Electronic table data processing method and system based on computer vision technology

Abstract

The present invention relates to the field of computer vision technology, and in particular, to a method and system for processing electronic table data based on computer vision technology. The method comprises the following steps: using computer vision technology to identify the structure of a table, using dynamic relationship modeling technology to establish a dynamic data relationship model between cells, and generating a data dependency graph; according to the data dependency graphs of different tables, using a cross-table reference intelligent analysis algorithm to detect and aggregate relevant data of multiple tables, monitoring the reference relationship between different tables to correct data errors; using complex relationship visualization and processing technology to construct a data flow graph; using topology and Dijkstra algorithm to analyze data dependency graphs and data flow graphs. The electronic table data processing method and system based on computer vision technology automatically identify and model the data relationship in the table through the combination of computer vision technology and intelligent analysis algorithm, and detect and correct reference errors between tables.

Description

Technical Field

[0001] The present invention relates to the field of computer vision technology, and in particular to an electronic table data processing method and system based on computer vision technology. Background Art

[0002] The electronic spreadsheet data processing method and system based on computer vision technology are designed to automatically identify complex table structures and efficiently process cross-table data dependencies. Through dynamic relationship modeling technology and cross-table reference intelligent analysis algorithms, the intelligent aggregation and error correction of table data are controlled to achieve accurate data dependency analysis, cross-table data integration and efficient data processing automation.

[0003] Existing spreadsheet data processing methods and systems are usually difficult to automatically capture complex cell dependencies, and due to the complex table structure and intricate reference relationships between tables, it will lead to inaccurate cross-table data aggregation, reference errors are difficult to automatically correct, and data analysis efficiency is low. Therefore, a spreadsheet data processing method and system based on computer vision technology are provided. Summary of the invention

[0004] The purpose of the present invention is to provide a spreadsheet data processing method and system based on computer vision technology to solve the problems raised in the above background technology, namely, inaccurate cross-table data aggregation, difficulty in automatically correcting reference errors, and low data analysis efficiency due to the complex table structure and intricate reference relationships between tables.

[0005] To achieve the above object, the present invention provides a method for processing electronic table data based on computer vision technology, comprising:

[0006] S1. Use computer vision technology to identify the structure of the table, use dynamic relationship modeling technology to establish a dynamic data relationship model between cells, and generate a data dependency graph;

[0007] S2. Based on the data dependency graphs of different tables, use the cross-table reference intelligent analysis algorithm to detect and aggregate the relevant data of multiple tables, monitor the reference relationships between different tables and correct data errors;

[0008] S3, construct data flow graphs using complex relationship visualization and processing techniques;

[0009] S4. Use topology and Dijkstra algorithm to analyze data dependency graphs and data flow graphs.

[0010] As a further improvement of the technical solution, in S1, the structure of the table is identified using computer vision technology, and the specific steps are as follows:

[0011] S1.1.1. Obtain image data of the electronic form by a scanner;

[0012] S1.1.2, use the Canny algorithm to identify the row and column borders of the table, use the Hough transform to identify the straight lines of the table, and determine the intersection points of rows and columns;

[0013] S1.1.3. After identifying the row-column framework, divide each cell according to the intersection to form a cell structure, use optical character recognition to extract the text in the cell, identify the row and column headers, and form an initial data set :

[0014] ;

[0015] ;

[0016] ;

[0017] in, is the row header collection; is a collection of column headers; is the initial data set, containing all combinations of row and column headers.

[0018] As a further improvement of the technical solution, the dynamic relationship modeling technology is implemented based on computer vision technology and data dependency analysis, and is used to identify and establish dynamic data relationships between cells in a spreadsheet;

[0019] In S1, a dynamic relationship modeling technology is used to establish a dynamic data relationship model between cells and generate a data dependency graph. The specific steps are as follows:

[0020] S1.2.1. Using the adjacency matrix To represent the relationship between cells:

[0021] ;

[0022] S1.2.2. Define a value vector :

[0023] ;

[0024] in, Indicates the total number of cells; Represents a value vector Middle Cells The value of

[0025] S1.2.3. Use dynamic relationship modeling techniques combined with adjacency matrix Sum value vector To build a dynamic data relationship model:

[0026] ;

[0027] in, For cells The dynamic output value of is the adjacency matrix Elements in For cells The bias term of Indicates the total number of cells; Represents a value vector Middle Cells The value of

[0028] S1.2.4, define the node set of the data dependency graph as :

[0029] ;

[0030] if ≠0, then add a slave node to the graph To Node The directed edges of :

[0031] ;

[0032] S1.2.5. For each directed edge , the weight is , using graph data structures to build data dependency graphs , and use the adjacency matrix to store the structure of the data dependency graph.

[0033] As a further improvement of the technical solution, the cross-table reference intelligent analysis algorithm is implemented based on the Jaccard similarity method and aggregation operation, and is used to automatically detect and integrate data of similar column headers in a spreadsheet;

[0034] In S2, according to the data dependency graph of different tables, a cross-table reference intelligent analysis algorithm is used to detect and aggregate the relevant data of multiple tables, monitor the reference relationship between different tables and correct data errors. The specific steps are as follows:

[0035] S2.1. Defining Similarity Metrics Based on Jaccard Similarity ;

[0036] S2.2. Setting similarity threshold , generating a collection of similar column headers ;

[0037] S2.3, according to and data dependency graph The reference relationship determines the aggregated data source;

[0038] S2.4. Using Data Dependency Graph The edges in the grid are used to detect the reference relationship between cells;

[0039] S2.5. After detecting that the cell reference relationship is incorrect, use the data dependency diagram Automatically correct data errors.

[0040] As a further improvement of the technical solution, in S2.1, the similarity measure is defined based on Jaccard similarity , the specific method is as follows:

[0041] ;

[0042] in, For column headers and column headers The number of elements in the intersection of For column headers and column headers The number of elements in the union of ;

[0043] In S2.2, a similarity threshold is set. , generating a collection of similar column headers , the specific method is as follows:

[0044] ;

[0045] In S2.3, according to and data dependency graph The reference relationship is used to determine the aggregated data source. The specific method is as follows:

[0046] ;

[0047] in, is an aggregated dataset; Represents an aggregation operation; For column headers The corresponding dataset; For column headers The corresponding dataset;

[0048] In S2.4, a data dependency graph is used The edges in the grid are used to detect the reference relationship between cells. The specific method is as follows:

[0049] ;

[0050] in, For cells Whether to reference a cell indicator variable of

[0051] S2.5: Using the data dependency graph Automatically correct data errors. The specific method is as follows:

[0052] ;

[0053] in, represents the revised data set; For cells Whether to reference a cell indicator variable of For cells Correction value of Represents the cell The total number of related cells.

[0054] As a further improvement of the present technical solution, the complex relationship visualization and processing technology is implemented based on a data dependency analysis method and a visualization algorithm, and is used to construct a data flow diagram.

[0055] S3. Use complex relationship visualization and processing technology to build a data flow graph. The specific steps are as follows:

[0056] S3.1. Define the data flow graph node set as the set is , introducing the state vector of each node ;

[0057] S3.2, define the edge set of the data flow graph as , introduce a weight function ;

[0058] S3.3. Building dynamic data flow functions ;

[0059] S3.4. Constructing the adjacency matrix of the data flow graph ;

[0060] ;

[0061] S3.5, the combined node set is , the edge set is and dynamic data flow functions , construct the mathematical model expression of the data flow graph.

[0062] As a further improvement of the present technical solution, in S3.1, the data flow graph node set is defined as , introduce the The state vector of a node , the specific method is as follows:

[0063] ;

[0064] ;

[0065] in, Represents all nodes in the data flow graph; For the The state vector of each node; Indicates The state vector of a node Dimensions; For the Node Dimensional data characteristics;

[0066] In S3.2, the edge set of the data flow graph is defined as , introduce a weight function , the specific method is as follows:

[0067] ;

[0068] ;

[0069] in, is the weight function; is the weight function between the calculation nodes; For the The state vector of each node;

[0070] In S3.3, construct a dynamic data flow function , the specific method is as follows:

[0071] ;

[0072] in, According to the weight function and state vector , calculate from all nodes Related nodes right The impact of is the activation function.

[0073] As a further improvement of the technical solution, in S3.5, the combined node set is , the edge set is and dynamic data flow functions , construct the mathematical model expression of the data flow graph, the specific method is as follows:

[0074] ;

[0075] in, Indicates the data stream output result; is the initial spreadsheet data; is the number of nodes to be processed in the data flow graph.

[0076] As a further improvement of the technical solution, in S4, topology and Dijkstra algorithm are used to analyze the data dependency graph and the data flow graph, and the specific steps are as follows:

[0077] S4.1. Analyze the topological structure of the data dependency graph and data flow graph, and calculate the in-degree and out-degree of the nodes;

[0078] S4.2. Use the Dijkstra algorithm to analyze the critical paths in the data dependency graph and data flow graph and identify the main routes for data flow.

[0079] On the other hand, the present invention provides an electronic spreadsheet data processing system based on computer vision technology, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-mentioned electronic spreadsheet data processing method based on computer vision technology.

[0080] Compared with the prior art, the present invention has the following beneficial effects:

[0081] 1. In the electronic spreadsheet data processing method and system based on computer vision technology, the table structure recognition and dynamic relationship modeling technology based on computer vision technology can automatically recognize the complex structure of the table, extract row and column headers, and establish a dynamic data relationship model between cells, and generate a data dependency diagram, thereby accurately capturing the data dependency and reference relationship in the table.

[0082] 2. In the electronic spreadsheet data processing method and system based on computer vision technology, through the cross-table reference intelligent analysis algorithm and Jaccard similarity analysis, automatic detection of similar column headers in multiple tables, intelligent aggregation of data, and automatic correction of data errors in cross-table reference relationships are achieved, thereby ensuring data consistency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] Figure 1 The figure is a flow chart of the overall method of the present invention. DETAILED DESCRIPTION

[0084] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0085] See also Figure 1 As shown, this embodiment provides a method for processing electronic table data based on computer vision technology, comprising the following steps:

[0086] S1. Use computer vision technology to identify the structure of the table, use dynamic relationship modeling technology to establish a dynamic data relationship model between cells, and generate a data dependency graph;

[0087] In S1, the structure of the table is identified using computer vision technology, and the specific steps are as follows:

[0088] S1.1.1. Obtain image data of the electronic form by a scanner;

[0089] S1.1.2, use the Canny algorithm to identify the row and column borders of the table, use the Hough transform to identify the straight lines of the table, and determine the intersection points of rows and columns;

[0090] S1.1.3. After identifying the row-column framework, divide each cell according to the intersection to form a cell structure, use optical character recognition to extract the text in the cell, identify the row and column headers, and form an initial data set :

[0091] ;

[0092] ;

[0093] ;

[0094] in, is the row header collection; is a collection of column headers; is the initial data set, containing all combinations of row and column headers.

[0095] The dynamic relationship modeling technology is implemented based on computer vision technology and data dependency analysis, and is used to identify and establish dynamic data relationships between cells in a spreadsheet;

[0096] In this embodiment S1, a dynamic relationship modeling technology is used to establish a dynamic data relationship model between cells and generate a data dependency graph. The specific steps are as follows:

[0097] S1.2.1. Using the adjacency matrix To represent the relationship between cells:

[0098] ;

[0099] S1.2.2. Define a value vector :

[0100] ;

[0101] in, Indicates the total number of cells; Represents a value vector Middle Cells The value of

[0102] S1.2.3. Use dynamic relationship modeling techniques combined with adjacency matrix Sum value vector To build a dynamic data relationship model:

[0103] ;

[0104] in, For cells The dynamic output value of is the adjacency matrix Elements in For cells The bias term of Indicates the total number of cells; Represents a value vector Middle Cells The value of

[0105] S1.2.4, define the node set of the data dependency graph as :

[0106] ;

[0107] if ≠0, then add a slave node to the graph To Node The directed edges of :

[0108] ;

[0109] S1.2.5. For each directed edge , the weight is , using graph data structures to build data dependency graphs , and use the adjacency matrix to store the structure of the data dependency graph.

[0110] S2. Based on the data dependency graphs of different tables, use the cross-table reference intelligent analysis algorithm to detect and aggregate the relevant data of multiple tables, monitor the reference relationships between different tables and correct data errors;

[0111] The cross-table reference intelligent analysis algorithm is implemented based on the Jaccard similarity method and aggregation operation, and is used to automatically detect and integrate data with similar column headers in a spreadsheet;

[0112] In this embodiment S2, according to the data dependency graph of different tables, a cross-table reference intelligent analysis algorithm is used to detect and aggregate the relevant data of multiple tables, monitor the reference relationship between different tables and correct data errors. The specific steps are as follows:

[0113] S2.1. Defining Similarity Metrics Based on Jaccard Similarity ;

[0114] S2.2. Setting similarity threshold , generating a collection of similar column headers ;

[0115] S2.3, according to and data dependency graph The reference relationship determines the aggregated data source;

[0116] S2.4. Using Data Dependency Graph The edges in the grid are used to detect the reference relationship between cells;

[0117] S2.5. After detecting that the cell reference relationship is incorrect, use the data dependency diagram Automatically correct data errors.

[0118] In this embodiment S2.1, the similarity metric is defined based on Jaccard similarity , the specific method is as follows:

[0119] ;

[0120] in, For column headers and column headers The number of elements in the intersection of For column headers and column headers The number of elements in the union of ;

[0121] In this embodiment S2.2, the similarity threshold is set , generating a collection of similar column headers , the specific method is as follows:

[0122] ;

[0123] In this embodiment S2.3, according to and data dependency graph The reference relationship is used to determine the aggregated data source. The specific method is as follows:

[0124] ;

[0125] in, is an aggregated dataset; Represents an aggregation operation; For column headers The corresponding dataset; For column headers The corresponding dataset;

[0126] In the data dependency graph, nodes represent cells and edges represent dependencies between cells. Through the structure of the graph, the reference relationship between each cell and other cells can be directly extracted.

[0127] In this embodiment S2.4, the data dependency graph is used The edges in the grid are used to detect the reference relationship between cells. The specific method is as follows:

[0128] ;

[0129] in, For cells Whether to reference a cell indicator variable of

[0130] In this embodiment S2.5, using the data dependency graph Automatically correct data errors. The specific method is as follows:

[0131] ;

[0132] in, represents the revised data set; For cells Whether to reference a cell indicator variable of For cells Correction value of Represents the cell The total number of related cells.

[0133] In this embodiment, The value of is the mean of historical data;

[0134] S3, construct data flow graphs using complex relationship visualization and processing techniques;

[0135] In this embodiment, the complex relationship visualization and processing technology is implemented based on the data dependency analysis method and visualization algorithm, and is used to construct a data flow graph;

[0136] S3. Use complex relationship visualization and processing technology to build a data flow graph. The specific steps are as follows:

[0137] S3.1, define the data flow graph node set as , introducing the state vector of each node ;

[0138] S3.2, define the edge set of the data flow graph as , introduce a weight function ;

[0139] S3.3. Building dynamic data flow functions ;

[0140] S3.4. Constructing the adjacency matrix of the data flow graph ;

[0141] ;

[0142] S3.5, the combined node set is , the edge set is and dynamic data flow functions , construct the mathematical model expression of the data flow graph.

[0143] In this embodiment S3.1, the data flow graph node set is defined as , introduce the The state vector of a node , the specific method is as follows:

[0144] ;

[0145] ;

[0146] in, Represents all nodes in the data flow graph; For the The state vector of each node; Indicates The state vector of a node Dimensions; For the Node Dimensional data characteristics;

[0147] The data flow diagram graphically displays each link in the data processing process, including data input, processing, output and error monitoring;

[0148] Nodes: represent the various steps of data processing, such as data input, cleaning, aggregation, and output.

[0149] Edge: represents the flow relationship of data between different nodes. Edges can contain weights to reflect the frequency or importance of data transmission.

[0150] In this embodiment S3.2, the edge set of the data flow graph is defined as , introduce a weight function , the specific method is as follows:

[0151] ;

[0152] ;

[0153] in, is the weight function; is the weight function between the calculation nodes; For the The state vector of each node;

[0154] is a function that includes various parameters of the node status, such as data volume, processing time, and data quality;

[0155] It is a slave node To Node The weight of Flow to The importance or strength of the node. The weight value depends on the and The state vector and The relationship between . Function Used to calculate weights, possibly considering similarities and dependencies between data features;

[0156] In this embodiment S3.3, a dynamic data flow function is constructed , the specific method is as follows:

[0157] ;

[0158] in, According to the weight function and state vector , calculate from all nodes Related nodes right The impact of is the activation function;

[0159] Data Stream Functions The output value of is determined by the weights and state vectors between the nodes; the choice of activation function can be adjusted according to the complexity of the data flow graph; if the data flow graph is relatively simple, a linear activation function can be used; if the data relationship is more complex, a nonlinear activation function can be selected.

[0160] In this embodiment S3.5, the combined node set is , the edge set is and dynamic data flow functions , construct the mathematical model expression of the data flow graph, the specific method is as follows:

[0161] ;

[0162] in, Indicates the data stream output result; is the initial spreadsheet data; is the number of nodes to be processed in the data flow graph;

[0163] The value of depends on the input data and the data processing steps of each node. is a vector or a matrix, depending on the type and dimension of the data. is the final result of the entire processing process, including the cleaned data, the aggregated data, and the calculated analysis results;

[0164] The initial spreadsheet data is a two-dimensional table containing mixed types of values, texts, and dates;

[0165] S4. Use topology and Dijkstra algorithm to analyze data dependency graphs and data flow graphs;

[0166] In this embodiment S4, topology and Dijkstra algorithm are used to analyze the data dependency graph and the data flow graph, and the specific steps are as follows:

[0167] S4.1. Analyze the topological structure of the data dependency graph and data flow graph, and calculate the in-degree and out-degree of the nodes;

[0168] S4.2. Use the Dijkstra algorithm to analyze the critical paths in the data dependency graph and data flow graph and identify the main routes for data flow.

[0169] This embodiment provides an electronic spreadsheet data processing system based on computer vision technology, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any one of the electronic spreadsheet data processing methods based on computer vision technology described above.

[0170] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments, and the above embodiments and descriptions are only preferred examples of the present invention, and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements all fall within the scope of the present invention to be protected.

Claims

1. A method for processing electronic spreadsheet data based on computer vision technology, characterized in that: The following steps are involved: S1. Use computer vision technology to identify the structure of the table, use dynamic relationship modeling technology to establish a dynamic data relationship model between cells, and generate a data dependency graph; S2. Based on the data dependency graphs of different tables, use the cross-table reference intelligent analysis algorithm to detect and aggregate the relevant data of multiple tables, monitor the reference relationships between different tables and correct data errors; S3, construct data flow graphs using complex relationship visualization and processing techniques; S4. Use topology and Dijkstra algorithm to analyze data dependency graphs and data flow graphs; The dynamic relationship modeling technology is implemented based on computer vision technology and data dependency analysis, and is used to identify and establish dynamic data relationships between cells in a spreadsheet; In S1, a dynamic relationship modeling technology is used to establish a dynamic data relationship model between cells and generate a data dependency graph. The specific steps are as follows: S1.2.

1. Using the adjacency matrix To represent the relationship between cells: ； S1.2.

2. Define a value vector : ； in, Indicates the total number of cells; Represents a value vector Middle Cells The value of S1.2.

3. Use dynamic relationship modeling techniques combined with adjacency matrix Sum value vector To build a dynamic data relationship model: ； in, For cells The dynamic output value of is the adjacency matrix Elements in For cells The bias term of Indicates the total number of cells; Represents a value vector Middle Cells The value of S1.2.4, define the node set of the data dependency graph as ; ； if , then add a slave node to the graph To Node The directed edges of : ； S1.2.

5. For each directed edge , the weight is , using graph data structures to build data dependency graphs , and use the adjacency matrix to store the structure of the data dependency graph; The cross-table reference intelligent analysis algorithm is implemented based on the Jaccard similarity method and aggregation operation, and is used to automatically detect and integrate data with similar column headers in a spreadsheet; In S2, according to the data dependency graph of different tables, a cross-table reference intelligent analysis algorithm is used to detect and aggregate the relevant data of multiple tables, monitor the reference relationship between different tables and correct data errors. The specific steps are as follows: S2.

1. Defining Similarity Metrics Based on Jaccard Similarity ; S2.

2. Setting similarity threshold , generating a collection of similar column headers ; S2.3, according to and data dependency graph The reference relationship determines the aggregated data source; S2.

4. Using Data Dependency Graph The edges in the grid are used to detect the reference relationship between cells; S2.

5. After detecting that the cell reference relationship is incorrect, use the data dependency diagram Automatically correct data errors.

2. The electronic table data processing method based on computer vision technology according to claim 1, characterized in that: In S1, the structure of the table is identified using computer vision technology, and the specific steps are as follows: S1.1.

1. Obtain image data of the electronic form by a scanner; S1.1.2, use the Canny algorithm to identify the row and column borders of the table, use the Hough transform to identify the straight lines of the table, and determine the intersection points of rows and columns; S1.1.

3. After identifying the row-column framework, divide each cell according to the intersection to form a cell structure, use optical character recognition to extract the text in the cell, identify the row and column headers, and form an initial data set : ； ； ； in, is the row header collection; is a collection of column headers; is the initial data set, containing all combinations of row and column headers.

3. The electronic table data processing method based on computer vision technology according to claim 2 is characterized in that: In S2.1, the similarity measure is defined based on Jaccard similarity. , the specific method is as follows: ; in, For column headers and column headers The number of elements in the intersection of For column headers and column headers The number of elements in the union of ; In S2.2, a similarity threshold is set. , generating a collection of similar column headers , the specific method is as follows: ； In S2.3, according to and data dependency graph The reference relationship is used to determine the aggregated data source. The specific method is as follows: ； in, is an aggregated dataset; Represents an aggregation operation; For column headers The corresponding dataset; For column headers The corresponding dataset; In S2.4, a data dependency graph is used The edges in the grid are used to detect the reference relationship between cells. The specific method is as follows: ； in, For cells Whether to reference cells indicator variable of S2.5: Using the data dependency graph Automatically correct data errors. The specific method is as follows: ； in, represents the revised data set; For cells Whether to reference cells indicator variable of For cells Correction value of Represents the cell The total number of related cells.

4. The electronic table data processing method based on computer vision technology according to claim 3 is characterized in that: The complex relationship visualization and processing technology is implemented based on the data dependency analysis method and visualization algorithm, and is used to construct a data flow graph; S3. Use complex relationship visualization and processing technology to build a data flow graph. The specific steps are as follows: S3.

1. Define the data flow graph node set as , introducing the state vector of each node ; S3.2, define the edge set of the data flow graph as , introduce a weight function ; S3.

3. Building dynamic data flow functions ; S3.

4. Constructing the adjacency matrix of the data flow graph ; ； S3.5, the combined node set is , the edge set is and dynamic data flow functions , construct the mathematical model expression of the data flow graph.

5. The electronic table data processing method based on computer vision technology according to claim 4 is characterized in that: In S3.1, the data flow graph node set is defined as , introduce the The state vector of a node , the specific method is as follows: ； ； in, Represents all nodes in the data flow graph; For the The state vector of each node; Indicates The state vector of a node Dimensions; For the Node Dimensional data characteristics; In S3.2, the edge set of the data flow graph is defined as , introduce a weight function , the specific method is as follows: ； ;in, is the weight function; is the weight function between the calculation nodes; For the The state vector of each node; In S3.3, construct a dynamic data flow function , the specific method is as follows: ； in, According to the weight function and state vector , calculate from all nodes Related nodes right The impact of is the activation function.

6. The electronic table data processing method based on computer vision technology according to claim 5 is characterized in that: In S3.5, the combined node set is , the edge set is and dynamic data flow functions , construct the mathematical model expression of the data flow graph, the specific method is as follows: ； in, Indicates the data stream output result; is the initial spreadsheet data; is the number of nodes to be processed in the data flow graph.

7. The electronic table data processing method based on computer vision technology according to claim 6 is characterized in that: In S4, topology and Dijkstra algorithm are used to analyze the data dependency graph and data flow graph. The specific steps are as follows: S4.

1. Analyze the topological structure of the data dependency graph and data flow graph, and calculate the in-degree and out-degree of the nodes; S4.

2. Use the Dijkstra algorithm to analyze the critical paths in the data dependency graph and data flow graph and identify the main routes for data flow.

8. An electronic spreadsheet data processing system based on computer vision technology, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: The processor executes the computer program to implement the electronic form data processing method based on computer vision technology as described in any one of claims 1 to 7.

Images