Accounting model for automatically generating voucher for bill data

Abstract

The invention discloses an accounting model of a bill data automatic generation voucher, and relates to the technical field of accounting processing. According to the invention, full-process automation from bill data input to voucher generation is realized, tedious and time-consuming multi-step manual input, classification and template setting links in a traditional mode are thoroughly eliminated, and the voucher generation period is remarkably shortened; through automatic two-dimensional code recognition, OCR information extraction, business type prediction and voucher template filling, manual intervention and manual operation of financial staff in a basic data processing link are reduced to the maximum extent, the labor cost is reduced, and input errors and information omission possibly caused by manual operation are avoided; the large model is used for carrying out deep learning and reasoning prediction on bill data, multiple possible business types and probabilities thereof in a complex scene can be identified, the first three possibilities are returned for reference, and the judgment capability which is more flexible than traditional rule matching and closer to actual business is provided.

Description

Technical Field

[0001] This invention relates to the field of accounting technology, specifically to an accounting model that automatically generates vouchers from invoice data. Background Technology

[0002] Negotiable instruments are legally created documents in a prescribed form that demonstrate an obligation to pay money. In a broad sense, they include various securities and certificates, such as stocks, treasury bonds, corporate bonds, invoices, and bills of lading. In a narrow sense, negotiable instruments refer only to those stipulated in the Negotiable Instruments Law. Accounting refers to the specific tasks involved in collecting, organizing, recording, calculating, and reporting original documents to achieve accounting processing. It requires standardization and accuracy to ensure the effective implementation of accounting, accounting supervision, and accounting standards.

[0003] Current accounting software typically uses a semi-automated approach to voucher generation, breaking it down into several steps that require manual intervention to set up or pre-categorize and retrieve invoice data. For cost and expense invoices that cannot be retrieved from the tax bureau's system, manual setup or data entry is necessary. Therefore, we propose an accounting model that automatically generates vouchers from invoice data. Summary of the Invention

[0004] The purpose of this invention is to provide an accounting model for automatically generating vouchers from invoice data in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention specifically adopts the following technical solution:

[0006] An accounting model for automatically generating vouchers from invoice data includes:

[0007] Value-added tax (VAT) invoice QR code scanning system: used to detect the outline of the QR code, accurately locate the position of the QR code in the VAT invoice image, and decode the QR code encoding to identify the five elements contained therein;

[0008] Key information extraction system: used to recognize text in invoice images. By building a large model through the prompt project, the main information in the invoice is extracted and formed into a key-value pair data structure.

[0009] The bill classification system uses bills with extracted key information as a basis for label training. After model deployment, it performs inference and classification on bill images.

[0010] Business Type Prediction Model System: After the model is deployed, it performs inference on paper document data such as bank statement data and VAT invoices to predict various possible business types and their proportions, and outputs the top three business types with the highest proportions.

[0011] Automatic voucher generation system: It adopts a message queue mechanism to process paper invoices uploaded by users or full invoice data obtained from tax authorities.

[0012] Furthermore, it also includes: an invoice management cloud platform, which builds an invoice management platform based on cloud computing to realize centralized storage, processing and analysis of invoice data, and supports multiple users to operate online at the same time.

[0013] Furthermore, it also includes an invoice data visualization unit, which uses data visualization technology to transform invoice data into intuitive charts and reports.

[0014] Furthermore, the QR code scanning system uses Halcon to perform QR code contour detection based on the black and white channels of the image.

[0015] Furthermore, the key information extraction system uses OCR technology to recognize text in the invoice image.

[0016] Furthermore, the key information extraction system includes an intelligent review unit and a data cleaning and preprocessing unit. The intelligent review unit uses deep learning algorithms to intelligently review the content of invoices. The data cleaning and preprocessing unit cleans and preprocesses invoice data from different sources.

[0017] Furthermore, the bill classification system utilizes OCR technology to recognize text in bill images.

[0018] An accounting model for automatically generating vouchers from invoice data includes the following steps:

[0019] Step 1: Invoice classification and processing. Invoices are scanned using a QR code scanning system and automatically classified according to their content.

[0020] Step 2: QR code and OCR text recognition, performing the scanning of VAT invoices and the recognition of text information;

[0021] Step 3: Invoice information verification and key data extraction. This involves verifying the information and extracting key data from the identified invoices.

[0022] Step 4: Business type prediction and analysis. Utilize the constructed large-scale business type prediction model system to perform predictive analysis of business types.

[0023] Step 5: Automatic voucher generation. Based on the predictive analysis results, accounting vouchers are automatically generated.

[0024] Step Six: Voucher Review and Verification. The system automatically reviews and verifies the generated accounting vouchers to ensure they comply with financial standards and regulations, reducing the time and cost of manual review.

[0025] Step 7: Data storage and archiving. The audited accounting vouchers will be securely stored in the system for easy retrieval and auditing. At the same time, the system will automatically archive the voucher data.

[0026] Furthermore, after predicting the business type in the business type prediction process, the corresponding system preset voucher template is applied, and the subject and amount of the template are automatically replaced. Finally, the voucher data is saved and the subject balance is updated.

[0027] The beneficial effects of this invention are as follows:

[0028] 1. This invention significantly improves accounting efficiency: it automates the entire process from inputting invoice data (scanning / uploading / capturing) to generating vouchers, completely eliminating the cumbersome and time-consuming multi-step manual input, classification, and template setting steps in the traditional method, and significantly shortening the voucher generation cycle.

[0029] 2. This invention effectively reduces labor costs and error rates: By automating QR code recognition, OCR information extraction, business type prediction, and voucher template filling, it minimizes the manual intervention and operation of financial personnel in the basic data processing stage, thereby reducing labor costs and avoiding input errors and information omissions that may be caused by human operation.

[0030] 3. This invention has a wide processing range and strong applicability: The model can simultaneously process multiple common invoice sources, including: paper invoices (through QR code scanning and OCR recognition), full invoice data from the tax bureau (through system integration or capture), electronic bank receipt data, and other cost and expense invoices that cannot be directly obtained from the tax bureau system (through OCR recognition and key information extraction), covering the main invoice types for daily accounting processing of enterprises.

[0031] 4. This invention improves the accuracy and consistency of voucher generation: Based on a combination of rules (QR code recognition) and large models (key information extraction, business type prediction), it ensures the accuracy of core information extracted from vouchers (such as the five elements of a VAT invoice) and business type judgment. It uses preset voucher templates to automatically replace accounts and amounts, ensuring the standardization and consistency of voucher entries, which meets the requirements of accounting standards.

[0032] 5. This invention enhances the intelligence level of business type judgment: By using a large model to perform deep learning and inference prediction on invoice data, it can identify multiple possible business types and their probabilities in complex scenarios, and return the top three possibilities for reference, providing a more flexible and more practical judgment capability than traditional rule matching.

[0033] 6. This invention realizes the processization and standardization of data processing: the entire processing flow is connected through message queues to form an efficient, orderly, and monitorable automated pipeline, ensuring the standardization and manageability of the data processing process.

[0034] 7. This invention updates accounting data in real time: After vouchers are generated, the balances of relevant accounts are automatically saved and updated to ensure the real-time nature and accuracy of accounting data, providing a reliable foundation for subsequent report preparation and financial analysis.

[0035] 8. This invention frees up the energy of finance personnel and enhances their value: It liberates finance personnel from heavy and repetitive data entry and basic voucher generation work, allowing them to focus on more valuable tasks such as financial auditing, analysis, decision support, and risk control. Attached Figure Description

[0036] Figure 1 This is a block diagram of the present invention;

[0037] Figure 2 This is a flowchart of the process of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0039] Please see Figure 1 This invention provides an accounting model for automatically generating vouchers from invoice data, comprising:

[0040] Value-added tax (VAT) invoice QR code scanning system: used to detect the outline of the QR code, accurately locate the position of the QR code in the VAT invoice image, and decode the QR code encoding to identify the five elements contained therein; ensuring the rapid and accurate acquisition of invoice information.

[0041] Key Information Extraction System: Used to recognize text in invoice images. Through the construction of a large-scale model in the prompt project, the main information in the invoice is extracted and formed into a key-value pair data structure. Deep learning technology is used to improve the efficiency and accuracy of information extraction.

[0042] The bill classification system uses bills with extracted key information as a basis for label training. After model deployment, it performs inference and classification on bill images.

[0043] The large-scale model system for predicting business types: After deployment, the model performs inference on paper document data such as bank receipts and VAT invoices to predict various possible business types and their proportions, and outputs the top three business types with the highest proportions. The document classification system and the large-scale model system for predicting business types further improve the intelligence level of document processing, enabling the system to automatically determine the type and business scope of a document based on its content.

[0044] Automatic voucher generation system: Employing a message queue mechanism, it processes paper invoices uploaded by users or all invoice data obtained from tax authorities. The system integrates this information and transforms it into vouchers that comply with financial regulations, greatly reducing the workload of financial personnel and improving the efficiency and accuracy of accounting processing.

[0045] In this embodiment, preferably, it also includes: an invoice management cloud platform, which constructs a cloud-based invoice management platform to realize centralized storage, processing, and analysis of invoice data, supports simultaneous online operation by multiple users, and improves the efficiency and security of invoice management. This platform can monitor invoice status in real time, prevent problems such as lost or duplicate reimbursement of invoices, and ensure the accuracy and compliance of corporate financial data; at the same time, the invoice management cloud platform provides powerful query and statistical functions to help financial personnel quickly locate the required invoice information and optimize financial management processes.

[0046] In this embodiment, preferably, it also includes an invoice data visualization unit. This unit utilizes data visualization technology to transform invoice data into intuitive charts and reports, helping users quickly understand the business implications behind the data and improving decision-making efficiency. This invoice data visualization unit can update data in real time, displaying key information such as the flow of invoices, amount distribution, and business type proportions. This allows users to grasp the company's financial situation and operational trends at a glance. Through charts and reports, complex financial data is simplified, enabling non-financial professionals to easily understand and participate in the company's financial management, further promoting internal communication and collaboration. The invoice data visualization unit also supports user-defined reports to meet the personalized needs of different departments and positions, enhancing the system's flexibility and practicality.

[0047] In this embodiment, preferably, the QR code scanning system uses Halcon to perform QR code contour detection based on the black and white channels of the image. This system can quickly and accurately identify and parse invoice information from QR codes, greatly improving the speed and accuracy of data entry. Through Halcon's powerful image processing algorithms, it can achieve stable and reliable recognition even in low light conditions or when the QR code is slightly damaged. The QR code scanning system also has an automatic error correction function, which can intelligently identify and correct errors in the QR code, ensuring the integrity and accuracy of the data. This feature enables the accounting model to maintain efficient and stable operation when processing large amounts of invoice data, providing strong support for enterprise financial management.

[0048] In this embodiment, preferably, the key information extraction system uses OCR technology to recognize text in invoice images. This system can efficiently and accurately extract key information from invoice images, such as invoice number, date, amount, and issuing unit, avoiding the tediousness and errors of manual input. The use of OCR technology enables the system to process invoice images of various formats and qualities, including handwritten, printed, or stamped invoices, greatly improving the breadth and accuracy of information extraction. The key information extraction system also supports multilingual recognition, adapting to the invoice processing needs of different countries and regions. This function provides great convenience for the financial management of multinational corporations, ensuring global consistency and accuracy in accounting processes.

[0049] In this embodiment, preferably, the key information extraction system includes an intelligent review unit and a data cleaning and preprocessing unit. The intelligent review unit uses a deep learning algorithm. The data cleaning and preprocessing unit cleans and preprocesses invoice data from different sources to ensure data quality and provide accurate basic data for subsequent data analysis and mining.

[0050] The intelligent auditing unit incorporates various auditing rules and logics, automatically identifying anomalies in invoices such as duplicate invoices, inconsistent amounts, and contradictory dates. This effectively prevents financial fraud and misoperation. Furthermore, this unit can flexibly customize auditing rules according to specific enterprise needs, adapting to the financial management standards of different industries and companies. This enables intelligent auditing of invoice content, ensuring the accuracy and compliance of invoice information and reducing the workload and error rate of manual auditing. The data cleaning and preprocessing unit utilizes advanced data cleaning technology to deduplicate, fill in missing values, and correct errors in invoice data, ensuring data integrity and consistency. This unit can also format invoice data, transforming it into a standardized data format for easier subsequent data processing and analysis. Through the collaborative work of these two units, the key information extraction system provides enterprises with high-quality, accurate invoice data, laying a solid foundation for the automated processing of accounting models.

[0051] In this embodiment, preferably, the invoice classification system utilizes OCR technology to recognize text in invoice images. This system can accurately and quickly recognize textual information on invoices, including but not limited to key fields such as invoice header, amount, date, and product name. Through advanced OCR technology, the invoice classification system can handle invoice images with various complex backgrounds, effectively overcoming recognition challenges such as image blur and text distortion, ensuring the accuracy and reliability of the recognition results. The recognized textual information will be automatically classified and stored, providing strong support for subsequent data processing and accounting model generation. This function greatly improves the efficiency and accuracy of invoice processing, reducing the tediousness and errors of manual input.

[0052] Please see Figure 2 The present invention also provides a working method for an accounting processing model that automatically generates accounting vouchers based on invoice data, comprising the following steps:

[0053] Step 1: Invoice classification and processing. Invoices are scanned using a QR code scanning system and automatically classified according to their content.

[0054] Step 2: QR code and OCR text recognition, performing the scanning of VAT invoices and the recognition of text information;

[0055] Step 3: Invoice information verification and key data extraction. This involves verifying the information and extracting key data from the identified invoices.

[0056] Step 4: Business type prediction and analysis. Utilize the constructed large-scale business type prediction model system to perform predictive analysis of business types.

[0057] Step 5: Automatic voucher generation. Based on the predictive analysis results, accounting vouchers are automatically generated.

[0058] Step Six: Voucher Review and Verification. The system automatically reviews and verifies the generated accounting vouchers to ensure they comply with financial standards and regulations, reducing the time and cost of manual review.

[0059] Step 7: Data storage and archiving. The audited accounting vouchers will be securely stored in the system for easy retrieval and auditing. At the same time, the system will automatically archive the voucher data to maintain data integrity and traceability. Through this series of refined processes, the accounting processing model not only improves work efficiency but also ensures the accuracy and security of data processing.

[0060] In this embodiment, preferably, after the business type is predicted in the business type prediction, the corresponding system preset voucher template is applied, and the subject and amount of the template are automatically replaced. Finally, the voucher data is saved and the subject balance is updated.

[0061] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily 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 invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An accounting model for automatically generating vouchers from invoice data, characterized in that, include: Value-added tax (VAT) invoice QR code scanning system: used to detect the outline of the QR code, accurately locate the position of the QR code in the VAT invoice image, and decode the QR code encoding to identify the five elements contained therein; Key information extraction system: used to recognize text in invoice images. By building a large model through the prompt project, the main information in the invoice is extracted and formed into a key-value pair data structure. The bill classification system uses bills with extracted key information as a basis for label training. After model deployment, it performs inference and classification on bill images. Business Type Prediction Model System: After the model is deployed, it performs inference on paper document data such as bank statement data and VAT invoices to predict various possible business types and their proportions, and outputs the top three business types with the highest proportions. Automatic voucher generation system: It adopts a message queue mechanism to process paper invoices uploaded by users or full invoice data obtained from tax authorities.

2. The accounting model for automatically generating vouchers from invoice data according to claim 1, characterized in that, Also includes: The invoice management cloud platform is a cloud-based invoice management platform that enables centralized storage, processing, and analysis of invoice data, and supports simultaneous online operation by multiple users.

3. The accounting model for automatically generating vouchers from invoice data according to claim 1, characterized in that, Also includes: The invoice data visualization unit uses data visualization technology to transform invoice data into intuitive charts and reports, helping users quickly understand the business implications behind the data and improve decision-making efficiency.

4. The accounting model for automatically generating vouchers from invoice data according to claim 1, characterized in that, The QR code scanning system uses Halcon to detect the outline of the QR code based on the black and white channels of the image.

5. The accounting model for automatically generating vouchers from invoice data according to claim 1, characterized in that, The key information extraction system uses OCR technology to recognize text in invoice images.

6. The accounting model for automatically generating vouchers from invoice data according to claim 5, characterized in that, The key information extraction system includes an intelligent review unit and a data cleaning and preprocessing unit. The intelligent review unit uses deep learning algorithms to intelligently review the content of invoices. The data cleaning and preprocessing unit cleans and preprocesses invoice data from different sources.

7. The accounting model for automatically generating vouchers from invoice data according to claim 1, characterized in that, The invoice classification system uses OCR technology to recognize text in invoice images.

8. An accounting model for automatically generating vouchers from invoice data according to any one of claims 1-7, characterized in that, The work includes the following steps: Step 1: Invoice classification and processing. Invoices are scanned using a QR code scanning system and automatically classified according to their content. Step 2: QR code and OCR text recognition, performing the scanning of VAT invoices and the recognition of text information; Step 3: Invoice information verification and key data extraction. This involves verifying the information and extracting key data from the identified invoices. Step 4: Business type prediction and analysis. Utilize the constructed large-scale business type prediction model system to perform predictive analysis of business types. Step 5: Automatic voucher generation. Based on the predictive analysis results, accounting vouchers are automatically generated. Step Six: Voucher Review and Verification. The system automatically reviews and verifies the generated accounting vouchers to ensure they comply with financial standards and regulations, reducing the time and cost of manual review. Step 7: Data storage and archiving. The audited accounting vouchers will be securely stored in the system for easy retrieval and auditing. At the same time, the system will automatically archive the voucher data.

9. The accounting model for automatically generating vouchers from invoice data according to claim 8, characterized in that, After predicting the business type, the system applies the corresponding preset voucher template, automatically replaces the account and amount in the template, and finally saves the voucher data and updates the account balance.

Images