Intelligent electric power financial bill management device

Abstract

The invention provides an intelligent electric power financial bill management device, and aims to solve the problems of low recognition accuracy, low processing speed, tedious auditing process and the like when an existing electric power financial bill management system processes bills. The device comprises an electronic bill uploading module, a preprocessing module, an information extraction module, an auditing module, a classification and filing module, a multi-dimensional query module and a statistics and analysis module. Through an automatic and intelligent mode, comprehensive, rapid and accurate processing of electronic bills is realized, functions of bill uploading, preprocessing, key information extraction, auditing verification, classification filing, multi-dimensional query, statistical analysis and the like are included, and efficient and convenient financial management services are provided for power enterprises.

Description

Technical Field

[0001] The present invention relates to the technical field, and in particular to an intelligent power financial bill management device. Background Art

[0002] The intelligent power financial bill management system is a financial management software designed specifically for power companies. It processes the financial bills of power companies in an automated and intelligent way, including electricity bills, receipts, payment vouchers, etc. The system combines advanced technologies such as big data and artificial intelligence to provide comprehensive, professional and intelligent financial management services for power companies.

[0003] The intelligent power financial bill management system is a financial management software designed for power companies. It plays a vital role in the daily operations of power companies. Traditionally, the financial bill processing of power companies relies on manual operations, which is not only inefficient but also prone to errors. In order to meet this challenge, some power financial bill management systems have appeared on the market. However, these systems still have some problems when processing bills, such as low recognition accuracy, slow processing speed, and cumbersome audit process. These problems not only affect the financial management efficiency of power companies, but also increase the operating costs of enterprises. Therefore, it is necessary to develop a more intelligent and efficient power financial bill management device to solve the deficiencies in the existing technology.

[0004] To this end, we propose an intelligent power financial bill management device Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent power financial bill management device to solve the above-mentioned deficiencies in the prior art.

[0006] In order to achieve the above-mentioned purpose, the present invention provides the following technical solution: an intelligent power financial bill management device, comprising:

[0007] Electronic invoice upload module: used to receive electronic invoice files uploaded by users through the user interface;

[0008] Preprocessing module: preprocesses the electronic receipt files received from the electronic receipt upload module to improve the accuracy of OCR recognition. The preprocessing steps include but are not limited to image denoising, adjusting the image size to meet the requirements of the OCR recognition engine, and rotation correction to ensure the correct direction of the text;

[0009] Information extraction module: uses OCR (optical character recognition technology) to extract key information from the pre-processed electronic invoice image, and the key information includes but is not limited to the invoicing unit, amount, date, invoice number, and invoicing category;

[0010] Audit module: verify the key information extracted by the information extraction module according to the preset audit rule base to determine whether the electronic bill meets the standards and requirements of the financial management of the power system. The audit rule base includes but is not limited to the amount range check and the date format check;

[0011] Classification and filing module: classifies electronic invoices according to the extracted key information such as the invoicing unit and the invoicing category, and stores the classified electronic invoice information in a classification database, which supports fast retrieval and statistical functions;

[0012] Multi-dimensional query module: provides a user interface to conduct combined queries based on multiple dimensions such as invoicing unit, invoicing category, time interval, amount interval, etc., and displays the query results, which include but are not limited to detailed information and classification information of electronic invoices;

[0013] Statistics and analysis module: automatically counts and analyzes the query results according to the user's query conditions and the selected statistical type, and displays the statistical results. The statistical types include but are not limited to total amount, total quantity, average amount, maximum / minimum amount.

[0014] As a further description of the above technical solution: the electronic invoice upload module adopts the file input control of HTML5, and the uploaded invoice format of the electronic invoice upload module includes but is not limited to PDF, JPEG, and PNG.

[0015] As a further description of the above technical solution: the specific manner in which the preprocessing module preprocesses the received electronic invoice file is:

[0016] A: Gaussian filtering algorithm is used to remove noise points in the image by setting filtering parameters, including but not limited to standard deviation, and the pre-processing module grayscales the image to reduce the amount of calculation and improve the processing speed;

[0017] B: When adjusting the image size to meet the requirements of the OCR recognition engine, the preprocessing module uses a bilinear interpolation algorithm to scale the image to ensure that the quality of the scaled image is not affected;

[0018] C: The rotation correction function of the pre-processing module detects the direction of text in the image by adopting a text direction detection algorithm based on Hough transform, and automatically rotates the image to a horizontal or vertical direction.

[0019] As a further description of the above technical solution: the specific construction method of the audit rule library in the audit module is divided into the following steps:

[0020] D: Define specific validation rules, including but not limited to amount range validation and date format validation;

[0021] E: Use programming language (Python) to write validation functions and convert the defined rules into computer code;

[0022] F: Store the encoded validation rules in a configuration file (JSON file), where each rule corresponds to a key-value pair, the key is the rule name, and the value is the path of the validation function;

[0023] G: Upload the verification test ticket through the electronic ticket upload module, and perform unit tests on each rule in the rule base to ensure that they can correctly verify the electronic ticket.

[0024] As a further description of the above technical solution: the specific method of verifying the information extracted by the information extraction module with the set rules in the audit rule library in the audit module is:

[0025] H: Perform data cleaning on the extracted information, including but not limited to removing spaces and converting data types to ensure that it matches the rules in the rule base;

[0026] I: The audit module traverses the extracted key information and verifies it according to the rules in the rule base. For each rule, it finds the corresponding verification function for verification and records the verification result returned by the verification function.

[0027] As a further description of the above technical solution: the specific method for the classification and filing module to classify the bill information includes the following steps:

[0028] J: Establish classification items, including but not limited to invoicing units and different invoicing categories, and then classify the key information extracted by the information extraction module according to the preset classification rules through a statistical clustering algorithm;

[0029] K: The classification results are stored in a classification database, which can be a relational database (MySQL, Oracle). During the storage process, a unique identifier is assigned to each classification result to facilitate subsequent retrieval and statistics.

[0030] In the above technical solution, the intelligent power financial bill management device provided by the present invention has the following beneficial effects:

[0031] Improve bill processing efficiency: By processing the financial bills of power companies, such as electricity invoices, receipts, payment vouchers, etc. in an automated and intelligent manner, the speed and efficiency of bill processing are significantly improved.

[0032] Enhanced bill recognition accuracy: Combining big data, artificial intelligence and other advanced technologies, OCR (optical character recognition technology) is used to extract key information from pre-processed electronic bill images, which improves the accuracy of bill recognition and reduces human errors.

[0033] Simplify the audit process: Verify the extracted key information according to the preset audit rule library to determine whether the electronic invoice meets the standards and requirements of the power system financial management, thereby simplifying the audit process and reducing audit costs.

[0034] Optimize bill classification and filing: Classify electronic bills according to the extracted key information such as billing units and billing categories, and store the classified electronic bill information in the classification database, supporting fast retrieval and statistical functions to facilitate subsequent management and analysis DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below.

[0036] The embodiment of the present invention provides a technical solution: comprising:

[0037] Electronic invoice upload module: used to receive electronic invoice files uploaded by users through the user interface;

[0038] The electronic invoice upload module uses HTML5 file input controls, and the formats of the uploaded invoices include but are not limited to PDF, JPEG, and PNG. The HTML5 file input control allows file reading and verification via JavaScript, but does not allow direct access to the user's file system, thus ensuring security to a certain extent.

[0039] Preprocessing module: preprocesses the electronic receipt files received from the electronic receipt upload module to improve the accuracy of OCR recognition. The preprocessing steps include but are not limited to image denoising, adjusting the image size to meet the requirements of the OCR recognition engine, and rotation correction to ensure the correct direction of the text;

[0040] The specific method in which the preprocessing module preprocesses the received electronic invoice file is as follows:

[0041] A: Gaussian filtering algorithm is used to remove noise points in the image by setting filtering parameters, including but not limited to standard deviation, and the pre-processing module grayscales the image to reduce the amount of calculation and improve the processing speed;

[0042] The noise removal method may also be adaptive filtering, non-local mean filtering, etc., which can better preserve image details and remove complex noise, or use convolutional neural network (CNN) for image denoising, which will not be described in detail;

[0043] B: When adjusting the image size to meet the requirements of the OCR recognition engine, the preprocessing module uses a bilinear interpolation algorithm to scale the image to ensure that the quality of the scaled image is not affected. When adjusting the image size, the aspect ratio of the image should be kept unchanged to avoid image deformation. Although the higher the resolution, the richer the image details, too high a resolution will increase the processing time. Therefore, the appropriate resolution should be selected according to the recommendations of the OCR recognition engine or the test results;

[0044] C: The rotation correction function of the pre-processing module detects the direction of text in the image by using a text direction detection algorithm based on Hough transform, and automatically rotates the image to a horizontal or vertical direction;

[0045] It should be noted that the preprocessing module also includes adjusting the contrast and brightness of the image to highlight the text information and converting the image into a binary image containing only black and white colors, which is helpful for subsequent character segmentation and recognition;

[0046] Information extraction module: uses OCR (optical character recognition technology) to extract key information from the pre-processed electronic invoice image, and the key information includes but is not limited to the invoicing unit, amount, date, invoice number, and invoicing category;

[0047] Audit module: verify the key information extracted by the information extraction module according to the preset audit rule base to determine whether the electronic bill meets the standards and requirements of the financial management of the power system. The audit rule base includes but is not limited to the amount range check and the date format check;

[0048] Among them, the specific construction method of the audit rule base in the audit module is divided into the following steps:

[0049] D: Define specific validation rules, including but not limited to amount range validation and date format validation;

[0050] E: Use programming language (Python) to write validation functions and convert the defined rules into computer code;

[0051] F: Store the encoded validation rules in a configuration file (JSON file), where each rule corresponds to a key-value pair, the key is the rule name, and the value is the path of the validation function;

[0052] G: Upload the verification test ticket through the electronic ticket upload module, and perform unit tests on each rule in the rule base to ensure that they can correctly verify the electronic ticket;

[0053] When storing the verification rules in the configuration file, a reasonable file structure and data format should be designed to facilitate dynamic updating and management of the rules. For example, the hierarchical structure of the JSON file can be used to organize verification rules of different categories, while providing functions such as rule enable / disable and priority setting. In the testing phase, a variety of test tickets should be used to verify the validity and accuracy of the rule base. These test tickets should cover a variety of possible legal and illegal situations to ensure that the rule base can comprehensively and accurately verify the electronic tickets. At the same time, the test process and results should be recorded to facilitate subsequent problem tracking and performance optimization;

[0054] Classification and filing module: classifies electronic invoices according to the extracted key information such as the invoicing unit and the invoicing category, and stores the classified electronic invoice information in a classification database, which supports fast retrieval and statistical functions;

[0055] The specific method of classifying bill information in the classification and filing module includes the following steps:

[0056] J: Establish classification items, including but not limited to invoicing units and different invoicing categories, and then classify the key information extracted by the information extraction module according to the preset classification rules through a statistical clustering algorithm;

[0057] K: The classification results are stored in a classification database, which can be a relational database (MySQL, Oracle). During the storage process, a unique identifier is assigned to each classification result to facilitate subsequent retrieval and statistics.

[0058] Multi-dimensional query module: provides a user interface to conduct combined queries based on multiple dimensions such as invoicing unit, invoicing category, time interval, amount interval, etc., and displays the query results, which include but are not limited to detailed information and classification information of electronic invoices;

[0059] Statistics and analysis module: automatically counts and analyzes the query results according to the user's query conditions and the statistical type selected, and displays the statistical results. The statistical types include but are not limited to total amount, total quantity, average amount, maximum / minimum amount;

[0060] It should be noted that the specific method for the audit module to verify the information extracted by the information extraction module with the set rules in the audit rule library is:

[0061] H: Perform data cleaning on the extracted information, including but not limited to removing spaces and converting data types to ensure that it matches the rules in the rule base;

[0062] I: The audit module traverses the extracted key information and verifies it according to the rules in the rule base. For each rule, it finds the corresponding verification function for verification and records the verification result returned by the verification function;

[0063] It should be noted that after the classification project is established, in addition to statistical clustering algorithms, you can also consider using unsupervised learning algorithms in machine learning, such as K-means, DBSCAN, etc. These algorithms can automatically discover potential patterns in the data and improve the accuracy of classification. The statistics and analysis module should provide a variety of statistical types, such as total amount, total quantity, average amount, maximum / minimum amount, etc., to meet the different needs of users.

[0064] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that, for those skilled in the art, the described embodiments can be modified in various ways without departing from the spirit and scope of the present invention. Therefore, the above description is illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An intelligent power financial bill management device, characterized in that: include: Electronic invoice upload module: used to receive electronic invoice files uploaded by users through the user interface; Preprocessing module: preprocesses the electronic receipt files received from the electronic receipt upload module to improve the accuracy of OCR recognition. The preprocessing steps include but are not limited to image denoising, adjusting the image size to meet the requirements of the OCR recognition engine, and rotation correction to ensure the correct direction of the text; Information extraction module: uses OCR (optical character recognition technology) to extract key information from the pre-processed electronic invoice image, and the key information includes but is not limited to the invoicing unit, amount, date, invoice number, and invoicing category; Audit module: verify the key information extracted by the information extraction module according to the preset audit rule base to determine whether the electronic bill meets the standards and requirements of the financial management of the power system. The audit rule base includes but is not limited to the amount range check and the date format check; Classification and filing module: classifies electronic invoices according to the extracted key information such as the invoicing unit and the invoicing category, and stores the classified electronic invoice information in a classification database, which supports fast retrieval and statistical functions; Multi-dimensional query module: provides a user interface to conduct combined queries based on multiple dimensions such as invoicing unit, invoicing category, time interval, amount interval, etc., and displays the query results, which include but are not limited to detailed information and classification information of electronic invoices; Statistics and analysis module: automatically counts and analyzes the query results according to the user's query conditions and the selected statistical type, and displays the statistical results. The statistical types include but are not limited to total amount, total quantity, average amount, maximum / minimum amount.

2. The intelligent power financial bill management device according to claim 1, characterized in that: The electronic invoice uploading module adopts the file input control of HTML5, and the uploaded invoice format of the electronic invoice uploading module includes but is not limited to PDF, JPEG, and PNG.

3. The intelligent power financial bill management device according to claim 1, characterized in that: The specific method in which the preprocessing module preprocesses the received electronic invoice file is as follows: A: Gaussian filtering algorithm is used to remove noise points in the image by setting filtering parameters, including but not limited to standard deviation, and the pre-processing module grayscales the image to reduce the amount of calculation and improve the processing speed; B: When adjusting the image size to meet the requirements of the OCR recognition engine, the preprocessing module uses a bilinear interpolation algorithm to scale the image to ensure that the quality of the scaled image is not affected; C: The rotation correction function of the pre-processing module detects the direction of text in the image by adopting a text direction detection algorithm based on Hough transform, and automatically rotates the image to a horizontal or vertical direction.

4. The intelligent power financial bill management device according to claim 1, characterized in that: The specific construction method of the audit rule base in the audit module is divided into the following steps: D: Define specific validation rules, including but not limited to amount range validation and date format validation; E: Use programming language (Python) to write validation functions and convert the defined rules into computer code; F: Store the encoded validation rules in a configuration file (JSON file), where each rule corresponds to a key-value pair, the key is the rule name, and the value is the path of the validation function; G: Upload the verification test ticket through the electronic ticket upload module, and perform unit tests on each rule in the rule base to ensure that they can correctly verify the electronic ticket.

5. The intelligent power financial bill management device according to claim 1, characterized in that: The specific method of verifying the information extracted by the information extraction module with the set rules in the audit rule library in the audit module is: H: Perform data cleaning on the extracted information, including but not limited to removing spaces and converting data types to ensure that it matches the rules in the rule base; I: The audit module traverses the extracted key information and verifies it according to the rules in the rule base. For each rule, it finds the corresponding verification function for verification and records the verification result returned by the verification function.

6. The intelligent power financial bill management device according to claim 1, characterized in that: The specific method for the classification and filing module to classify the bill information includes the following steps: J: Establish classification items, including but not limited to invoicing units and different invoicing categories, and then classify the key information extracted by the information extraction module according to the preset classification rules through a statistical clustering algorithm; K: The classification results are stored in a classification database, which can be a relational database (MySQL, Oracle). During the storage process, a unique identifier is assigned to each classification result to facilitate subsequent retrieval and statistics.