Visual customization method, system and storage medium for industrial Internet applications

Abstract

The present invention relates to the field of industrial Internet technology, and in particular to a visual customization method, system and storage medium for industrial Internet applications. The method comprises the following steps: performing process metadata aggregation processing on target production process document data to obtain industrial production process metadata; performing multi-granular hierarchical abstract processing based on the industrial production process metadata to generate dynamic view abstract tree data; performing spatial view layout processing based on the dynamic view abstract tree data to generate initial view layout data; performing layout adaptability optimization on the initial view layout data, and performing real-time data injection to generate visual production process data; performing user style customization processing on the visual production process data to obtain visual production process customization instance data. The present invention realizes a highly customized visual production process by performing multi-granular hierarchical abstract processing on the industrial production process.

Description

Technical Field

[0001] The present invention relates to the field of industrial Internet technology, and in particular to a visual customization method, system and storage medium for industrial Internet applications. Background Art

[0002] With the vigorous development of the industrial Internet, the degree of digitization and networking of the production process continues to increase, and massive amounts of data are generated and accumulated in the industrial production process. How to effectively use this data to achieve real-time monitoring, intelligent analysis and optimization of the production process has become one of the core goals of industrial Internet applications. Data visualization, as a bridge connecting people and data, can present complex and abstract data in an intuitive and easy-to-understand graphical way, helping users quickly understand the laws and values ​​behind the data, and plays a vital role in industrial Internet applications. However, traditional visualization customization methods for industrial Internet applications usually use fixed charts and dashboards, which are difficult to meet the urgent needs for visualization customization of production processes in industrial Internet application scenarios. Summary of the invention

[0003] Based on this, the present invention provides a visual customization method, system and storage medium for industrial Internet applications to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a visual customization method for industrial Internet applications includes the following steps:

[0005] Step 1: Perform process element entity recognition on the target production process document data to generate process element relationship network data; perform process metadata aggregation processing based on the process element relationship network data to obtain industrial production process metadata;

[0006] Step S2: Perform multi-granularity hierarchical abstract processing based on industrial production process metadata to generate multi-level granularity mapping matrix data; construct a fractal dimension tree based on the multi-level granularity mapping matrix data to obtain preliminary view abstract tree data; perform view granularity adaptive adjustment based on the preliminary view abstract tree data to generate dynamic view abstract tree data;

[0007] Step S3: performing spatial view layout processing according to the dynamic view abstract tree data to generate initial view layout data; performing layout adaptability optimization on the initial view layout data to generate high-quality view layout solution data;

[0008] Step S4: performing view component matching on the high-quality view layout scheme data to generate visualization component data; performing component binding processing on the high-quality view layout scheme data through the visualization component data, and performing real-time data injection to generate visualization production process data;

[0009] Step S5: Perform user style customization processing on the visualized production process data to generate customized style template data; perform customized style dynamic rendering processing according to the customized style template data to obtain customized instance data of the visualized production process.

[0010] The present invention automatically identifies key process elements from unstructured or semi-structured process documents, clarifies the association between process elements, realizes accurate identification of process elements and construction of relationship network, and extracts core industrial production process metadata. Through multi-granular hierarchical abstract processing, different levels and details of the production process are effectively divided, and the entire production process is divided into multiple processes. Each process can be further divided into multiple operation steps, which is convenient for users to understand the process from different abstract levels. According to user needs and data characteristics, the preliminary view abstract tree is dynamically adjusted. For example, users can expand or shrink nodes at a specific level as needed to achieve adaptive changes in view granularity and improve user experience. The screen layout is optimized according to the dynamic view abstract tree data, ensuring the reasonable distribution of visual elements and improving the logic and aesthetics of information display. According to the view layout scheme data, appropriate visualization components are selected, such as flowchart components, data chart components, text components, etc. Through component matching and binding, abstract data is converted into intuitive visualization components such as charts and icons, which greatly promotes intuitive transmission and rapid understanding of information. The real-time data injection mechanism ensures the timeliness of the displayed content, enabling users to monitor the production process status in real time and make decisions in time. According to the view layout scheme data, select appropriate visualization components, such as flowchart components, data chart components, text components, etc. This helps to present abstract process elements in a specific visualization form, while allowing users to personalize the visualization interface according to their preferences and needs, thereby improving users' understanding of the process and operating efficiency. Therefore, a visualization customization method for industrial Internet applications of the present invention understands process documents from a semantic level, builds a multi-level granularity process model, and performs adaptive view construction and layout optimization based on terminal devices and user interactions, and finally performs personalized style rendering based on user preferences, thereby achieving full process customization from data to view.

[0011] Preferably, step S1 comprises the following steps:

[0012] Step S11: performing deep semantic analysis on the target production process document data to generate production process semantic node data;

[0013] Step S12: semantic role labeling is performed on the production process semantic node data using preset process semantic role labeling rule data to obtain process role labeling data;

[0014] Step S13: Perform process element entity recognition based on process role labeling data, and perform context semantic relationship mining to generate process element relationship network data;

[0015] Step S14: Performing temporal process logic link analysis based on process element relationship network data to generate process flow link data;

[0016] Step S15: Perform process metadata aggregation processing based on the process flow link data to obtain industrial production process metadata.

[0017] Through deep semantic analysis, the present invention can convert text information in production process document data into semantic node data, and can understand the deep meaning of the text, including the semantic relationship between words, the logical relationship between sentence components, etc., so as to more accurately identify process elements and relationships. A predefined semantic role label is assigned to each semantic node to clarify the function and role of each node in the process. A predefined semantic role label is assigned to each semantic node to clarify the function and role of each node in the process. Semantic role labeling can understand the complex semantic relationship in the process, so as to more accurately construct a process element relationship network. Using the semantic role labeling information, the key entities in the process document can be more accurately identified to avoid misidentification caused by keyword ambiguity. The time sequence and logical relationship between the process elements are analyzed, and the discrete process elements are connected in series into a complete process flow link, clearly showing the execution path and steps of the process. The attribute information, semantic role, relationship information, etc. of the process elements are integrated into a unified metadata model to form structured industrial production process metadata.

[0018] Preferably, step S2 comprises the following steps:

[0019] Step S21: performing production process topology decomposition based on industrial production process metadata to generate production process topology data;

[0020] Step S22: Perform multi-granularity hierarchical abstract processing according to the production process topology data to generate multi-level granularity mapping matrix data;

[0021] Step S23: performing hierarchical attribute weight allocation according to the multi-level granularity mapping matrix data, and calculating the correlation strength, to generate multi-level process granularity correlation strength data;

[0022] Step S24: Based on the preset process association threshold, the multi-level process granularity association strength data is used to screen the multi-level granularity mapping matrix data for strongly associated process granularity, and view node generation is performed to generate view node data;

[0023] Step S25: constructing a fractal dimension tree according to the view node data to obtain preliminary view abstract tree data;

[0024] Step S26: performing a view tree stability assessment based on the preliminary view abstract tree data to generate view tree stability assessment data, performing view node optimization based on the view tree stability assessment data to generate an optimized view abstract tree;

[0025] Step S27: Adaptively adjust the view granularity of the optimized view abstract tree through a preset user interaction operation strategy to generate dynamic view abstract tree data, wherein the preset user interaction operation strategy includes a view zoom operation, a view filtering operation, a view drill-down operation, and a view selection operation.

[0026] The present invention converts the original linear process description information into a networked topological structure, and clearly displays the association relationship between various elements in the process. The complex process topological structure is hierarchically divided according to different granularities, so that users can understand the process from different abstract levels. Through the hierarchical attribute weight allocation, the importance of process elements at different granularity levels can be quantified according to factors such as the type, attribute, and role of the process elements, so as to highlight them in the visualization process. By setting the process association threshold, process elements with weak associations can be filtered out, key paths and core nodes can be retained, and the visualization interface can be avoided from being too complicated, thereby improving the user's attention to the key information of the process. The view nodes are organized into a tree structure using fractal theory to clearly display the hierarchical relationship of the process and the parent-child relationship between nodes. By evaluating the structural stability of the view tree, structural problems that cause visual confusion can be identified. According to the view tree stability evaluation results, the view nodes are optimized to generate an optimized view abstract tree with a clearer structure and a more reasonable layout. Users are allowed to dynamically adjust the granularity level and display content of the view according to their own needs and concerns through operations such as zooming, filtering, drilling down, and selecting, so as to realize personalized process information browsing and analysis.

[0027] Preferably, step S22 includes the following steps:

[0028] Step S221: labeling entity roles according to production process topology data to generate process role labeling data;

[0029] Step S222: performing workshop-level functional process decomposition on the production process topology data by using the process role labeling data to obtain workshop-level functional decomposition data;

[0030] Step S223: Performing production line-level operation link analysis according to the workshop-level function decomposition data to generate production line-level operation link data;

[0031] Step S224: identifying workstation-level activity nodes according to the production line-level operation link data, and generating workstation-level activity node data;

[0032] Step S225: extracting equipment-level configuration based on the workshop-level function decomposition data, the production line-level operation link data, and the workstation-level activity node data to obtain equipment-level configuration parameters;

[0033] Step S226: construct a granularity mapping relationship matrix based on the workshop-level function decomposition data, the production line-level operation link data, the workstation-level activity node data, and the equipment-level configuration parameters to generate multi-level granularity mapping matrix data.

[0034] The present invention assigns a specific role label to each entity in the production process topology data, clarifies the role and function of each entity in the production process, and thus performs targeted analysis and optimization. According to the process role labeling information, the complex production process is functionally divided according to the workshop level. For example, the functions such as "raw material processing", "parts production" and "product assembly" are divided into different workshops to form workshop-level functional decomposition data, which helps users understand the organizational structure of the production process from a macro perspective. For the production activities in each workshop, the connection relationship and execution order between each process are analyzed, and the production line-level operation link data is constructed to describe the production process in each workshop in detail. The production line-level operation link data can be used to analyze the production line bottleneck, optimize the production rhythm, balance the production load, etc., to improve the production efficiency and flexibility of the production line. The production line-level operation link is refined to the specific workstation level, and the specific operation activities that need to be performed at each workstation are identified to achieve a refined description of the production process. The equipment-level configuration parameters provide a basis for realizing the interconnection and interoperability of equipment. For example, the data interface can be configured according to the equipment parameters to realize data exchange and collaborative work between different equipment. Integrate process information of different granularities into a unified matrix structure, establish a mapping relationship between different granularity levels, support users to analyze and optimize production processes from different granularity levels, and go deep into specific workstations or equipment levels to troubleshoot problems and improve efficiency.

[0035] Preferably, step S3 comprises the following steps:

[0036] Step S31: performing tree structure traversal processing on the dynamic view abstract tree data, and performing view topology projection transformation processing to generate view projection transformation data;

[0037] Step S32: using the terminal device to obtain screen size limitation parameters;

[0038] Step S33: constructing a constraint space for the view projection transformation data using screen size restriction parameters to generate visualization layout constraint space data;

[0039] Step S34: performing spatial view layout processing on the view projection transformation data using the visualization layout constraint space data, and generating initial view layout data and view layout adaptability score data respectively;

[0040] Step S35: optimizing the layout of the initial view layout data using preset aesthetic rule data, and performing visual aesthetic scoring, thereby generating view layout scoring data and optimized view layout data;

[0041] Step S36: performing comprehensive weighted scoring processing according to the view layout scoring data and the view layout adaptability scoring data to generate view layout comprehensive scoring data;

[0042] Step S37: Perform view scheme scoring judgment based on the view layout comprehensive scoring data. When the view layout comprehensive scoring data is higher than or equal to the preset layout scoring threshold, mark the optimized view layout data as high-quality view layout scheme data; when the view layout comprehensive scoring data is lower than the preset layout scoring threshold, mark the optimized view layout data as low-quality view layout scheme data, eliminate the low-quality view layout scheme data, and return to step S34 to readjust the view layout.

[0043] The present invention can convert each node in the tree structure into a point in a two-dimensional space by performing tree structure traversal processing on the dynamic view abstract tree data. The screen size information of the terminal device is obtained to constrain the spatial range of the view layout, ensure that the generated view layout can be completely and clearly displayed on the user's device, and improve the user experience. According to the screen size restriction parameter, the boundary range of the view layout is determined to ensure that it will not exceed the screen range, and avoid incomplete information display or exceeding the boundary. The view node is placed in the visual layout constraint space, that is, the specific position and size information of each node on the screen. According to the preset aesthetic rules, the initial view layout is fine-tuned to improve the aesthetics and readability of the view. The scoring data of the two aspects of layout adaptability and visual aesthetics are weighted averaged, and the rationality and aesthetics of the layout are comprehensively considered to avoid the evaluation deviation caused by a single indicator. The view scheme is scored and judged according to the comprehensive scoring data of the view layout, and the quality of the view layout scheme can be judged according to the scoring result, and corresponding processing is performed to ensure that the final generated view layout is both reasonable and beautiful.

[0044] Preferably, step S34 includes the following steps:

[0045] Step S341: constructing a view node adjacency matrix according to the view projection transformation data;

[0046] Step S342: recursively partition and layout the visualization layout constraint space data through the view node adjacency matrix to generate view partition space domain data;

[0047] Step S343: Calculate screen space utilization according to the view segmentation space domain data to generate screen space utilization data;

[0048] Step S344: Calculate the node importance using the view node adjacency matrix to generate node importance data; calculate the ideal node spacing of the view node adjacency matrix using the screen space utilization data and the node importance data to generate the ideal node spacing data;

[0049] Step S345: performing node coordinate mapping processing on the view node ideal spacing data through the view node adjacency matrix to generate view node layout coordinate data;

[0050] Step S346: performing connection path planning processing according to the view node layout coordinate data to generate view node connection line data;

[0051] Step S347: performing edge line intersection detection on the view node connection line data, and performing edge line trend optimization processing through Bezier curves to generate node edge line trend data;

[0052] Step S348: adjusting the view node layout coordinate data using the node edge direction data to generate initial view layout data;

[0053] Step S349: Evaluate the layout adaptability of the node edge direction data using the view node ideal spacing data to generate view layout adaptability score data.

[0054] The present invention stores the connection relationship between view nodes in the form of a matrix, and each element in the matrix represents whether there is a connection between two nodes, as well as information such as the weight or type of the connection. According to the connection relationship between the nodes, the visual layout constraint space is recursively divided into multiple sub-areas, each sub-area contains one or more nodes, and closely connected nodes are placed in adjacent sub-areas as much as possible to avoid the nodes being too far apart or crossing too much, thereby improving space utilization. The screen space utilization rate is calculated based on the view segmentation space domain data, and the space utilization efficiency of the current view layout can be evaluated. According to the status, type, attributes and other information of the node in the process, the importance weight of each node is calculated to give more space and a more eye-catching position during layout. According to the ideal node spacing and the view segmentation space domain data, each node is mapped to a specific coordinate position in the layout space, and the ideal spacing between nodes is met as much as possible to avoid node overlap or overcrowding. According to the coordinate position of the node, the path of the connection line in the layout space is planned. Whether there is intersection between the connection lines is detected, and the path of the connection line is optimized using a smooth curve such as a Bezier curve to minimize the situation of edge line intersection, so that the connection line is smoother and more beautiful. According to the optimized connection line path, the coordinate position of the node is fine-tuned to better adapt to the direction of the connection line and avoid nodes and connection lines being too close or overlapping. The fitness of the layout is evaluated based on the ideal spacing of the nodes and the direction of the connection line to judge the fitness of the layout.

[0055] Preferably, step S4 comprises the following steps:

[0056] Step S41: performing view function structure analysis on the high-quality view layout solution data to generate view structure analysis data;

[0057] Step S42: performing visual element index processing according to the view structure parsing data to generate visual element index data;

[0058] Step S43: using the visual element index data to match the view components of the preset visualization component library to generate visualization component data;

[0059] Step S44: performing component binding processing on the high-quality view layout solution data through the visualization component data to obtain production process view component data;

[0060] Step S45: using the industrial Internet platform to collect real-time production process data and generate real-time production process data;

[0061] Step S46: Perform feature engineering processing according to the real-time production process data to generate standard production process data;

[0062] Step S47: constructing an interactive data injection channel based on the production process view component data and the standard production process data;

[0063] Step S48: Based on the interactive data injection channel, the standard production process data is injected into the production process view component data in real time to generate visual production process data.

[0064] The high-quality view layout scheme data of the present invention performs view functional structure analysis, analyzes the high-quality view layout scheme data, identifies the various functional elements contained in the view, and the hierarchical relationship and attribute information between them. A unique index identifier is assigned to each functional element in the view, and the index identifier is associated with the type, attribute, position and other information of the element, so as to facilitate the subsequent rapid search and access to specific view elements. According to the type, function and data characteristics of the visual element, the most suitable component is selected from the preset visual component library for matching, and the visual component data contains the visual component information corresponding to each visual element. The matched visual component is bound to the corresponding element in the view layout scheme, and the component's attributes, style, event and other information are associated with the element information, so that the static view layout is transformed into an interactive and dynamically updated view interface. Various sensors, equipment and systems at the production site are connected through the industrial Internet platform, and various data in the production process are collected in real time to provide raw data for subsequent data injection. The collected real-time production process data is cleaned, filtered, converted, aggregated and other processes are performed to meet the data format requirements of the visual component and improve the data quality and visualization effect. Establish a mapping relationship between standard production process data and production process view component data to ensure that data can be accurately injected into the corresponding components. Transmit the processed standard production process data to the corresponding visualization component in real time through the data injection channel, and trigger the component update mechanism, so that the view can dynamically reflect the real-time status and change trend of the production process, helping users to monitor the production process in real time, find problems in time, make decisions, and improve production efficiency and management level.

[0065] Preferably, step S5 comprises the following steps:

[0066] Step S51: using the terminal device to perform visual display processing on the visualized production process data, and collecting style demand feedback from users to generate user demand feedback data;

[0067] Step S52: performing style preference portrait analysis based on user demand feedback data to generate style preference portrait data;

[0068] Step S53: performing style template customization processing according to the style preference portrait data to generate customized style template data;

[0069] Step S54: performing customized style dynamic rendering processing on the visualized production process data through the customized style template data, thereby obtaining the visualized production process customized instance data.

[0070] The present invention utilizes a terminal device to perform visual display processing on the visualized production process data, and can display the production process information to the user in an intuitive and easy-to-understand manner. By collecting user style demand feedback, the user's opinions and suggestions on the current view style can be collected. The user demand feedback data provides a basis for subsequent style customization, and the style can be adjusted according to the user's feedback information to meet the user's personalized needs. According to the user demand feedback data, a style preference portrait analysis is performed to analyze the user's style preferences and identify the user's common and personalized needs. According to the user's style preference portrait data, the system presets a variety of visual style templates to adjust and combine, and generate customized style template data that meets the user's personalized needs. The visualized production process data is rendered using the customized style template data, that is, the data is combined with customized visual elements, layouts, interactive methods, etc., and the results are presented on the user's terminal device, so that the user can view and analyze the production process information in his favorite style, thereby obtaining a personalized aesthetic and customary view experience.

[0071] Preferably, the present invention further provides a visual customization system for industrial Internet applications, which executes the visual customization method for industrial Internet applications as described above, and the visual customization system for industrial Internet applications comprises:

[0072] The process element parsing module is used to identify the process element entities of the target production process document data and generate process element relationship network data; process metadata aggregation is performed based on the process element relationship network data to obtain industrial production process metadata;

[0073] The multi-granularity view construction module is used to perform multi-granularity hierarchical abstract processing based on industrial production process metadata to generate multi-level granularity mapping matrix data; construct a fractal dimension tree based on the multi-level granularity mapping matrix data to obtain preliminary view abstract tree data; perform view granularity adaptive adjustment based on the preliminary view abstract tree data to generate dynamic view abstract tree data;

[0074] The view layout optimization module is used to perform spatial view layout processing according to the dynamic view abstract tree data to generate initial view layout data; perform layout adaptability optimization on the initial view layout data to generate high-quality view layout solution data;

[0075] The view component rendering module is used to match the high-quality view layout scheme data with view components to generate visual component data; perform component binding processing on the high-quality view layout scheme data through the visual component data, and perform real-time data injection to generate visual production process data;

[0076] The user style customization module is used to perform user style customization processing on the visualized production process data and generate customized style template data; perform customized style dynamic rendering processing according to the customized style template data to obtain the visualized production process customized instance data.

[0077] Preferably, a visual customization storage medium for industrial Internet applications stores a computer program, and when the computer program is executed, it implements a visual customization method for industrial Internet applications as described in any one of the above. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 A schematic diagram of the steps of the visual customization method for industrial Internet applications of the present invention;

[0079] Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart;

[0080] Figure 3 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0081] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0082] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0083] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0084] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0085] To achieve this, please refer to Figures 1 to 3 The present invention provides a visual customization method for industrial Internet applications, comprising the following steps:

[0086] Step 1: Perform process element entity recognition on the target production process document data to generate process element relationship network data; perform process metadata aggregation processing based on the process element relationship network data to obtain industrial production process metadata;

[0087] Step S2: Perform multi-granularity hierarchical abstract processing based on industrial production process metadata to generate multi-level granularity mapping matrix data; construct a fractal dimension tree based on the multi-level granularity mapping matrix data to obtain preliminary view abstract tree data; perform view granularity adaptive adjustment based on the preliminary view abstract tree data to generate dynamic view abstract tree data;

[0088] Step S3: performing spatial view layout processing according to the dynamic view abstract tree data to generate initial view layout data; performing layout adaptability optimization on the initial view layout data to generate high-quality view layout solution data;

[0089] Step S4: performing view component matching on the high-quality view layout scheme data to generate visualization component data; performing component binding processing on the high-quality view layout scheme data through the visualization component data, and performing real-time data injection to generate visualization production process data;

[0090] Step S5: Perform user style customization processing on the visualized production process data to generate customized style template data; perform customized style dynamic rendering processing according to the customized style template data to obtain customized instance data of the visualized production process.

[0091] The present invention automatically identifies key process elements from unstructured or semi-structured process documents, clarifies the association between process elements, realizes accurate identification of process elements and construction of relationship network, and extracts core industrial production process metadata. Through multi-granular hierarchical abstract processing, different levels and details of the production process are effectively divided, and the entire production process is divided into multiple processes. Each process can be further divided into multiple operation steps, which is convenient for users to understand the process from different abstract levels. According to user needs and data characteristics, the preliminary view abstract tree is dynamically adjusted. For example, users can expand or shrink nodes at a specific level as needed to achieve adaptive changes in view granularity and improve user experience. The screen layout is optimized according to the dynamic view abstract tree data, ensuring the reasonable distribution of visual elements and improving the logic and aesthetics of information display. According to the view layout scheme data, appropriate visualization components are selected, such as flowchart components, data chart components, text components, etc. Through component matching and binding, abstract data is converted into intuitive visualization components such as charts and icons, which greatly promotes intuitive transmission and rapid understanding of information. The real-time data injection mechanism ensures the timeliness of the displayed content, enabling users to monitor the production process status in real time and make decisions in time. According to the view layout scheme data, select appropriate visualization components, such as flowchart components, data chart components, text components, etc. This helps to present abstract process elements in a specific visualization form, while allowing users to personalize the visualization interface according to their preferences and needs, thereby improving users' understanding of the process and operating efficiency. Therefore, a visualization customization method for industrial Internet applications of the present invention understands process documents from a semantic level, builds a multi-level granularity process model, and performs adaptive view construction and layout optimization based on terminal devices and user interactions, and finally performs personalized style rendering based on user preferences, thereby achieving full process customization from data to view.

[0092] In the embodiment of the present invention, reference Figure 1 The above is a schematic diagram of the steps of the visual customization method for industrial Internet applications of the present invention. In this embodiment, the visual customization method for industrial Internet applications includes the following steps:

[0093] Step 1: Perform process element entity recognition on the target production process document data to generate process element relationship network data; perform process metadata aggregation processing based on the process element relationship network data to obtain industrial production process metadata;

[0094] In an embodiment of the present invention, natural language processing technology can be used to analyze the target production process document. For example, for the stamping production line process document of a certain automobile manufacturing plant, natural language processing tools (such as SpaCy, HanLP) are used to perform text analysis on the document to identify process element entities, such as "stamping process", "robot handling", and "quality inspection". The relationship between entities is extracted, for example, "stamping process" is followed by "robot handling", and "robot handling" is connected to "quality inspection". The extracted entities and relationships are constructed into process element relationship network data. Subsequently, these entities and relationships are aggregated using graph databases or knowledge graph technologies to summarize the metadata model of the industrial production process, including but not limited to core information such as process nodes, flow direction, and time cost. For example, all attribute information of the "stamping process" (such as process name, equipment number, operator, processing time, etc.) is aggregated to the entity node to form industrial production process metadata.

[0095] Step S2: Perform multi-granularity hierarchical abstract processing based on industrial production process metadata to generate multi-level granularity mapping matrix data; construct a fractal dimension tree based on the multi-level granularity mapping matrix data to obtain preliminary view abstract tree data; perform view granularity adaptive adjustment based on the preliminary view abstract tree data to generate dynamic view abstract tree data;

[0096] In the embodiment of the present invention, the complex industrial production process is hierarchically divided according to different granularities, and a hierarchical clustering algorithm or a top-down decomposition strategy can be used to perform multi-granular hierarchical abstraction on the process. For example, multiple entity nodes such as "stamping process", "robot handling", and "quality inspection" are abstracted into higher-level abstract nodes such as "stamping workshop", "welding workshop", and "painting workshop", and multi-level granularity mapping matrix data is established to record the mapping relationship between nodes of different granularities. According to the multi-level granularity mapping matrix data, a fractal dimension tree is constructed using a tree structure construction strategy to generate preliminary view abstract tree data. The preliminary view abstract tree data reflects the hierarchical relationship between process nodes of different granularities, for example, "stamping workshop" is used as a parent node, and its child nodes include "stamping process", "robot handling", "quality inspection", etc. Subsequently, the tree structure is pruned and merged, such as merging nodes with high similarity and deleting nodes with small information content, to form a preliminary view abstract tree. Finally, according to user interaction or data characteristics, the tree structure is dynamically adjusted, such as expanding or shrinking certain nodes, to obtain a dynamic view abstract tree that can be adaptively adjusted.

[0097] Step S3: performing spatial view layout processing according to the dynamic view abstract tree data to generate initial view layout data; performing layout adaptability optimization on the initial view layout data to generate high-quality view layout solution data;

[0098] In an embodiment of the present invention, a tree layout algorithm (such as the Reingold-Tilford algorithm) is used to perform spatial view layout processing to generate initial view layout data. For example, the "stamping shop" node is placed in the center of the view, and sub-nodes such as "stamping process", "robot handling", and "quality inspection" are arranged in a ring around it. In order to avoid node overlap and line crossing, the initial view layout data is optimized for layout adaptability, such as avoiding line crossing, optimizing information flow, etc. The node positions are iteratively adjusted to maintain a reasonable spacing between nodes and minimize line crossing, and finally generate high-quality view layout solution data.

[0099] Step S4: performing view component matching on the high-quality view layout scheme data to generate visualization component data; performing component binding processing on the high-quality view layout scheme data through the visualization component data, and performing real-time data injection to generate visualization production process data;

[0100] In an embodiment of the present invention, for example, the "stamping process" node is matched as a rectangular frame component, the "robot handling" node is matched as a robot icon component, and the "quality inspection" node is matched as a dashboard component, and the style and properties of the component are set to generate visual component data. The high-quality view layout scheme data is subjected to component binding processing using the visual component data. For example, the rectangular frame component of the "stamping process" node is bound to its corresponding coordinate position, and the attribute information of the node (such as process name, equipment number) is displayed on the component. Production data is collected in real time through the industrial Internet platform, and real-time data injection is performed. For example, the robot icon component of the "robot handling" node is bound to the actual robot's operating status data, and the robot's operating trajectory, speed and other information are displayed in real time, and finally visual production process data is generated.

[0101] Step S5: Perform user style customization processing on the visualized production process data to generate customized style template data; perform customized style dynamic rendering processing according to the customized style template data to obtain customized instance data of the visualized production process.

[0102] In an embodiment of the present invention, a user can select different theme colors, font styles, layout methods, etc. for personalized configuration to generate customized style template data. For example, a user can select a blue theme, Microsoft YaHei font, a tree layout, and other configurations. Customized style dynamic rendering processing is performed according to the customized style template data. For example, according to the blue theme selected by the user, the colors of all nodes and lines are set to blue; according to the Microsoft YaHei font selected by the user, all text information is set to Microsoft YaHei font; according to the tree layout selected by the user, all nodes are arranged in a tree structure, and finally the visualized production process customized instance data is obtained to meet the personalized needs of the user.

[0103] Preferably, step S1 comprises the following steps:

[0104] Step S11: performing deep semantic analysis on the target production process document data to generate production process semantic node data;

[0105] Step S12: semantic role labeling is performed on the production process semantic node data using preset process semantic role labeling rule data to obtain process role labeling data;

[0106] Step S13: Perform process element entity recognition based on process role labeling data, and perform context semantic relationship mining to generate process element relationship network data;

[0107] Step S14: Performing temporal process logic link analysis based on process element relationship network data to generate process flow link data;

[0108] Step S15: Perform process metadata aggregation processing based on the process flow link data to obtain industrial production process metadata.

[0109] In an embodiment of the present invention, for example, for a CNC machine tool processing process document of a certain factory, the document is deeply semantically parsed using dependency syntax analysis and semantic role labeling tools. Each semantic node contains key actions, participants and objects in the process, forming production process semantic node data. Based on the preset process semantic role labeling rule data, for example, verbs such as "start" and "operate" correspond to the "executor" role, and nouns such as "machine tool" and "parameter" correspond to the "resource" role, and the production process semantic node data generated in step S11 are semantically role labeled. For example, the "operator" in the node (operator, start, machine tool) is labeled as the "executor" role, and the "machine tool" is labeled as the "resource" role. Finally, the process role labeling data is obtained, for example: (operator-executor, start, machine tool-resource), (operator-executor, setting, parameter-resource). The semantic nodes marked as specific roles are identified as process element entities. For example, all nodes marked with the "resource" role ("machine tool-resource", "parameter-resource") are identified as process element entities. The contextual semantic relationship between process element entities is mined by using methods such as syntactic analysis and semantic similarity calculation. For example, by analyzing the sentence "After the operator starts the machine tool, parts processing is performed", it can be identified that there is a temporal "follow-up" relationship between "starting the machine tool" and "part processing". The identified process element entities and relationships are constructed into process element relationship network data. Combined with the temporal relationship between process element entities (for example, "starting" precedes "processing", and "setting parameters" and "starting" are parallel relationships), temporal process logic link analysis is performed. For example, it can be identified that "setting parameters" -> "starting the machine tool" -> "part processing" constitutes a complete process link. The process metadata related to each process is aggregated. For example, the processing time, processing accuracy, required equipment and other information of the "part processing" process are aggregated to the process node, and finally the industrial production process metadata is obtained.

[0110] Preferably, step S2 comprises the following steps:

[0111] Step S21: performing production process topology decomposition based on industrial production process metadata to generate production process topology data;

[0112] Step S22: Perform multi-granularity hierarchical abstract processing according to the production process topology data to generate multi-level granularity mapping matrix data;

[0113] Step S23: performing hierarchical attribute weight allocation according to the multi-level granularity mapping matrix data, and calculating the correlation strength, to generate multi-level process granularity correlation strength data;

[0114] Step S24: Based on the preset process association threshold, the multi-level process granularity association strength data is used to screen the multi-level granularity mapping matrix data for strongly associated process granularity, and view node generation is performed to generate view node data;

[0115] Step S25: constructing a fractal dimension tree according to the view node data to obtain preliminary view abstract tree data;

[0116] Step S26: performing a view tree stability assessment based on the preliminary view abstract tree data to generate view tree stability assessment data, performing view node optimization based on the view tree stability assessment data to generate an optimized view abstract tree;

[0117] Step S27: Adaptively adjust the view granularity of the optimized view abstract tree through a preset user interaction operation strategy to generate dynamic view abstract tree data, wherein the preset user interaction operation strategy includes a view zoom operation, a view filtering operation, a view drill-down operation, and a view selection operation.

[0118] As an example of the present invention, refer to Figure 2 As shown, Figure 1 Detailed implementation steps of step S2 in the embodiment are shown in the flowchart. In this embodiment, step S2 includes:

[0119] Step S21: performing production process topology decomposition based on industrial production process metadata to generate production process topology data;

[0120] In an embodiment of the present invention, for example, for a factory's production process metadata, it contains multiple process links such as raw material processing, parts assembly, and product testing, as well as the input and output, equipment resources, process parameters, and other information of each link. A graph theory algorithm (such as a directed graph) is used to abstract each process link in the process metadata into a node, and the input and output relationship between the links is abstracted into a directed edge, to construct a production process topology graph, and to store it as production process topology data. For example, the "raw material processing" link is used as the starting node, connected to the "parts assembly" link, and then connected to the "product testing" link, to form a simple production process topology structure.

[0121] Step S22: Perform multi-granularity hierarchical abstract processing according to the production process topology data to generate multi-level granularity mapping matrix data;

[0122] In an embodiment of the present invention, a complex process is subdivided into sub-processes at different levels, and each level focuses on details at different levels. By quantifying the characteristics and importance of each sub-process, a multi-level granularity mapping matrix data is formed, which clearly shows the mapping relationship and inclusion hierarchy between each level of granularity. For example, multiple similar "component assembly" links can be abstracted into a higher-level "assembly workshop" node, and multiple "product testing" links can be abstracted into a "test center" node. Establish multi-level granularity mapping matrix data to record the mapping relationship between nodes of different granularities, such as which "component assembly" links are included in the "assembly workshop" node, which "product testing" links are included in the "test center" node, and other information.

[0123] Step S23: performing hierarchical attribute weight allocation according to the multi-level granularity mapping matrix data, and calculating the correlation strength, to generate multi-level process granularity correlation strength data;

[0124] In an embodiment of the present invention, attribute weights are assigned to process nodes at different levels based on the domain knowledge or expert experience of the process. For example, the weight of the "raw material processing" link is 0.8, the weight of the "parts assembly" link is 0.6, and the weight of the "assembly workshop" node is 0.4. At the same time, the strength of the association relationship is calculated based on factors such as the connection relationship between the nodes, the frequency of data interaction, and the number of shared resources. For example, if the frequency of data interaction between the "raw material processing" link and the "parts assembly" link is very high, the strength of the association relationship between them is relatively high. The calculated hierarchical attribute weights and association relationship strengths are integrated into the multi-level granularity mapping matrix data to form multi-level process granularity association strength data.

[0125] Step S24: Based on the preset process association threshold, the multi-level process granularity association strength data is used to screen the multi-level granularity mapping matrix data for strongly associated process granularity, and view node generation is performed to generate view node data;

[0126] In an embodiment of the present invention, for example, the preset process association threshold is 0.5, and based on the multi-level process granularity association strength data, process node pairs with an association strength higher than the threshold are screened out. For example, if the association strength between the "raw material processing" link and the "parts assembly" link is 0.7, which is higher than the threshold, the connection relationship is retained. The multi-level granularity mapping matrix data is trimmed according to the screening results, and the strongly associated process nodes are retained and used as view nodes to generate view node data. For example, a strong association is formed between the "raw material processing" link, the "parts assembly" link, and the "assembly workshop" nodes, and these three nodes are retained as view nodes.

[0127] Step S25: constructing a fractal dimension tree according to the view node data to obtain preliminary view abstract tree data;

[0128] In the embodiment of the present invention, the view node data is used to construct the preliminary view abstract tree data through fractal theory and recursive algorithm. This tree structure shows the organization of the process in a hierarchical manner, and the nodes at each level represent the view overview of different granularities. For example, the "assembly workshop" node is used as the parent node, and the "raw material processing" link and the "parts assembly" link are used as its child nodes to construct a simple tree structure.

[0129] Step S26: performing a view tree stability assessment based on the preliminary view abstract tree data to generate view tree stability assessment data, performing view node optimization based on the view tree stability assessment data to generate an optimized view abstract tree;

[0130] In an embodiment of the present invention, a view tree stability evaluation is performed on preliminary view abstract tree data according to preset evaluation indicators, such as the depth of the tree, the number of nodes, the number of associations, etc., to generate view tree stability evaluation data. For example, if the depth of the tree is too large and the number of nodes is too large, the tree structure is considered to be too complex and the stability is poor. Based on the evaluation results, the view nodes are optimized, such as merging or deleting some nodes, adjusting the connection relationship between nodes, etc., to generate an optimized view abstract tree. For example, some nodes with weaker associations can be merged, or some nodes that have little impact on the overall structure can be deleted, thereby simplifying the tree structure and improving stability.

[0131] Step S27: Adaptively adjust the view granularity of the optimized view abstract tree through a preset user interaction operation strategy to generate dynamic view abstract tree data, wherein the preset user interaction operation strategy includes a view zoom operation, a view filtering operation, a view drill-down operation, and a view selection operation.

[0132] In the embodiment of the present invention, the preset user interaction operation strategy includes but is not limited to: "View zooming operation: the user can zoom in or out the view by using the mouse wheel or zoom gesture, enlarge or reduce the view ratio, and view process information of different granularities. For example, when the user zooms in the view, the detailed information of a node can be viewed; when the user reduces the view, the overview information of the entire process can be viewed; View filtering operation: the user can filter the view according to conditions such as keywords, time range, node type, etc., and only display the nodes and connection relationships that meet the conditions. For example, the user can set to display only nodes related to "raw material processing", or only display data interaction relationships that occur within a specific time period." etc. According to the preset user interaction operation strategy, the optimized view abstract tree is adaptively adjusted to the view granularity, and dynamic view abstract tree data is generated. For example, when the user performs a view zooming operation, the size and position of the node, as well as the number of nodes displayed and the amount of information are dynamically adjusted according to the zoom ratio; when the user performs a view filtering operation, the corresponding nodes and connection relationships are dynamically displayed or hidden according to the filtering conditions. The dynamic view abstract tree data can adjust the granularity and content of the view in real time according to the user's interactive operation.

[0133] Preferably, step S22 includes the following steps:

[0134] Step S221: labeling entity roles according to production process topology data to generate process role labeling data;

[0135] Step S222: performing workshop-level functional process decomposition on the production process topology data by using the process role labeling data to obtain workshop-level functional decomposition data;

[0136] Step S223: Performing production line-level operation link analysis according to the workshop-level function decomposition data to generate production line-level operation link data;

[0137] Step S224: identifying workstation-level activity nodes according to the production line-level operation link data, and generating workstation-level activity node data;

[0138] Step S225: extracting equipment-level configuration based on the workshop-level function decomposition data, the production line-level operation link data, and the workstation-level activity node data to obtain equipment-level configuration parameters;

[0139] Step S226: construct a granularity mapping relationship matrix based on the workshop-level function decomposition data, the production line-level operation link data, the workstation-level activity node data, and the equipment-level configuration parameters to generate multi-level granularity mapping matrix data.

[0140] In an embodiment of the present invention, for example, the production process topology data of a certain automobile manufacturing plant includes multiple production links such as stamping, welding, painting, and assembly, as well as the logistics and information flow relationships between each link. Using a pre-defined entity role labeling system, such as "raw materials", "semi-finished products", "finished products", "equipment", "process", etc., each node in the production process topology data is labeled with an entity role. For example, the "steel plate storage" node is labeled as "raw materials", the "stamping forming" node is labeled as "process", and the "welding robot" node is labeled as "equipment" to generate process role labeling data. Nodes with the same or similar functions are grouped and classified into corresponding workshops. For example, all nodes related to "stamping", including "steel plate cutting", "stamping forming", etc., are classified into "stamping workshop". The functional process of each workshop is abstracted into a subgraph to form workshop-level functional decomposition data. For example, the "stamping workshop" subgraph analyzes the temporal relationship and logical relationship between each node and identifies the operation link therein. For example, in the "Stamping Workshop" subgraph, we can identify the operation links such as "steel plate storage" -> "steel plate cutting" -> "stamping forming" -> "stamping parts inspection" -> "stamping parts out of storage", which represents the complete stamping process of steel plates from storage to out of storage. Each node is further refined into specific station-level activity nodes. For example, the "steel plate cutting" node can be refined into station-level activity nodes such as "loading", "positioning", "cutting", and "unloading". All station-level activity nodes and their sequence are recorded to generate station-level activity node data. The equipment resources and configuration parameters required for each station-level activity node are analyzed, and the equipment resources and configuration parameters required for each station-level activity node are extracted to form equipment-level configuration parameter data. According to the workshop-level functional decomposition data, production line-level operation link data, station-level activity node data, and equipment-level configuration parameters, a granularity mapping relationship matrix is ​​constructed. This matrix shows the detailed mapping logic from macro to micro by encoding the association relationship between different levels (workshop, production line, station, equipment). For example, an entry in the matrix can indicate which workstations are included in a production line of a workshop, and which equipment and their parameters are configured at these workstations.

[0141] Preferably, step S3 comprises the following steps:

[0142] Step S31: performing tree structure traversal processing on the dynamic view abstract tree data, and performing view topology projection transformation processing to generate view projection transformation data;

[0143] Step S32: using the terminal device to obtain screen size limitation parameters;

[0144] Step S33: constructing a constraint space for the view projection transformation data using screen size restriction parameters to generate visualization layout constraint space data;

[0145] Step S34: performing spatial view layout processing on the view projection transformation data using the visualization layout constraint space data, and generating initial view layout data and view layout adaptability score data respectively;

[0146] Step S35: optimizing the layout of the initial view layout data using preset aesthetic rule data, and performing visual aesthetic scoring, thereby generating view layout scoring data and optimized view layout data;

[0147] Step S36: performing comprehensive weighted scoring processing according to the view layout scoring data and the view layout adaptability scoring data to generate view layout comprehensive scoring data;

[0148] Step S37: Perform view scheme scoring judgment based on the view layout comprehensive scoring data. When the view layout comprehensive scoring data is higher than or equal to the preset layout scoring threshold, mark the optimized view layout data as high-quality view layout scheme data; when the view layout comprehensive scoring data is lower than the preset layout scoring threshold, mark the optimized view layout data as low-quality view layout scheme data, eliminate the low-quality view layout scheme data, and return to step S34 to readjust the view layout.

[0149] In an embodiment of the present invention, a tree structure is traversed using a depth-first or breadth-first algorithm to obtain the hierarchical relationship and position information of each node. The node information of the tree structure is mapped to a two-dimensional plane coordinate system to perform a view topology projection transformation. For example, a tree layout algorithm (such as the Reingold-Tilford algorithm) can be used to convert the tree structure into a two-dimensional layout in which a certain distance is maintained between nodes and the lines are as non-crossed as possible. The coordinate information and connection relationship of each node are saved in the view projection transformation data. In order to meet the display requirements of different terminal devices, the system needs to obtain the screen size restriction parameters of the target display device, which can be achieved through the JavaScript screen object or other device APIs. Parameters include but are not limited to screen resolution, aspect ratio, available display area, etc. According to the screen size restriction parameters, visual layout constraint space data is constructed to limit the scope of the view layout. For example, the screen width and height can be used as the maximum values ​​of the horizontal and vertical coordinates, respectively, to construct a rectangular area as the visual layout constraint space, and ensure that all nodes and connecting lines fall within the area. The view projection transformation data is adaptively adjusted to generate initial view layout data to ensure that the layout is suitable for the target screen. At the same time, the adaptability of the layout is scored through algorithm evaluation (such as entropy-based evaluation method, space utilization index), and view layout adaptability scoring data is formed. At the same time, the initial view layout data is scored for adaptability based on indicators such as node distribution uniformity, number of connection line crossings, and whether the node exceeds the boundary, and view layout adaptability scoring data is generated. For example, the more uniform the node distribution and the fewer connection line crossings, the higher the score. The preset aesthetic rule data includes rules on node size, color, shape, connection line style, layout symmetry, etc. For example, the size of important nodes should be larger and the color should be brighter; the connection line should be as smooth as possible to avoid sharp bends. The initial view layout data is optimized according to the aesthetic rule data, such as adjusting the node size, color, shape, connection line style, etc., to generate optimized view layout data. The optimized layout is scored for visual aesthetics according to the aesthetic rule data to generate view layout scoring data. For example, the more symmetrical the layout and the more beautiful the style of the nodes and connection lines, the higher the score. According to the view layout adaptability scoring data and the view layout scoring data, a comprehensive weighted scoring process is performed to generate view layout comprehensive scoring data. For example, different weights may be assigned to the adaptability score and the aesthetic score, such as 0.6 and 0.4, and then the weighted sum is used to obtain the comprehensive score. The preset layout score threshold is 0.8. The view scheme score is determined based on the view layout comprehensive score data generated in step S36. If the view layout comprehensive score data is higher than or equal to the preset layout score threshold, it is considered that the layout scheme has high quality, and the optimized view layout data is marked as high-quality view layout scheme data, and the layout scheme is output.If the comprehensive view layout score data is lower than the preset layout score threshold, the layout scheme is considered to have low quality, the optimized view layout data is marked as low-quality view layout scheme data, and the low-quality view layout scheme data is eliminated, and the process returns to step S34. The view projection transformation data is re-layouted using different layout parameters or random perturbations, and steps S34 to S37 are repeated until a high-quality view layout scheme that meets the score threshold is found.

[0150] Preferably, step S34 includes the following steps:

[0151] Step S341: constructing a view node adjacency matrix according to the view projection transformation data;

[0152] Step S342: recursively partition and layout the visualization layout constraint space data through the view node adjacency matrix to generate view partition space domain data;

[0153] Step S343: Calculate screen space utilization according to the view segmentation space domain data to generate screen space utilization data;

[0154] Step S344: Calculate the node importance using the view node adjacency matrix to generate node importance data; calculate the ideal node spacing of the view node adjacency matrix using the screen space utilization data and the node importance data to generate the ideal node spacing data;

[0155] Step S345: performing node coordinate mapping processing on the view node ideal spacing data through the view node adjacency matrix to generate view node layout coordinate data;

[0156] Step S346: performing connection path planning processing according to the view node layout coordinate data to generate view node connection line data;

[0157] Step S347: performing edge line intersection detection on the view node connection line data, and performing edge line trend optimization processing through Bezier curves to generate node edge line trend data;

[0158] Step S348: adjusting the view node layout coordinate data using the node edge direction data to generate initial view layout data;

[0159] Step S349: Evaluate the layout adaptability of the node edge direction data using the view node ideal spacing data to generate view layout adaptability score data.

[0160] In an embodiment of the present invention, according to the view projection transformation data generated in step S31, the connection relationship between all nodes is extracted, and the view node adjacency matrix is ​​constructed. For example, if there is a connection relationship between node A and node B, then in the adjacency matrix, the element value corresponding to row A, column B and row B, column A is 1, otherwise it is 0. This matrix is ​​used to represent the topological relationship between nodes in the view. Using the view node adjacency matrix constructed in step S341, the visualization layout constraint space data constructed in step S33 is spatially recursively segmented and laid out. For example, a quadtree algorithm can be used to recursively segment the visualization layout constraint space into four equal subspaces until each subspace contains at most one node. Each node is assigned to the corresponding subspace to generate view segmentation space domain data. According to the view segmentation space domain data generated in step S342, the area of ​​each subspace and the sum of the areas of all subspaces are calculated, and divided by the total area of ​​the visualization layout constraint space to obtain screen space utilization data. Using the view node adjacency matrix constructed in step S341, the degree centrality, betweenness centrality and other indicators of each node are calculated as node importance data. For example, the higher the degree centrality of a node, the more nodes it connects and the higher its importance. According to the screen space utilization data calculated in step S343 and the node importance data calculated in step S344, the ideal node spacing of the view node adjacency matrix is ​​calculated. For example, an ideal spacing value can be calculated for each node pair according to the screen space utilization and the node importance. The larger the spacing value, the greater the space interval required between the nodes. The calculated ideal node spacing is used to construct the view node ideal spacing matrix data. Through the view node adjacency matrix, the calculated ideal view node spacing data is mapped to the actual screen coordinate system, and specific layout coordinates are assigned to each node to generate view node layout coordinate data, thereby ensuring that the ideal distance between nodes is maintained while meeting the physical limitations of the screen space. Path planning is performed on the connection relationship between nodes to generate view node connection line data. For example, adjacent nodes can be connected by a straight line path, a broken line path, or a curved path. For example, the Bentley-Ottmann algorithm can be used to detect line segment intersections. For intersecting edge lines, Bezier curves can be used to optimize the edge line direction. For example, the intersecting line segments can be replaced with smooth Bezier curves to avoid segment intersection. Using the node edge line direction data generated in step S347, the view node layout coordinate data generated in step S345 is fine-tuned to avoid overlapping nodes and edge lines, and the initial view layout data is generated. For example, the node coordinates can be fine-tuned according to the control points of the Bezier curve to ensure that the nodes do not overlap with the curve. Based on the ideal spacing data of the view nodes calculated in step S344 and the node edge line direction data generated in step S347, the layout is evaluated for fitness and view layout fitness score data is generated.For example, the degree of deviation between the actual spacing and the ideal spacing between nodes, as well as indicators such as the number of edge crossings and the degree of curvature, can be calculated as a basis for evaluating the fitness of the layout.

[0161] Preferably, step S4 comprises the following steps:

[0162] Step S41: performing view function structure analysis on the high-quality view layout solution data to generate view structure analysis data;

[0163] Step S42: performing visual element index processing according to the view structure parsing data to generate visual element index data;

[0164] Step S43: using the visual element index data to match the view components of the preset visualization component library to generate visualization component data;

[0165] Step S44: performing component binding processing on the high-quality view layout solution data through the visualization component data to obtain production process view component data;

[0166] Step S45: using the industrial Internet platform to collect real-time production process data and generate real-time production process data;

[0167] Step S46: Perform feature engineering processing according to the real-time production process data to generate standard production process data;

[0168] Step S47: constructing an interactive data injection channel based on the production process view component data and the standard production process data;

[0169] Step S48: Based on the interactive data injection channel, the standard production process data is injected into the production process view component data in real time to generate visual production process data.

[0170] As an example of the present invention, refer to Figure 3 As shown, Figure 1 Detailed implementation steps of step S4 in the flowchart, in this example, step S4 includes:

[0171] Step S41: performing view function structure analysis on the high-quality view layout solution data to generate view structure analysis data;

[0172] In the embodiment of the present invention, the high-quality view layout solution data outputted in step S37 is analyzed for information such as node types, connection relationships, and hierarchical structures, and the functional structure of the view is parsed, for example, identifying which nodes represent production workshops, which nodes represent production lines, which nodes represent equipment, etc. The parsing results are expressed in the form of functional semantics.

[0173] Step S42: performing visual element index processing according to the view structure parsing data to generate visual element index data;

[0174] In the embodiment of the present invention, each functional element in the view is indexed to generate visual element index data. For example, a unique ID can be used to identify each functional element, and its type, name, location, size and other information can be recorded. For example, the "stamping shop" node can be identified as "node_01", the type is "workshop", the name is "stamping shop", the location is (100, 100), and the size is (200, 150).

[0175] Step S43: using the visual element index data to match the view components of the preset visualization component library to generate visualization component data;

[0176] In the embodiment of the present invention, a visualization component library is pre-built, which contains various types of visualization components, such as charts, maps, dashboards, icons, etc. Each component defines its functions, data interfaces, style attributes, and other information. The visualization component library is matched using visual element index data, and a suitable visualization component is selected for each functional element to generate visualization component data. For example, a "workshop" type of functional element can be matched to a "map" component, a "production line" type of functional element can be matched to a "flowchart" component, and a "equipment" type of functional element can be matched to an "icon" component.

[0177] Step S44: performing component binding processing on the high-quality view layout solution data through the visualization component data to obtain production process view component data;

[0178] In the embodiment of the present invention, each visualization component is instantiated and bound to the corresponding functional element in the high-quality view layout solution data output in step S37 to generate production process view component data. For example, the "map" component is instantiated and bound to the "stamping shop" node, and the location, size and other information of the node are passed to the map component for setting the display area and size of the map.

[0179] Step S45: using the industrial Internet platform to collect real-time production process data and generate real-time production process data;

[0180] In the embodiment of the present invention, various sensors, controllers and other equipment deployed at the factory production site are used to collect production process data in real time through the industrial Internet platform, such as equipment status data, product output data, energy consumption data, etc., to generate real-time production process data. For example, the operating status, temperature, pressure and other data of the equipment can be read from the PLC through the OPC UA protocol.

[0181] Step S46: Perform feature engineering processing according to the real-time production process data to generate standard production process data;

[0182] In the embodiment of the present invention, the real-time production process data collected in step S45 is cleaned, converted, and feature extracted to generate standard production process data that conforms to the visualization component data interface specification. For example, the equipment status data can be converted into status codes such as "running", "stopped", and "fault", and the product output data can be statistically summarized to calculate indicators such as output and yield rate in each time period.

[0183] Step S47: constructing an interactive data injection channel based on the production process view component data and the standard production process data;

[0184] In the embodiment of the present invention, an interactive data injection channel is constructed based on the production process view component data generated in step S44 and the standard production process data generated in step S46. For example, a real-time data transmission channel can be established using WebSocket technology to push the standard production process data to the corresponding visualization component in real time.

[0185] Step S48: Based on the interactive data injection channel, the standard production process data is injected into the production process view component data in real time to generate visual production process data.

[0186] In the embodiment of the present invention, the interactive data injection channel constructed in step S47 is used to inject the standard production process data generated in step S46 into the production process view component data generated in step S44 in real time. For example, the real-time production data of the "stamping workshop" node is injected into the "map" component bound to it, and the map component dynamically updates the production display on the map according to the real-time data. After the real-time data of all components are updated, the final visualized production process data is generated and displayed on the terminal device.

[0187] Preferably, step S5 comprises the following steps:

[0188] Step S51: using the terminal device to perform visual display processing on the visualized production process data, and collecting style demand feedback from users to generate user demand feedback data;

[0189] Step S52: performing style preference portrait analysis based on user demand feedback data to generate style preference portrait data;

[0190] Step S53: performing style template customization processing according to the style preference portrait data to generate customized style template data;

[0191] Step S54: performing customized style dynamic rendering processing on the visualized production process data through the customized style template data, thereby obtaining the visualized production process customized instance data.

[0192] In an embodiment of the present invention, the visualized production process data is rendered into an interactive web page for display using the browser of the user's terminal device (such as a computer, tablet, or mobile phone), for example, using visualization libraries such as D3.js and ECharts to render the data into charts, maps, and the like. At the same time, a user feedback function is provided on the page, for example, the user can express his or her preferences for the current visualization effect by clicking a button, a drop-down menu, a color selector, and the like, such as preferences in terms of color theme, layout style, and chart type. The user's feedback information is recorded to generate user demand feedback data. The user's style preference is analyzed in combination with a pre-established style preference portrait model. For example, the user's feedback information can be quantified into a series of features, for example, color preferences can be represented by RGB values, layout styles can be represented by indicators such as compactness and symmetry, and chart types can be represented by categories such as bar charts, line charts, and pie charts. Then, a machine learning algorithm (such as a clustering algorithm or a classification algorithm) is used to analyze the user's features, identify the style category to which it belongs, such as "technical style", "business style", "simple style", and the like, and generate style preference portrait data. A set of style template libraries is prepared in advance, which contains visualization templates of various styles, such as "tech style" templates, "business style" templates, "simple style" templates, etc. Each template defines a complete set of visualization style rules, such as color schemes, font styles, layout methods, chart types, etc. According to the style preference portrait data generated in step S52, a style template that best matches the user's preference is selected from the style template library, and the template is fine-tuned according to the user's specific feedback information. For example, the color scheme in the template can be adjusted according to the color theme selected by the user; the layout method in the template can be adjusted according to the layout style selected by the user, and finally the customized style template data is generated. The customized style template data generated in step S53 is used to dynamically render the visualized production process data generated in step S48 in a customized style. For example, JavaScript code is used to read the style rules in the template and apply them to the visualization component, and the color, font, layout, chart type and other properties of the component are dynamically modified, and finally the customized instance data of the visualized production process is obtained and displayed on the terminal device.

[0193] Preferably, the present invention further provides a visual customization system for industrial Internet applications, which executes the visual customization method for industrial Internet applications as described above, and the visual customization system for industrial Internet applications comprises:

[0194] The process element parsing module is used to identify the process element entities of the target production process document data and generate process element relationship network data; process metadata aggregation is performed based on the process element relationship network data to obtain industrial production process metadata;

[0195] The multi-granularity view construction module is used to perform multi-granularity hierarchical abstract processing based on industrial production process metadata to generate multi-level granularity mapping matrix data; construct a fractal dimension tree based on the multi-level granularity mapping matrix data to obtain preliminary view abstract tree data; perform view granularity adaptive adjustment based on the preliminary view abstract tree data to generate dynamic view abstract tree data;

[0196] The view layout optimization module is used to perform spatial view layout processing according to the dynamic view abstract tree data to generate initial view layout data; perform layout adaptability optimization on the initial view layout data to generate high-quality view layout solution data;

[0197] The view component rendering module is used to match the high-quality view layout scheme data with view components to generate visual component data; perform component binding processing on the high-quality view layout scheme data through the visual component data, and perform real-time data injection to generate visual production process data;

[0198] The user style customization module is used to perform user style customization processing on the visualized production process data and generate customized style template data; perform customized style dynamic rendering processing according to the customized style template data to obtain the visualized production process customized instance data.

[0199] Preferably, a visual customization storage medium for industrial Internet applications stores a computer program, and when the computer program is executed, it implements a visual customization method for industrial Internet applications as described in any one of the above.

[0200] The beneficial effect of this application is that a highly flexible and detailed process model is constructed through in-depth analysis of production process documents, semantic role labeling and multi-level granular abstract processing. Then, topological decomposition, functional decomposition and granularity mapping techniques are used to generate a view structure that adapts to different perspectives, ensuring the rationality of the layout and the clear communication of information. On this basis, dynamic view abstract tree construction and optimization are implemented, and adaptive adjustment of views is achieved in combination with user interaction strategies. Through spatial layout optimization, aesthetic scoring and adaptability scoring, the view is ensured to be both beautiful and practical. Furthermore, using a data-driven approach, intelligent matching of view components and injection of real-time data, as well as personalized style customization based on user feedback are achieved, and a highly customized visual production process is finally generated, which not only improves the transparency and management efficiency of the production process, but also greatly enhances the user experience and meets the diversified and personalized visualization needs in the industrial Internet environment.

[0201] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0202] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A visual customization method for industrial Internet applications, characterized in that: The following steps are involved: Step 1: Perform process element entity recognition on the target production process document data to generate process element relationship network data; Process metadata is aggregated based on process element relationship network data to obtain industrial production process metadata; Step S2: Perform multi-granularity hierarchical abstract processing based on industrial production process metadata to generate multi-level granularity mapping matrix data; Construct a fractal dimension tree based on multi-level granularity mapping matrix data to obtain preliminary view abstract tree data; Adaptively adjust the view granularity according to the preliminary view abstract tree data to generate dynamic view abstract tree data; wherein step S2 is specifically as follows: Step S21: performing production process topology decomposition based on industrial production process metadata to generate production process topology data; Step S22: Perform multi-granularity hierarchical abstract processing according to the production process topology data to generate multi-level granularity mapping matrix data; wherein step S22 is specifically as follows: Step S221: labeling entity roles according to production process topology data to generate process role labeling data; Step S222: performing workshop-level functional process decomposition on the production process topology data by using the process role labeling data to obtain workshop-level functional decomposition data; Step S223: Performing production line-level operation link analysis according to the workshop-level function decomposition data to generate production line-level operation link data; Step S224: identifying workstation-level activity nodes according to the production line-level operation link data, and generating workstation-level activity node data; Step S225: extracting equipment-level configuration based on the workshop-level function decomposition data, the production line-level operation link data, and the workstation-level activity node data to obtain equipment-level configuration parameters; Step S226: constructing a granularity mapping relationship matrix according to the workshop-level function decomposition data, the production line-level operation link data, the workstation-level activity node data, and the equipment-level configuration parameters to generate multi-level granularity mapping matrix data; Step S23: performing hierarchical attribute weight allocation according to the multi-level granularity mapping matrix data, and calculating the correlation strength, to generate multi-level process granularity correlation strength data; Step S24: Based on the preset process association threshold, the multi-level process granularity association strength data is used to screen the multi-level granularity mapping matrix data for strongly associated process granularity, and view node generation is performed to generate view node data; Step S25: constructing a fractal dimension tree according to the view node data to obtain preliminary view abstract tree data; Step S26: performing a view tree stability assessment based on the preliminary view abstract tree data to generate view tree stability assessment data, performing view node optimization based on the view tree stability assessment data to generate an optimized view abstract tree; Step S27: Adaptively adjust the view granularity of the optimized view abstract tree through a preset user interaction operation strategy to generate dynamic view abstract tree data, wherein the preset user interaction operation strategy includes a view zoom operation, a view filtering operation, a view drill-down operation, and a view selection operation; Step S3: performing spatial view layout processing according to the dynamic view abstract tree data to generate initial view layout data; performing layout adaptability optimization on the initial view layout data to generate high-quality view layout solution data; Step S4: performing view component matching on the high-quality view layout scheme data to generate visualization component data; performing component binding processing on the high-quality view layout scheme data through the visualization component data, and performing real-time data injection to generate visualization production process data; Step S5: Perform user style customization processing on the visualized production process data to generate customized style template data; perform customized style dynamic rendering processing according to the customized style template data to obtain customized instance data of the visualized production process.

2. The visual customization method for industrial Internet applications according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: performing deep semantic analysis on the target production process document data to generate production process semantic node data; Step S12: semantic role labeling is performed on the production process semantic node data using preset process semantic role labeling rule data to obtain process role labeling data; Step S13: Perform process element entity recognition based on process role labeling data, and perform context semantic relationship mining to generate process element relationship network data; Step S14: Performing temporal process logic link analysis based on process element relationship network data to generate process flow link data; Step S15: Perform process metadata aggregation processing based on the process flow link data to obtain industrial production process metadata.

3. The visual customization method for industrial Internet applications according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing tree structure traversal processing on the dynamic view abstract tree data, and performing view topology projection transformation processing to generate view projection transformation data; Step S32: using the terminal device to obtain screen size limitation parameters; Step S33: constructing a constraint space for the view projection transformation data using screen size restriction parameters to generate visualization layout constraint space data; Step S34: performing spatial view layout processing on the view projection transformation data using the visualization layout constraint space data, and generating initial view layout data and view layout adaptability score data respectively; Step S35: optimizing the layout of the initial view layout data using preset aesthetic rule data, and performing visual aesthetic scoring, thereby generating view layout scoring data and optimized view layout data; Step S36: performing comprehensive weighted scoring processing according to the view layout scoring data and the view layout adaptability scoring data to generate view layout comprehensive scoring data; Step S37: Perform view scheme scoring judgment based on the view layout comprehensive scoring data. When the view layout comprehensive scoring data is higher than or equal to the preset layout scoring threshold, mark the optimized view layout data as high-quality view layout scheme data; when the view layout comprehensive scoring data is lower than the preset layout scoring threshold, mark the optimized view layout data as low-quality view layout scheme data, eliminate the low-quality view layout scheme data, and return to step S34 to readjust the view layout.

4. The visual customization method for industrial Internet applications according to claim 3 is characterized in that: Step S34 includes the following steps: Step S341: constructing a view node adjacency matrix according to the view projection transformation data; Step S342: recursively partition and layout the visualization layout constraint space data through the view node adjacency matrix to generate view partition space domain data; Step S343: Calculate screen space utilization according to the view segmentation space domain data to generate screen space utilization data; Step S344: Calculate the node importance using the view node adjacency matrix to generate node importance data; calculate the ideal node spacing of the view node adjacency matrix using the screen space utilization data and the node importance data to generate the ideal node spacing data; Step S345: performing node coordinate mapping processing on the view node ideal spacing data through the view node adjacency matrix to generate view node layout coordinate data; Step S346: performing connection path planning processing according to the view node layout coordinate data to generate view node connection line data; Step S347: performing edge line intersection detection on the view node connection line data, and performing edge line trend optimization processing through Bezier curves to generate node edge line trend data; Step S348: adjusting the view node layout coordinate data using the node edge direction data to generate initial view layout data; Step S349: Evaluate the layout adaptability of the node edge direction data using the view node ideal spacing data to generate view layout adaptability score data.

5. The visual customization method for industrial Internet applications according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: performing view function structure analysis on the high-quality view layout solution data to generate view structure analysis data; Step S42: performing visual element index processing according to the view structure parsing data to generate visual element index data; Step S43: using the visual element index data to match the view components of the preset visualization component library to generate visualization component data; Step S44: performing component binding processing on the high-quality view layout solution data through the visualization component data to obtain production process view component data; Step S45: using the industrial Internet platform to collect real-time production process data and generate real-time production process data; Step S46: Perform feature engineering processing according to the real-time production process data to generate standard production process data; Step S47: constructing an interactive data injection channel based on the production process view component data and the standard production process data; Step S48: Based on the interactive data injection channel, the standard production process data is injected into the production process view component data in real time to generate visualized production process data.

6. The visual customization method for industrial Internet applications according to claim 1, characterized in that: Step S5 includes the following steps: Step S51: using the terminal device to perform visual display processing on the visualized production process data, and collecting style demand feedback from users to generate user demand feedback data; Step S52: performing style preference portrait analysis based on user demand feedback data to generate style preference portrait data; Step S53: performing style template customization processing according to the style preference portrait data to generate customized style template data; Step S54: performing customized style dynamic rendering processing on the visualized production process data through the customized style template data, thereby obtaining the visualized production process customized instance data.

7. A visual customization system for industrial Internet applications, characterized in that: For executing the visual customization method for industrial Internet applications according to claim 1, the visual customization system for industrial Internet applications comprises: The process element parsing module is used to identify the process element entities of the target production process document data and generate process element relationship network data; process metadata aggregation is performed based on the process element relationship network data to obtain industrial production process metadata; The multi-granularity view construction module is used to perform multi-granularity hierarchical abstract processing based on industrial production process metadata to generate multi-level granularity mapping matrix data; construct a fractal dimension tree based on the multi-level granularity mapping matrix data to obtain preliminary view abstract tree data; perform view granularity adaptive adjustment based on the preliminary view abstract tree data to generate dynamic view abstract tree data; The view layout optimization module is used to perform spatial view layout processing according to the dynamic view abstract tree data to generate initial view layout data; perform layout adaptability optimization on the initial view layout data to generate high-quality view layout solution data; The view component rendering module is used to match the high-quality view layout scheme data with view components to generate visual component data; perform component binding processing on the high-quality view layout scheme data through the visual component data, and perform real-time data injection to generate visual production process data; The user style customization module is used to perform user style customization processing on the visualized production process data and generate customized style template data; perform customized style dynamic rendering processing according to the customized style template data to obtain the visualized production process customized instance data.

8. A visual customized storage medium for industrial Internet applications, storing a computer program, characterized in that: When the computer program is executed by a processor, it implements a visual customization method for industrial Internet applications as described in any one of claims 1 to 6.

Images