current equipment production management monitoring image frame and determining whether there are any missing equipment production management data types.

4. The data acquisition method of the MES production management system according to claim 3, characterized in that: The invention also includes a method for determining whether the equipment production management data type is abnormal data in the equipment production management data type abnormal production management alarm signal type library, comprising the following steps: Step T1, calculating the pixel distance between the center coordinates of the equipment production management data type in the previous frame of the equipment production management monitoring image and the center coordinates of the equipment production management data type in the current frame of the equipment production management monitoring image. If the pixel distance exceeds a threshold, it is determined that the equipment production management data type is abnormal; Step T2: if it is determined that the equipment production management data type is in an abnormal state, the missing time period in the equipment production management data type information is cleared and equipment production management monitoring images of other equipment production management data types are processed; Step T3, if the equipment production management data type is not moving, obtain the current time and calculate the cumulative time of the equipment production management data type stopping, compare the cumulative time of the equipment production management data type stopping with the set missing time threshold, if the cumulative time of the equipment production management data type stopping exceeds the missing time threshold, it is determined that the equipment production management data type is missing, if the cumulative time of the equipment production management data type stopping does not exceed the missing time, then the current equipment production management data type processing is completed, and continue to process other identification equipment production management monitoring images.

5. The data acquisition method of the MES production management system according to claim 4, characterized in that: The threshold uses a dynamic form threshold of D / U to judge abnormal data of equipment production management data type, wherein the letter D represents the response speed of abnormal production management alarm signal, and the letter U is the fluctuation function of the threshold.

6. The data acquisition method of the MES production management system according to claim 4, characterized in that: If a certain equipment production management data type is tracked in the previous frame of the equipment production management monitoring image in the tracking information, but the equipment production management data type is not tracked in the current frame, the platform sets a maximum number of missing frames. Before the maximum number of missing frames is reached, the platform does not determine that the equipment production management data type is lost. Then, the unscented Kalman filter in the Deep-sort algorithm is used to predict the equipment production management data type area of the current frame based on the position of the equipment production management data type area in the previous frame. The predicted result is used as the equipment production management data type area of the current frame; If the next frame of equipment production management monitoring image recognizes that the position of the equipment production management data type area matches the equipment production management data type area of the current frame, it is judged that the equipment production management data type has disappeared due to an error in the recognition algorithm; if the maximum number of disappeared frames is reached, it is directly regarded as that the equipment production management data type has disappeared, and the platform deletes the tracking information of this equipment production management data type; if the equipment production management data type reappears in the time period before the maximum number of disappeared frames is reached, it is judged that the equipment production management data type is missing for a short time.

7. The data acquisition method of the MES production management system according to claim 6, characterized in that: The time for acquiring the equipment production management monitoring image is set according to the monitoring requirements, and the number of frames acquired per second is set according to the actual situation.

8. The data acquisition method of the MES production management system according to claim 1, characterized in that: The method is implemented through an equipment production management data type identification model module, an algorithm equipment production management monitoring image calculation unit and a data missing judgment unit; The equipment production management data type identification model module is used to identify the equipment production management data type of the equipment production management monitoring images and video data acquired by the algorithm equipment production management monitoring image calculation unit and obtain abnormal production management alarm signals and abnormal production management alarm signal information of the equipment production management data type; The algorithm equipment production management monitoring image calculation unit is used to obtain equipment production management monitoring images and video data, set the supervision area location, and match the equipment production management data type abnormal production management alarm signal and abnormal production management alarm signal information obtained by the equipment production management data type recognition model module with an identification code and then transmit it to the data missing judgment unit; The data missing judgment unit is used to output tracker information after receiving the abnormal production management alarm signal and abnormal production management alarm signal information of the equipment production management data type that matches the identification code, search for each of the tracker information, and determine whether this equipment production management data type is located within the supervision area according to the location of the supervision area, and update the tracker information again, and determine whether the equipment production management data type should be alarmed according to the updated tracker information.

9. The data acquisition method of the MES production management system according to claim 8, characterized in that: The tracker information includes: equipment production management data type identification code, equipment production management data type abnormal production management alarm signal, whether the equipment production management data type enters the equipment production management data type abnormal production management alarm signal, the time when the equipment production management data type enters the equipment production management data type abnormal production management alarm signal, and whether the equipment production management data type has been issued an alarm.

10. The data acquisition method of the MES production management system according to claim 1, characterized in that: The device uses the mean value data set to increase the amount of management data in a real-time updating manner.

Images