A data graph management system and method based on digitalized industrial chain

Abstract

The present invention discloses a data map management system and method based on a digital and intelligent industrial chain, and relates to the technical field of data map information management. The data map management system based on the digital and intelligent industrial chain includes: a data map generation module, a feature extraction module, a data analysis module and a map management module. The present invention generates a data map by acquiring image data of a node in the digital and intelligent industrial chain and inputting it into a map structure, then acquiring frequency feature information and evaluating it to obtain key feature evaluation values, then performing state analysis on the image data and forming a data state curve based on the standard deviation and average value of the image data, and finally comparing the actual value of the image data with the fluctuation range of the image data to determine whether to execute a data map update, thereby achieving an improvement in the accuracy of image data state analysis in the data map management process, and solving the problem of low accuracy of image data state analysis in the data map management process in the prior art.

Description

Technical Field

[0001] The present invention relates to the field of data graph information management technology, and in particular to a data graph management system and method based on a digital industry chain. Background Art

[0002] With the development of technologies such as artificial intelligence, 5G, and industrial Internet, all industries are accelerating their transformation to digitalization and intelligence. This transformation not only improves production efficiency, but also optimizes resource allocation and promotes the overall improvement of the industrial chain. In this process, data has become a key production factor, and how to efficiently manage and utilize this data has become an important challenge facing enterprises and industries. Therefore, a data graph management system based on the digital and intelligent industrial chain has emerged. It can help enterprises and industries better manage and utilize data and promote the in-depth development of digital and intelligent transformation.

[0003] Existing technologies collect various types of data in the digitized industrial chain and process the data to obtain data maps, then conduct distributed analysis on the data in the data maps to identify distributed trends in the industrial chain, and finally visualize the results of the data analysis in the form of charts, thereby enabling enterprises to manage the operation process of the industrial chain.

[0004] For example, the invention patent announcement with announcement number: CN116187868B discloses a method and device for evaluating the quality of industrial chain development based on a knowledge graph, including: obtaining the association between two entities in the industrial chain development entity list and weighting them to obtain weighted associations between entities; constructing an industrial chain knowledge graph based on the industrial chain development entity list and the weighted associations between entities; constructing an industrial chain key element importance evaluation model through a random walk strategy based on the industrial chain knowledge graph to evaluate the importance of key elements in the industrial chain knowledge graph; and obtaining a development quality score for the industrial chain to be evaluated based on the coverage of key elements in the industrial chain to be evaluated.

[0005] For example, the patent application with publication number: CN117952423A discloses a digital-based industrial chain risk management system and method, including: collecting data in the industrial chain field; building an industrial chain knowledge graph; identifying risks in the industrial chain; issuing early warnings to high-risk enterprises; providing targeted risk response plans; and generating risk management reports.

[0006] However, in the process of implementing the technical solution of the invention in the embodiments of the present application, the present application found that the above technology has at least the following technical problems:

[0007] In the existing technology, data in the industrial chain often exists in various forms such as images and texts. The analysis of these data usually involves a complex pattern recognition process, which is easily affected by factors such as noise and light, resulting in inaccurate data analysis results. There is a problem of low accuracy in image data status analysis during data map management. Summary of the invention

[0008] The embodiments of the present application solve the problem of low accuracy of image data status analysis in the data graph management process in the prior art by providing a data graph management system and method based on the digital industry chain, thereby improving the accuracy of image data status analysis in the data graph management process.

[0009] The embodiment of the present application provides a data graph management system based on a digital and intelligent industrial chain, including: a data graph generation module, a feature extraction module, a data analysis module and a graph management module; wherein the data graph generation module is used to obtain image data of a node in the digital and intelligent industrial chain within a preset time period and encode it, and input the encoded image data into a graph structure to generate a data graph, the nodes in the digital and intelligent industrial chain include raw material suppliers, manufacturers, distributors and retailers, the image data include production line flow rate data, production line output data and inventory item quantity data, and the data graph is used to visualize the connection and interaction relationship between the nodes in the digital and intelligent industrial chain; the feature extraction module is used to perform feature extraction on the generated data graph to obtain frequency feature information, and at the same time evaluate the acquired frequency feature information. The key feature evaluation value is obtained by estimating the frequency feature information, wherein the frequency feature information includes high-frequency feature information and low-frequency feature information, and the key feature evaluation value is used to reflect the importance of the frequency feature information in the data map; the data analysis module is used to perform state analysis on the image data in the data map according to the obtained key feature evaluation value, and at the same time form a data state curve according to the standard deviation and average value of the image data in the state analysis process, and the data state curve is used to visualize the fluctuation of the image data within a preset time period; the map management module is used to obtain the actual value of the image data in the data state curve, and compare the actual value of the image data with the fluctuation range of the image data to determine whether to execute the data map update and perform authority visualization, and the actual value of the image data is used to reflect the operation status of the digital industry chain node within a preset time period.

[0010] Furthermore, the specific acquisition process of the key feature evaluation value is as follows: A1, obtain the feature map and perform downsampling to obtain the downsampling rate, and monitor the change of the frequency feature information in real time, and judge whether the frequency feature information is high-frequency feature information. If so, directly execute A2, otherwise record it as low-frequency feature information and add a padding parameter to execute A2, wherein the padding parameter makes the output size of the feature map in the fully connected layer the same as the input size, and the downsampling rate is the ratio of the receiving dimension of the feature map to the initial dimension;

[0011] A2, input the downsampled feature map into the fully connected layer in the convolutional neural network for feature combination and classification, and obtain the resolution and step parameters of the data map through the convolutional neural network input layer interface, and obtain the feature combination weight factor through the convolutional neural network fully connected layer interface. The step parameter is used to reflect the distance that the sliding convolution kernel in the convolutional neural network slides on the data map, and the feature combination weight factor is used to quantify the frequency of interaction between nodes in the feature map; A3, combine the padding parameter and the obtained resolution, step parameter, feature combination weight factor, and downsampling rate to obtain the key feature evaluation value.

[0012] Furthermore, the specific limiting expression of the key feature evaluation value is:

[0013] ;

[0014] In the formula, j is the number of the data map, , J is the total number of data maps, e is a natural constant, represents the key feature evaluation value of the feature graph corresponding to the jth data graph, represents the receiving dimension of the feature map corresponding to the jth data map, represents the initial dimension of the feature map corresponding to the jth data map, represents the resolution of the feature map corresponding to the jth data map, Represents the step size parameter of the feature map corresponding to the jth data map, Represents the feature combination weight factor of the feature graph corresponding to the jth data graph, represents the maximum allowable filling parameter of the feature map corresponding to the jth data map, Indicates the filling parameters of the feature map corresponding to the j-th data map, Indicates the type of frequency feature information corresponding to the jth data spectrum, Represents high-frequency feature information, Represents low-frequency feature information.

[0015] Furthermore, the specific steps of performing state analysis on the image data in the data map based on the acquired key feature evaluation value include: E1, determining whether the acquired key feature evaluation value is within the feature evaluation threshold range, and if so, continuing to monitor the changes of the image data in the data map within a preset time period, otherwise executing E2; E2, recording the image data that is not within the feature evaluation threshold range as an abnormal state, and adjusting the parameters of the feature map in combination with the image data fluctuation range, and the parameter adjustment includes size parameter adjustment, step parameter adjustment and filling parameter adjustment.

[0016] Furthermore, the atlas management module includes an image data actual value acquisition unit and a deviation assessment value acquisition unit; the image data actual value acquisition unit is used to read the production line flow rate data, production line output data and inventory item quantity data on the image data status curve in real time within a preset time period; the deviation assessment value acquisition unit is used to obtain a deviation assessment value in combination with the production line flow rate fluctuation range, production line output fluctuation range and inventory item quantity fluctuation range on the image data status curve, and the deviation assessment value is used to reflect the degree of deviation between the image data actual value and the image data fluctuation range.

[0017] Furthermore, the specific process of comparing the actual value of the image data with the fluctuation range of the image data to determine whether to execute the data map update is as follows: L1, determine whether the production line flow rate data is within the production line flow rate fluctuation range, if so, continue to monitor the changes of the production line flow rate data on the image data status curve, otherwise execute L2; L2, determine whether the production line output data is within the production line output fluctuation range, if so, continue to monitor the changes of the production line output data on the image data status curve, otherwise execute L3; L3, determine whether the inventory item quantity data is within the inventory item quantity fluctuation range, if so, continue to monitor the changes of the inventory item quantity data on the image data status curve, otherwise execute L4; L4, update the preprocessing parameters of the data map according to the abnormal proportion duration of the actual value of the image data and regenerate the image data status curve, the preprocessing parameters are used to suppress the abnormal duration of the actual value of the image data.

[0018] Furthermore, the preprocessing parameters of the data map are updated according to the abnormal proportion duration of the actual value of the image data, and then it also includes obtaining the map management efficiency score; the map management efficiency score is used to reflect the data map management efficiency of the digital industry chain within a preset time period; the map management efficiency score is obtained by the following method: according to the duration of the abnormality and the total duration of the preset time period, the first proportion duration coefficient, the second proportion duration coefficient and the third proportion duration coefficient are obtained and the total proportion duration coefficient is obtained by addition operation, the first proportion duration coefficient represents the ratio of the first proportion duration to the total duration of the preset time period and the first proportion duration weight. The product of the weight factor, the second proportion time coefficient represents the ratio of the second proportion time to the total time of the preset time period and the second proportion time weight factor, the third proportion time coefficient represents the ratio of the third proportion time to the total time of the preset time period and the third proportion time weight factor; the number of parameter updates of the preprocessing parameters of the data map within the preset time period is obtained and the parameter update coefficient is obtained in combination with the corresponding reference value, and the map management efficiency score is obtained according to the result of the addition operation of the total proportion time coefficient, the parameter update coefficient and the maximum value of the deviation evaluation, and the parameter update coefficient is the ratio of the parameter update number to the reference parameter update number.

[0019] The embodiment of the present application provides a data graph management method based on a digital and intelligent industrial chain, including the following steps: S1, obtaining image data of a node in the digital and intelligent industrial chain within a preset time period and encoding it, and inputting the encoded image data into a graph structure to generate a data graph, wherein the nodes in the digital and intelligent industrial chain include raw material suppliers, manufacturers, distributors and retailers, and the image data include production line flow rate data, production line output data and inventory item quantity data, and the data graph is used to visualize the connection and interaction relationship between the nodes in the digital and intelligent industrial chain; S2, performing feature extraction on the generated data graph to obtain frequency feature information, and evaluating the acquired frequency feature information to obtain key feature evaluation values, wherein the frequency The rate feature information includes high-frequency feature information and low-frequency feature information, and the key feature evaluation value is used to reflect the importance of the frequency feature information in the data map; S3, perform state analysis on the image data in the data map according to the acquired key feature evaluation value, and form a data state curve according to the standard deviation and average value of the image data in the state analysis process, and the data state curve is used to visualize the fluctuation of the image data within a preset time period; S4, obtain the actual value of the image data in the data state curve, compare the actual value of the image data with the image data fluctuation range to determine whether to execute the data map update and perform authority visualization, and the actual value of the image data is used to reflect the operation status of the digital industry chain node within a preset time period.

[0020] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0021] 1. A data map is generated by acquiring image data of nodes in the digital industry chain and inputting it into a map structure. Frequency feature information is then acquired and evaluated to obtain key feature evaluation values. The image data in the data map is then analyzed in status and a data status curve is formed based on the standard deviation and average value of the image data. Finally, the actual value of the image data is compared with the fluctuation range of the image data to determine whether to update the data map. This improves the accuracy and reliability of data status curve generation, and further improves the accuracy of image data status analysis in the data map management process. This effectively solves the problem of low accuracy of image data status analysis in the data map management process in the prior art.

[0022] 2. By judging whether the key feature evaluation value obtained is within the feature evaluation threshold range, if so, continue to monitor the changes of image data in the data map within the preset time period, otherwise the image data that is not within the feature evaluation threshold range will be recorded as an abnormal state, and the parameters of the feature map will be adjusted in combination with the image data fluctuation range, thereby achieving more accurate judgment of the image data and improving the reliability of image data status analysis.

[0023] 3. The first proportion duration coefficient, the second proportion duration coefficient and the third proportion duration coefficient are obtained through the duration of the abnormality and the total duration of the preset time period, and the total proportion duration coefficient is obtained by adding them together. Then, the number of parameter updates of the preprocessing parameters of the data map within the preset time period is obtained and the parameter update coefficient is obtained in combination with the corresponding reference value. The map management efficiency score is obtained according to the result of the addition operation of the total proportion duration coefficient, the parameter update coefficient and the maximum value of the deviation assessment, thereby achieving more accurate acquisition of the proportion duration coefficient, and further improving the accuracy of obtaining the map management efficiency score. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of the structure of a data graph management system based on a digitalized industrial chain provided in an embodiment of the present application;

[0025] Figure 2 An analysis flow chart for updating the data map provided in the embodiment of the present application;

[0026] Figure 3 A flowchart of a data graph management method based on a digitalized industrial chain provided in an embodiment of the present application;

[0027] Figure 4 This is a flow chart for analyzing the data graph provided in the embodiments of the present application. DETAILED DESCRIPTION

[0028] The embodiment of the present application solves the problem of low accuracy of image data status analysis in the data map management process in the prior art by providing a data map management system and method based on the digital industry chain. The image data of the digital industry chain node within a preset time period is obtained and encoded through a data map generation module, and the encoded image data is input into the map structure to generate a data map. The feature extraction module then performs feature extraction on the generated data map to obtain frequency feature information, and the obtained frequency feature information is evaluated to obtain key feature evaluation values. The data analysis module then performs status analysis on the image data in the data map according to the obtained key feature evaluation values, and a data status curve is formed according to the standard deviation and average value of the image data in the status analysis process. Finally, the map management module obtains the actual value of the image data in the data status curve, and compares the actual value of the image data with the image data fluctuation range to determine whether to execute the data map update and perform permission visualization, thereby improving the accuracy of image data status analysis in the data map management process.

[0029] The technical solution in the embodiment of the present application is to solve the problem of low accuracy of image data state analysis in the above-mentioned data atlas management process. The overall idea is as follows:

[0030] By acquiring and encoding the image data of the nodes in the digital industry chain, the encoded image data is input into the graph structure to generate a data graph, and then the frequency feature information is acquired and evaluated to obtain the key feature evaluation value. Next, the image data in the data graph is subjected to state analysis and a data state curve is formed according to the standard deviation and average value of the image data. Finally, the actual value of the image data is compared with the fluctuation range of the image data to determine whether to execute the data graph update, thereby achieving the effect of improving the accuracy of image data state analysis in the data graph management process.

[0031] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0032] like Figure 1As shown, it is a structural schematic diagram of a data map management system based on a digital and intelligent industrial chain provided in an embodiment of the present application. A data map management system based on a digital and intelligent industrial chain provided in an embodiment of the present application includes: a data map generation module, a feature extraction module, a data analysis module and a map management module; wherein the data map generation module is used to obtain and encode the image data of the digital and intelligent industrial chain nodes within a preset time period, and input the encoded image data into the map structure to generate a data map. The digital and intelligent industrial chain nodes include raw material suppliers, manufacturers, distributors and retailers. The image data includes production line flow rate data, production line output data and inventory item quantity data. The map structure is the basic framework of the data map to be managed. The data map is used to visualize the connection and interaction relationship between the nodes in the digital and intelligent industrial chain; the feature extraction module is used to perform feature extraction on the generated data map to obtain frequency feature information, and at the same time evaluate the acquired frequency feature information to obtain key feature evaluation values. The frequency feature information includes high-frequency features The high-frequency feature information is used to reflect the texture features between entity nodes in the data map. The low-frequency feature information includes the edge features, color features and brightness features of the data map. The key feature evaluation value is used to reflect the importance of the frequency feature information in the data map. The data analysis module is used to perform state analysis on the image data in the data map according to the acquired key feature evaluation values, and at the same time form a data state curve according to the standard deviation and average value of the image data in the state analysis process. The state analysis is used to reflect the state changes of the image data within a preset time period, and the data state curve is used to visualize the fluctuation of the image data within a preset time period. The map management module is used to obtain the actual value of the image data in the data state curve, compare the actual value of the image data with the fluctuation range of the image data to determine whether to execute the data map update and perform permission visualization. The actual value of the image data is used to reflect the operation status of the digital industry chain node within the preset time period, and the permission visualization represents the authorized user's access and query to the visualized data map.

[0033] In this embodiment, the production line flow rate data (i.e., the transmission speed of the products on the production line) is obtained through sensors deployed at preset monitoring points (set by preset personnel) on the production line of the digital industry chain. For example, there are ten monitoring points on a production line, and the distances between adjacent monitoring points are the same; the production line output data (i.e., the number of products delivered by the production line within a preset time period) is obtained through sensors deployed at the assembly line interface on the production line. It should be noted that the duration of each preset time period is the same, such as hourly output and daily output; the inventory item quantity data (i.e., the number of products in the digital industry chain inventory within a preset time period) is monitored in real time through sensors on the inventory management system interface of the digital industry chain.

[0034] Specifically, in the digital industry chain, high-frequency feature information only includes texture features between entity nodes in the data graph. Edge features, color features, and brightness features usually contained in the data graph are low-frequency feature information. Each node can be regarded as an entity in the data graph. Each node has a unique identifier and attributes to describe the characteristics and status of the node. The relationship between nodes (such as raw material supply relationships, production relationships, and distribution relationships) can be represented as edges in the data graph. Use visualization tools (such as Graphviz) to display the data graph in a graphical way to reflect the changes of each node in the industrial chain in real time, thereby improving the accuracy of image data status analysis in the data graph management process.

[0035] Furthermore, the specific process of obtaining the key feature evaluation value is as follows: A1, obtain the feature map and downsample it to obtain the downsampling rate, and monitor the changes in the frequency feature information in real time to determine whether the frequency feature information is high-frequency feature information. If so, do not add the padding parameter and directly execute A2, otherwise record it as low-frequency feature information and add the padding parameter before executing A2. The feature map is obtained by transmitting the generated data spectrum to the input layer in the convolutional neural network and inputting it into the convolution layer for convolution operation. The convolution operation means extracting the feature map of the frequency components in the spectrum data through the sliding convolution kernel in the convolution layer. The sliding convolution kernel contains a size parameter, a step parameter and a padding parameter. The padding parameter makes the output size of the feature map in the fully connected layer the same as the input size. Downsampling is used to reduce the dimension of the feature map. The downsampling rate is the ratio of the receiving dimension of the feature map to the initial dimension. , the receiving dimension is obtained through the sensor deployed by the fully connected layer interface of the convolutional neural network, and the initial dimension is obtained through the sensor deployed by the pooling layer interface of the convolutional neural network; A2, the downsampled feature map is input into the fully connected layer in the convolutional neural network for feature combination and classification, and the resolution and step parameters of the data map are obtained through the convolutional neural network input layer interface, and the feature combination weight factor is obtained through the convolutional neural network fully connected layer interface. The fully connected layer is used to map the frequency components in the feature map to the frequency characteristics of the fully connected layer. The step parameter is used to reflect the distance that the sliding convolution kernel in the convolutional neural network slides on the data map. The feature combination weight factor is used to quantify the frequency of interaction between nodes in the feature map; A3, the key feature evaluation value is obtained by combining the padding parameter and the obtained resolution, step parameter, feature combination weight factor, and downsampling rate.

[0036] In this embodiment, the dimension generally refers to the amount of graph data processed in a feature graph at one time; when the output size of the feature graph is different from the input size, it is necessary to increase the padding parameter to keep the size consistent. It should be noted that since high-frequency feature information is insensitive to size changes, when the frequency feature information is high-frequency feature information, there is no need to consider increasing or decreasing the padding parameter; the feature combination weight factor ranges from 0 to 1. Within a preset time period (1 hour), it is assumed that: the number of interactions between node 1 and node 2 is 10 times, and the number of interactions between node 2 and node 3 is 5 times, then the frequency of interaction between node 1 and node 2 is greater than the frequency of interaction between node 2 and node 3, thereby achieving efficient processing and analysis of the data graph, and further achieving more accurate acquisition of key feature evaluation values.

[0037] Furthermore, the specific limiting expression of the key feature evaluation value is:

[0038] ;

[0039] In the formula, j is the number of the data map, , J is the total number of data maps, e is a natural constant, represents the key feature evaluation value of the feature graph corresponding to the jth data graph, represents the receiving dimension of the feature map corresponding to the jth data map, represents the initial dimension of the feature map corresponding to the jth data map, represents the resolution of the feature map corresponding to the jth data map, Represents the step size parameter of the feature map corresponding to the jth data map, Represents the feature combination weight factor of the feature graph corresponding to the jth data graph, represents the maximum allowable filling parameter of the feature map corresponding to the jth data map, Indicates the filling parameters of the feature map corresponding to the j-th data map, Indicates the type of frequency feature information corresponding to the jth data spectrum, Represents high-frequency feature information, Represents low-frequency feature information.

[0040] In this embodiment, the maximum allowable filling parameter is represented by the sum and average of the maximum values ​​of the historical filling parameters of the convolutional neural network in the preset database. When , the key feature evaluation value decreases with the increase of the downsampling rate, the resolution coefficient (the product of the complement of the resolution and the inverse of the step size parameter) and the feature combination weight factor; when When , the key feature evaluation value decreases with the increase of downsampling rate, resolution coefficient, feature combination weight factor and filling parameter (not 0). The change statistics of key feature evaluation value are shown in Table 1:

[0041] Table 1 Statistics of changes in key feature evaluation values

[0042]

[0043] It is important to understand that when When the padding parameter is increased (decreased), the corresponding resolution will decrease (increase), while the step parameter will increase (decrease), thereby affecting the value of the key feature evaluation value. By considering the indirect effect of the padding parameter on the resolution and step parameter, the accuracy of frequency feature extraction can be improved, thereby improving the accuracy of data map management, thereby achieving an improvement in the accuracy of image data state analysis in the process of data map management, and effectively solving the problem of low accuracy of image data state analysis in the process of data map management in the prior art.

[0044] Furthermore, the specific steps of performing state analysis on the image data in the data map based on the acquired key feature evaluation value include: E1, determining whether the acquired key feature evaluation value is within the feature evaluation threshold range, if so, recording it as a normal state and continuing to monitor the changes of the image data in the data map within a preset time period, otherwise executing E2; E2, recording the image data that is not within the feature evaluation threshold range as an abnormal state, and adjusting the parameters of the feature map in combination with the image data fluctuation range, and the parameter adjustment includes size parameter adjustment, step parameter adjustment and filling parameter adjustment.

[0045] In this embodiment, the feature evaluation threshold range represents the range corresponding to the maximum and minimum values ​​of the historical key feature evaluation values ​​in the historical time period in the preset database. Assume that the initial values ​​of the height and width of the feature map are 64*64, the initial value of the step parameter is 1, and the initial value of the padding parameter is 0.8. When the key feature evaluation value is not within the feature evaluation threshold range, it is necessary to adjust the height and width of the feature map from the original 64*64 to 128*128 to increase the fine-grainedness of feature extraction, adjust the step parameter of the sliding convolution kernel from the original 1 to 0.8 to reduce the information loss during feature extraction, and adjust the padding parameter from the original 0.8 to 1 to reduce the impact of the boundary effect of the feature map on feature extraction. Dynamic parameter adjustment through the feature map helps the key feature evaluation value to be closer to the set feature evaluation threshold range, thereby improving the real-time and reliability of the image data state analysis on the feature map.

[0046] Furthermore, the specific steps of forming a data state curve according to the standard deviation and average value of image data in the state analysis process are as follows: obtaining pixel values ​​of the image data within a preset time period and averaging them to obtain a pixel average value, and recording the square root of the square operation of the deviation between the pixel average value and the pixel value as the pixel standard deviation; generating an image data state curve according to the obtained pixel average value and pixel standard deviation, the image data state curve including a first image data state curve and a second image data state curve; the first image data state curve is a graph showing that the pixel average value changes with changes at various time points within a preset time period; the second image data state curve is a graph showing that the pixel standard deviation changes with changes at various time points within the preset time period.

[0047] In this embodiment, the actual values ​​of the image data on the first image data status curve (the horizontal axis is the preset time points within the preset time period, and the vertical axis is the pixel average value) and the second image data status curve (the horizontal axis is the preset time points within the preset time period, and the vertical axis is the pixel standard deviation) are marked with different colors, and the pixel average values ​​and pixel standard deviations at the same time points on the first image data status curve and the second image data status curve are marked with the same color. It should be noted that the time intervals between the preset time points are the same, such as there are 6 preset time points in 1 hour (10 minutes, 20 minutes, 30 minutes, 40 minutes, 50 minutes, and 60 minutes).

[0048] This embodiment can intuitively display the changing trends of the mean and standard deviation of image data over time by generating a first image data status curve and a second image data status curve. The actual values ​​of the image data marked with different colors help to quickly identify abnormal data points. By considering the pixel mean and pixel standard deviation, the accuracy and representativeness of the image data status curve are improved. The pixel mean and pixel standard deviation at the same time point are marked with the same color, which is convenient for comparing and analyzing the relationship between the two, helping to ensure the real-time and timeliness of the atlas data and improving the efficiency of data atlas management.

[0049] Furthermore, the atlas management module includes an image data actual value acquisition unit and a deviation assessment value acquisition unit; the image data actual value acquisition unit is used to read the production line flow rate data, production line output data and inventory item quantity data on the image data status curve in real time within a preset time period; the deviation assessment value acquisition unit is used to obtain a deviation assessment value in combination with the production line flow rate fluctuation range, production line output fluctuation range and inventory item quantity fluctuation range on the image data status curve, and the deviation assessment value is used to reflect the degree of deviation between the image data actual value and the image data fluctuation range.

[0050] Among them, the specific process of obtaining the deviation evaluation value is: obtain the first deviation coefficient, the second deviation coefficient and the third deviation coefficient and perform addition operation to obtain the deviation evaluation value; the first deviation coefficient represents the ratio of the absolute value of the difference between the production line flow rate data and the reference production line flow rate data to the reference production line flow rate data; the second deviation coefficient represents the ratio of the absolute value of the difference between the production line output data and the reference production line output data to the reference production line output data; the third deviation coefficient represents the ratio of the absolute value of the difference between the inventory item quantity data and the reference inventory item quantity data to the reference inventory item quantity data.

[0051] In this embodiment, the production line flow rate fluctuation range represents the range corresponding to the maximum and minimum values ​​of the historical production line flow rate within the historical time period in the preset database, the production line output fluctuation range represents the range corresponding to the maximum and minimum values ​​of the historical production line output within the historical time period in the preset database, and the inventory item quantity fluctuation range represents the range corresponding to the maximum and minimum values ​​of the historical inventory item quantity within the historical time period in the preset database; the reference production line flow rate data is represented by the sum and average of the historical production line flow rates within the historical time period in the preset database, the reference production line output data is represented by the sum and average of the historical production line outputs within the historical time period in the preset database, and the reference inventory item quantity data is represented by the sum and average of the historical inventory item quantities within the historical time period in the preset database.

[0052] Specifically, when the deviation evaluation value is within the deviation threshold range, it indicates that the degree of deviation between the actual value of the image data and the image data fluctuation range is within the allowable deviation range (set by the preset personnel); otherwise, it indicates that the degree of deviation between the actual value of the image data and the image data fluctuation range is not within the allowable deviation range. At this time, it is necessary to adjust the amount of image data within the preset time period. For example, when the production line output data is not within the production line output fluctuation range, it is necessary to re-acquire the production line output data within the preset time period and reduce the amount of production line output data. When the re-acquired production line output data is within the production line output fluctuation range, the adjustment is stopped. The deviation threshold range represents the range corresponding to the maximum and minimum values ​​of the historical deviation evaluation values ​​within the historical time period in the preset database.

[0053] Specifically, the specific limiting expression of the deviation evaluation value is:

[0054] ;

[0055] In the formula, i is the number of the preset time point, , U is the total number of preset time points, e is a natural constant, represents the deviation evaluation value at the i-th preset time point on the image data state curve, represents the production line flow rate data at the i-th preset time point on the image data state curve, represents the reference production line flow rate data, Represents the production line output data at the i-th preset time point on the image data state curve, represents the reference production line output data, Represents the inventory item quantity data at the i-th preset time point on the image data state curve, Represents reference inventory item quantity data.

[0056] Specifically, when , and When , the deviation evaluation value increases with the increase of production line flow rate data, production line output data and inventory item quantity data. , and When , the corresponding deviation evaluation value is 0; when , and When the deviation evaluation value increases, the deviation evaluation value decreases as the production line flow rate data, production line output data and inventory item quantity data increase.

[0057] To simplify the analysis, we define , , , where represents the first deviation coefficient of the i-th preset time point on the image data state curve, represents the second deviation coefficient of the i-th preset time point on the image data state curve, The third deviation coefficient of the i-th preset time point on the image data state curve is represented by the simplified calculation formula of the deviation evaluation value: , the statistical table of the change of deviation evaluation value is shown in Table 2:

[0058] Table 2 Statistics of changes in deviation evaluation values

[0059]

[0060] It should be understood that the deviation evaluation value increases with the increase of the first deviation coefficient, the second deviation coefficient and the second deviation coefficient. , and When the first deviation coefficient increases with the increase of production line flow rate data, the second deviation coefficient increases with the increase of production line output data, and the third deviation coefficient increases with the increase of inventory item quantity data. , and When the first deviation coefficient decreases with the increase of the production line flow rate data, the second deviation coefficient decreases with the increase of the production line output data, and the third deviation coefficient decreases with the increase of the inventory item quantity data. This embodiment introduces the concept of deviation evaluation value, which helps to more clearly reflect the degree of deviation between the actual value of the image data and the fluctuation range of the image data, thereby improving the accuracy of data map management.

[0061] It should be noted that the key feature evaluation value can indirectly affect the deviation evaluation value by affecting the first deviation coefficient, the second deviation coefficient and the third deviation coefficient in the deviation evaluation value. For example, if the production line flow rate data increases, the first deviation coefficient also increases accordingly, thereby increasing the deviation evaluation value. Similarly, if the production line output data or the inventory item quantity data increases, the second deviation coefficient or the third deviation coefficient also increases accordingly, thereby increasing the deviation evaluation value. By considering the correlation between the key feature evaluation value and the deviation evaluation value, the degree of deviation between the actual value of the image data in the data map and the image data fluctuation range can be more accurately reflected, thereby achieving an improvement in the accuracy of image data state analysis in the data map management process, and effectively solving the problem of low accuracy of image data state analysis in the data map management process in the prior art.

[0062] Further, such as Figure 2 As shown, it is an analysis flow chart of the data map update provided by the embodiment of the present application. The specific process of comparing the actual value of the image data with the fluctuation range of the image data to determine whether to execute the data map update is as follows: L1, determine whether the production line flow rate data is within the production line flow rate fluctuation range. If so, continue to monitor the changes of the production line flow rate data on the image data state curve, otherwise execute L2; L2, determine whether the production line output data is within the production line output fluctuation range. If so, continue to monitor the changes of the production line output data on the image data state curve, otherwise execute L3; L3, determine whether the inventory item quantity data is within the inventory item quantity fluctuation range. If so, continue to monitor the inventory items. The change of quantity data on the image data status curve, otherwise execute L4; L4, update the preprocessing parameters of the data map according to the abnormal proportion duration of the actual value of the image data and regenerate the image data status curve, the abnormal duration includes the first proportion duration, the second proportion duration and the third proportion duration, the first proportion duration is the duration during which the production line flow rate data is not within the production line flow rate fluctuation range within the preset time period, the second proportion duration is the duration during which the production line output data is not within the production line output fluctuation range within the preset time period, the third proportion duration is the duration during which the inventory item quantity data is not within the inventory item quantity fluctuation range within the preset time period, the preprocessing parameters are used to suppress the abnormal duration of the actual value of the image data.

[0063] In this embodiment, in the actual application of the digital industry chain, the preprocessing parameters usually refer to smoothing parameters (usually set by preset personnel). When an abnormal proportion duration appears on the image data status curve (i.e., any one of the first proportion duration, the second proportion duration, and the third proportion duration), the smoothing parameters on the feature graph are increased (such as increasing the initial value of the smoothing parameter on the feature graph from 0.6 to 0.8) to reduce the noise and fluctuation amplitude of the actual value of the image data, thereby achieving more accurate suppression of the duration of the abnormality of the actual value of the image data.

[0064] Furthermore, the preprocessing parameters of the data map are updated according to the abnormal proportion duration of the actual value of the image data, and then the map management efficiency score is obtained; the map management efficiency score is used to reflect the data map management efficiency of the digital industry chain within a preset time period; the map management efficiency score is obtained by the following method: according to the duration of the abnormality and the total duration of the preset time period, the first proportion duration coefficient, the second proportion duration coefficient and the third proportion duration coefficient are obtained and added to obtain the total proportion duration coefficient, the first proportion duration coefficient represents the ratio of the first proportion duration to the total duration of the preset time period and the first proportion duration weight The product of the factors, the second proportion time coefficient represents the ratio of the second proportion time to the total time of the preset time period and the second proportion time weight factor, the third proportion time coefficient represents the ratio of the third proportion time to the total time of the preset time period and the third proportion time weight factor; the number of parameter updates of the preprocessing parameters of the data map within the preset time period is obtained and the parameter update coefficient is obtained in combination with the corresponding reference value, the map management efficiency score is obtained according to the result of the addition operation of the total proportion time coefficient, the parameter update coefficient and the maximum value of the deviation evaluation, and the parameter update coefficient is the ratio of the parameter update number to the reference parameter update number.

[0065] In this embodiment, the first proportion duration weight factor is the weight factor corresponding to the first proportion duration coefficient in the preset database, which represents the numerical value of the influence of the first proportion duration coefficient on the graph management efficiency score. When used, the weight factor corresponding to the first proportion duration coefficient can be directly obtained from the preset database, and the corresponding relationship can be a pre-set mapping relationship. For example, the first proportion duration coefficient corresponding to the proportion of the first proportion duration in the preset time period and the weight factor corresponding to the first proportion duration coefficient in the preset database form a mapping set, and the real-time first proportion duration coefficient is input into the mapping set to obtain the corresponding weight factor, wherein the mapping relationship can be one-to-one or many-to-one. The value range of the weight factor of the first proportion duration coefficient in this example is .

[0066] The second proportion duration weight factor is the weight factor corresponding to the second proportion duration coefficient in the preset database, which represents the numerical value of the influence of the second proportion duration coefficient on the graph management efficiency score. When used, the weight factor corresponding to the second proportion duration coefficient can be directly obtained from the preset database. The corresponding relationship can be a pre-set mapping relationship. For example, the second proportion duration coefficient corresponding to the proportion of the second proportion duration in the preset time period and the weight factor corresponding to the second proportion duration coefficient in the preset database form a mapping set, and the real-time second proportion duration coefficient is input into the mapping set to obtain the corresponding weight factor. The mapping relationship can be one-to-one or many-to-one. The value range of the second proportion duration coefficient weight factor in this example is .

[0067] The sum of the first proportion duration coefficient weight factor, the second proportion duration coefficient weight factor and the third proportion duration coefficient weight factor is 1.

[0068] Specifically, the specific restricted expression of the graph management efficiency score is:

[0069] ;

[0070] ;

[0071] In the formula, i is the number of the preset time point, , U is the total number of preset time points, e is a natural constant, Indicates the image data management efficiency score of the image data on the image data status curve within a preset time period. Represents the first proportion duration weight factor, Indicates the first proportion duration of the production line flow rate data on the image data status curve within a preset time period, Indicates the total duration of the preset time period. Represents the second proportion duration weight factor, Indicates the second proportion duration of the production line output data on the image data status curve within the preset time period, Represents the weight factor of the third proportion duration, Indicates the third proportion duration of the inventory item quantity data on the image data status curve within the preset time period, Indicates the number of parameter updates of the preprocessing parameters of the data map within the preset time period. Indicates the number of reference parameter updates, Indicates the maximum deviation evaluation value of the deviation evaluation value on the image data status curve within a preset time period, Represents the deviation evaluation value of the image data on the image data state curve at the Uth preset time point.

[0072] Specifically, when the obtained graph management efficiency score is within the graph management threshold range, it indicates that the data graph management efficiency of the digital industry chain within the preset time period meets the preset data graph management requirements (set by the preset personnel); otherwise, it indicates that the data graph management efficiency of the digital industry chain within the preset time period does not meet the preset data graph management requirements. At this time, the data graph management process needs to be adjusted. For example, increase the number of data graph backups in the data graph management process and reduce the amount of image data within the preset time period. After each data graph management process adjustment, the graph management efficiency score is re-obtained until the obtained graph management efficiency score is within the graph management threshold range, where the graph management threshold range represents the range corresponding to the maximum and minimum values ​​of the historical graph management efficiency within the historical time period in the preset database.

[0073] Specifically, the number of reference parameter updates is represented by the sum and average of the number of historical parameter updates in the historical time period in the preset database; the graph management efficiency score decreases with the increase of the first proportion time, the second proportion time, the third proportion time, the number of parameter updates and the maximum deviation evaluation value. It should be noted that this embodiment introduces the concept of the maximum deviation evaluation value to make the data graph management efficiency score more comprehensive, which can more truly reflect the actual situation in the data graph management process, and realize the improvement of the management efficiency of the graph management module.

[0074] It should be noted that the maximum deviation evaluation value also indirectly affects the number of parameter updates. Assume that in the data map management of a production line, when the maximum deviation evaluation value of the production line flow rate data increases (this may be due to a failure of a device on the production line), in order to maintain the accuracy and timeliness of the data map, it is necessary to increase the number of parameter updates to adapt to changes on the production line. In this case, the increase in the number of parameter updates can be regarded as a response to the change in the maximum deviation evaluation value. By updating the parameters more frequently, the system can better adapt to changes on the production line, thereby improving the accuracy of the data map, thereby achieving an improvement in the accuracy of the image data state analysis in the data map management process, and effectively solving the problem of low accuracy of the image data state analysis in the data map management process in the prior art.

[0075] like Figure 3As shown, it is a flow chart of a data map management method based on a digital and intelligent industrial chain provided in an embodiment of the present application. A data map management method based on a digital and intelligent industrial chain provided in an embodiment of the present application includes the following steps: S1, generating a data map: obtaining image data of a node in the digital and intelligent industrial chain within a preset time period and encoding it, and inputting the encoded image data into a map structure to generate a data map. The nodes in the digital and intelligent industrial chain include raw material suppliers, manufacturers, distributors and retailers. The image data includes production line flow rate data, production line output data and inventory item quantity data. The map structure is the basic framework of the data map to be managed. The data map is used to visualize the connection and interaction relationship between each node in the digital and intelligent industrial chain; S2, obtaining key feature evaluation values: performing feature extraction on the generated data map to obtain frequency feature information, and evaluating the obtained frequency feature information to obtain key feature evaluation values. The frequency feature information includes high-frequency feature information and low-frequency feature information. The high-frequency feature information is used To reflect the texture features between entity nodes in the data map, the low-frequency feature information includes the edge features, color features and brightness features of the data map, and the key feature evaluation value is used to reflect the importance of the frequency feature information in the data map; S3, generate a data status curve: perform a status analysis on the image data in the data map according to the acquired key feature evaluation value, and form a data status curve according to the standard deviation and average value of the image data during the status analysis process. The status analysis is used to reflect the status changes of the image data within a preset time period, and the data status curve is used to visualize the fluctuation of the image data within a preset time period; S4, data map update judgment: obtain the actual value of the image data in the data status curve, compare the actual value of the image data with the fluctuation range of the image data to determine whether to execute the data map update and perform permission visualization. The actual value of the image data is used to reflect the operating status of the digital industry chain node within a preset time period, and the permission visualization represents the authorized user's access and query to the visualized data map.

[0076] like Figure 4 As shown, this is an analysis flow chart of the data map provided in an embodiment of the present application. This embodiment reflects the state changes of the image data within a preset time period by performing state analysis on the image data, and forms a data state curve in combination with the standard deviation and the average value, which can dynamically display the fluctuations in the operating state of the industrial chain nodes. This dynamic and real-time analysis helps to timely discover trends in the operation process, and achieves improved accuracy and real-time performance of image data state analysis.

[0077] To summarize, the embodiment of the present application generates a data map by acquiring image data of nodes in the digital industry chain and inputting it into a map structure, then acquiring frequency feature information and evaluating it to obtain key feature evaluation values, then performing state analysis on the image data in the data map and forming a data state curve based on the standard deviation and average value of the image data, and finally comparing the actual value of the image data with the image data fluctuation range to determine whether to perform a data map update, thereby achieving improved accuracy and reliability in generating the data state curve, and further achieving improved accuracy in image data state analysis in the process of data map management, effectively solving the problem of low accuracy in image data state analysis in the process of data map management in the prior art.

[0078] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0079] The present invention is described with reference to flowcharts and / or block diagrams of systems, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0080] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0081] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0082] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0083] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A data graph management system based on a digitalized industrial chain, characterized in that: include: Data graph generation module, feature extraction module, data analysis module and graph management module; The data map generation module is used to obtain and encode the image data of the digital industry chain nodes within a preset time period, and input the encoded image data into the map structure to generate a data map. The digital industry chain nodes include raw material suppliers, manufacturers, distributors and retailers. The image data includes production line flow rate data, production line output data and inventory item quantity data. The data map is used to visualize the connection and interaction relationship between the nodes in the digital industry chain. The feature extraction module is used to perform feature extraction on the generated data map to obtain frequency feature information, and at the same time evaluate the obtained frequency feature information to obtain a key feature evaluation value, wherein the frequency feature information includes high-frequency feature information and low-frequency feature information, wherein the high-frequency feature information is used to reflect the texture features between entity nodes in the data map, and the low-frequency feature information includes edge features, color features and brightness features of the data map, and the key feature evaluation value is used to reflect the importance of the frequency feature information in the data map; The data analysis module is used to perform state analysis on the image data in the data map according to the acquired key feature evaluation values, and to form a data state curve according to the standard deviation and average value of the image data during the state analysis process, wherein the data state curve is used to visualize the fluctuation of the image data within a preset time period; The graph management module is used to obtain the actual value of the image data in the data status curve, compare the actual value of the image data with the image data fluctuation range to determine whether to execute the data graph update and perform authority visualization. The actual value of the image data is used to reflect the operating status of the digital industry chain node within a preset time period; The specific process of obtaining the key feature evaluation value is as follows: A1, obtain the feature map and perform downsampling to obtain the downsampling rate, and monitor the change of the frequency feature information in real time, and determine whether the frequency feature information is high-frequency feature information. If so, directly execute A2, otherwise record it as low-frequency feature information and add padding parameters to execute A2. The padding parameters make the output size of the feature map in the fully connected layer the same as the input size, and the downsampling rate is the ratio of the receiving dimension of the feature map to the initial dimension; A2, input the downsampled feature map into the fully connected layer in the convolutional neural network for feature combination and classification, and obtain the resolution and step parameters of the data map through the convolutional neural network input layer interface, and obtain the feature combination weight factor through the convolutional neural network fully connected layer interface. The step parameter is used to reflect the distance that the sliding convolution kernel in the convolutional neural network slides on the data map, and the feature combination weight factor is used to quantify the frequency of interaction between nodes in the feature map; A3, combining the filling parameters and the obtained resolution, step size parameters, feature combination weight factor, and downsampling rate to obtain the key feature evaluation value; The specific limiting expression of the key feature evaluation value is: ； In the formula, j is the number of the data map, , J is the total number of data maps, e is a natural constant, represents the key feature evaluation value of the feature graph corresponding to the jth data graph, represents the receiving dimension of the feature map corresponding to the jth data map, represents the initial dimension of the feature map corresponding to the jth data map, represents the resolution of the feature map corresponding to the jth data map, Represents the step size parameter of the feature map corresponding to the jth data map, Represents the feature combination weight factor of the feature graph corresponding to the jth data graph, represents the maximum allowable filling parameter of the feature map corresponding to the jth data map, Indicates the filling parameters of the feature map corresponding to the j-th data map, Indicates the type of frequency feature information corresponding to the jth data spectrum, Represents high-frequency feature information, Represents low-frequency feature information.

2. A data graph management system based on a digitalized industrial chain as claimed in claim 1, characterized in that: The specific steps of performing state analysis on the image data in the data map according to the acquired key feature evaluation values ​​include: E1, determine whether the obtained key feature evaluation value is within the feature evaluation threshold range, if so, continue to monitor the changes of the image data in the data map within the preset time period, otherwise execute E2; E2, record the image data that is not within the feature evaluation threshold range as an abnormal state, and adjust the parameters of the feature map in combination with the image data fluctuation range, wherein the parameter adjustment includes size parameter adjustment, step parameter adjustment and padding parameter adjustment.

3. A data graph management system based on a digitalized industrial chain as claimed in claim 1, characterized in that: The specific steps of forming the data state curve according to the standard deviation and average value of the image data in the state analysis process are: Obtaining pixel values ​​of the image data within a preset time period and averaging them to obtain a pixel average value, and recording the square root of the deviation between the pixel average value and the pixel value as the pixel standard deviation; Generate an image data state curve according to the acquired pixel average value and pixel standard deviation, wherein the image data state curve includes a first image data state curve and a second image data state curve; The first image data state curve is a graph showing how the pixel average value changes as each time point changes within a preset time period; The second image data state curve is a curve graph showing how the pixel standard deviation changes with each time point within a preset time period.

4. A data graph management system based on a digitalized industrial chain as claimed in claim 1, characterized in that: The atlas management module includes an image data actual value acquisition unit and a deviation evaluation value acquisition unit; The image data actual value acquisition unit is used to read the production line flow rate data, production line output data and inventory item quantity data on the image data state curve in real time within a preset time period; The deviation evaluation value acquisition unit is used to obtain a deviation evaluation value by combining the production line flow rate fluctuation range, the production line output fluctuation range and the inventory item quantity fluctuation range on the image data status curve. The deviation evaluation value is used to reflect the degree of deviation between the actual value of the image data and the image data fluctuation range.

5. A data graph management system based on a digitalized industrial chain as claimed in claim 4, characterized in that: The specific process of obtaining the deviation evaluation value is as follows: Obtaining a first deviation coefficient, a second deviation coefficient, and a third deviation coefficient and performing an addition operation to obtain a deviation evaluation value; The first deviation coefficient represents the ratio of the absolute value of the difference between the production line flow rate data and the reference production line flow rate data to the reference production line flow rate data; The second deviation coefficient represents the ratio of the absolute value of the difference between the production line output data and the reference production line output data to the reference production line output data; The third deviation coefficient represents the ratio of the absolute value of the difference between the inventory item quantity data and the reference inventory item quantity data to the reference inventory item quantity data.

6. A data graph management system based on a digitalized industrial chain as claimed in claim 1, characterized in that: The specific process of comparing the actual value of the image data with the fluctuation range of the image data to determine whether to perform the data map update is as follows: L1, determine whether the production line flow rate data is within the production line flow rate fluctuation range, if so, continue to monitor the change of the production line flow rate data on the image data state curve, otherwise execute L2; L2, determine whether the production line output data is within the production line output fluctuation range, if so, continue to monitor the changes of the production line output data on the image data state curve, otherwise execute L3; L3, determine whether the inventory item quantity data is within the inventory item quantity fluctuation range, if so, continue to monitor the change of the inventory item quantity data on the image data state curve, otherwise execute L4; L4, updating the preprocessing parameters of the data map according to the abnormal proportion duration of the actual value of the image data and regenerating the image data status curve, wherein the preprocessing parameters are used to suppress the abnormal duration of the actual value of the image data.

7. A data graph management system based on a digitalized industrial chain as claimed in claim 6, characterized in that: The preprocessing parameters of the data map are updated according to the abnormal proportion duration of the actual value of the image data, and then the map management efficiency score is obtained; The graph management efficiency score is used to reflect the data graph management efficiency of the digital industry chain within a preset time period; The graph management efficiency score is obtained by the following method: According to the duration of the abnormality and the total duration of the preset time period, a first proportion duration coefficient, a second proportion duration coefficient and a third proportion duration coefficient are obtained and added to obtain a total proportion duration coefficient, wherein the first proportion duration coefficient represents the product of the ratio of the first proportion duration to the total duration of the preset time period and the first proportion duration weight factor, the second proportion duration coefficient represents the product of the ratio of the second proportion duration to the total duration of the preset time period and the second proportion duration weight factor, and the third proportion duration coefficient represents the product of the ratio of the third proportion duration to the total duration of the preset time period and the third proportion duration weight factor; Obtain the number of parameter updates of the preprocessing parameters of the data map within a preset time period and obtain the parameter update coefficient in combination with the corresponding reference value. Obtain the map management efficiency score based on the result of the addition operation of the total coefficient of the proportion duration, the parameter update coefficient and the maximum value of the deviation evaluation. The parameter update coefficient is the ratio of the number of parameter updates to the number of reference parameter updates.

8. A data graph management method based on a digitalized industrial chain, characterized in that: The following steps are involved: S1, obtaining and encoding image data of digital industry chain nodes within a preset time period, and inputting the encoded image data into a graph structure to generate a data graph, wherein the digital industry chain nodes include raw material suppliers, manufacturers, distributors and retailers, and the image data include production line flow rate data, production line output data and inventory item quantity data, and the data graph is used to visualize the connection and interaction relationship between the nodes in the digital industry chain; S2, extracting features from the generated data map to obtain frequency feature information, and evaluating the obtained frequency feature information to obtain a key feature evaluation value, wherein the frequency feature information includes high-frequency feature information and low-frequency feature information, wherein the high-frequency feature information is used to reflect the texture features between entity nodes in the data map, and the low-frequency feature information includes edge features, color features, and brightness features of the data map, and the key feature evaluation value is used to reflect the importance of the frequency feature information in the data map; S3, performing state analysis on the image data in the data map according to the acquired key feature evaluation values, and forming a data state curve according to the standard deviation and average value of the image data during the state analysis process, wherein the data state curve is used to visualize the fluctuation of the image data within a preset time period; S4, obtaining the actual value of the image data in the data status curve, comparing the actual value of the image data with the image data fluctuation range to determine whether to execute the data map update and perform authority visualization, wherein the actual value of the image data is used to reflect the operation status of the digital industry chain node within a preset time period; The specific process of obtaining the key feature evaluation value is as follows: A1, obtain the feature map and perform downsampling to obtain the downsampling rate, and monitor the change of the frequency feature information in real time, and determine whether the frequency feature information is high-frequency feature information. If so, directly execute A2, otherwise record it as low-frequency feature information and add padding parameters to execute A2. The padding parameters make the output size of the feature map in the fully connected layer the same as the input size, and the downsampling rate is the ratio of the receiving dimension of the feature map to the initial dimension; A2, input the downsampled feature map into the fully connected layer in the convolutional neural network for feature combination and classification, and obtain the resolution and step parameters of the data map through the convolutional neural network input layer interface, and obtain the feature combination weight factor through the convolutional neural network fully connected layer interface. The step parameter is used to reflect the distance that the sliding convolution kernel in the convolutional neural network slides on the data map, and the feature combination weight factor is used to quantify the frequency of interaction between nodes in the feature map; A3, combining the filling parameters and the obtained resolution, step size parameters, feature combination weight factor, and downsampling rate to obtain the key feature evaluation value; The specific limiting expression of the key feature evaluation value is: ； In the formula, j is the number of the data map, , J is the total number of data maps, e is a natural constant, represents the key feature evaluation value of the feature graph corresponding to the jth data graph, represents the receiving dimension of the feature map corresponding to the jth data map, represents the initial dimension of the feature map corresponding to the jth data map, represents the resolution of the feature map corresponding to the jth data map, Represents the step size parameter of the feature map corresponding to the jth data map, Represents the feature combination weight factor of the feature graph corresponding to the jth data graph, represents the maximum allowable filling parameter of the feature map corresponding to the jth data map, Indicates the filling parameters of the feature map corresponding to the j-th data map, Indicates the type of frequency feature information corresponding to the jth data spectrum, Represents high-frequency feature information, Represents low-frequency feature information.

Images