Implementation method, system and device of cross-border trade service platform and storage medium

Abstract

The invention discloses an implementation method, system and device of a cross-border trade service platform and a storage medium, and mainly relates to the technical field of trade services. The method includes acquiring image samples; establishing an OCR model, converting text data in the collected image sample into an electronic text, and generating an electronic form; and generating an RPA flow according to the electronic form. The method has the beneficial effects that the document processing efficiency and accuracy can be greatly improved, and the operation cost of an enterprise can be remarkably reduced.

Description

Technical Field

[0001] The present invention relates to the field of trade service technology, and specifically to an implementation method, system, device and storage medium of a cross-border trade service platform. Background Art

[0002] With the acceleration of global economic integration, cross-border trade activities are becoming more frequent, and the processing of a large number of cross-border business documents has become increasingly important. These documents include invoices, customs declarations, contracts, bills of lading and other types. They are not only huge in number but also in various formats, and they also involve multiple languages. Traditional manual processing methods are not only inefficient but also prone to errors, especially in the case of multiple languages ​​and complex formats. These problems are particularly prominent.

[0003] The artificial intelligence technology currently used to solve the above problems faces the following problems:

[0004] 1. Low processing efficiency: Current document processing systems can only process documents in specific formats, and have limited processing capabilities for documents in diverse and complex formats;

[0005] 2. Insufficient multi-language support: Most systems only support a few mainstream languages ​​and lack support for minority languages, which limits their global application;

[0006] 3. Low degree of automation: Although some systems have achieved partial automation, manual intervention is still required for tasks such as document classification and data verification.

[0007] Therefore, there is an urgent need for an implementation method, system, device and storage medium of a cross-border trade service platform to solve the above problems. Summary of the invention

[0008] The purpose of the present invention is to provide a method, system, device and storage medium for implementing a cross-border trade service platform, which can not only greatly improve the efficiency and accuracy of document processing, but also significantly reduce the operating costs of enterprises.

[0009] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0010] On the one hand, a method for implementing a cross-border trade service platform is provided, comprising the following steps:

[0011] S1: collect image samples;

[0012] S2: Establish an OCR model to convert text data in the collected image samples into electronic text and generate an electronic form;

[0013] S3: Generate RPA process based on electronic form.

[0014] Preferably, the collecting of image samples in step S1 is specifically:

[0015] Collect multilingual documents in the cross-border transaction process, including handwritten documents and computer-generated documents, and save the handwritten documents and computer-generated documents in image format.

[0016] Preferably, the step S2 is specifically:

[0017] S21: Select the PyTorch deep learning framework, build an OCR model based on convolutional neural network and recurrent neural network, and use the collected multilingual documents to pre-train the OCR model;

[0018] S22: Establishing recognition algorithm models for tables and lists;

[0019] S23: Establish multilingual model and language detection model;

[0020] S24: Establish recognition models for handwriting and blurred documents;

[0021] S25: Outputting an electronic form according to the established model.

[0022] Preferably, the establishment of the recognition algorithm model for the table and the list is specifically as follows: establishing a TSRFormer algorithm model, predicting the segmentation line by direct regression, using a number of sampling points on each segmentation line to represent the segmentation line, and allowing the model to directly regress the coordinates of the sampling points on each segmentation line;

[0023] The establishment of a multilingual model is specifically as follows:

[0024] Use multilingual text data for pre-training to enable the model to learn rich language features; fine-tune on multilingual tasks to make the model more suitable for specific application scenarios; use the Transformer model to capture the correlation between different languages ​​through the self-attention mechanism;

[0025] The language detection model is established by: extracting features such as vocabulary frequency and grammatical structure from the text for training the language detection model; selecting a support vector machine to build the language detection model; using annotated multilingual data sets to train the model, and evaluating the performance of the model through a cross-validation method.

[0026] Preferably, the establishment of a recognition model for handwriting and fuzzy documents comprises the following steps:

[0027] Use MNIST to collect handwritten samples and fuzzy document sample data and annotate the data;

[0028] Use rotation, scaling, translation, and noise addition to enhance data to improve the generalization ability of the model;

[0029] For blurry documents, a denoising algorithm is used to improve image quality;

[0030] Convert grayscale images into binary images to reduce computational complexity;

[0031] Unify the image size and normalize the pixel values;

[0032] Use CNN to automatically extract high-level features;

[0033] Convert the trained model to ONNX format.

[0034] Preferably, the RPA generation process is specifically as follows:

[0035] Analyze existing business processes and select repetitive and regular workflows for marking;

[0036] Use UiPath to edit the PRA process;

[0037] Perform optimization tests on the generated PRA process.

[0038] On the other hand, the implementation system of the cross-border trade service platform is provided, including:

[0039] The data acquisition module is used to: acquire image samples;

[0040] The model building module is used to: build an OCR model, convert text data in the collected image samples into electronic text, and generate an electronic form;

[0041] The data output module is used to generate RPA processes based on electronic forms.

[0042] On the other hand, a device for implementing a cross-border trade service platform is provided, comprising a processor and a memory for storing a computer program, wherein when the processor executes the computer program, the steps of any of the methods described above are implemented.

[0043] On the other hand, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by one or more processors, the steps of any of the above methods are implemented.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] 1. Apply OCR and RPA technologies to the field of cross-border trade information technology. Through the application of new technology scenarios, we can solve the problems of cumbersome, inefficient and error-prone form processing in cross-border trade, large differences in form formats in different countries and regions, and manual processing affecting the progress of trade processes;

[0046] 2. Introducing OCR and RPA technologies: OCR can quickly and accurately identify text information in forms, overcoming the difficulties of different languages ​​and handwriting recognition; RPA simulates manual operations, automatically classifies, organizes and enters data, and realizes automated form processing;

[0047] 3. Improve efficiency and speed up the progress of trade process. Improve accuracy and reduce the incidence of human errors. Easy to operate and reduce the difficulty of operation. Save manpower and reduce enterprise costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of the method of the present invention.

[0049] Figure 2 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0050] The present invention will be further described below in conjunction with specific embodiments. It should be understood that these embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art can make various changes or modifications to the present invention, and these equivalent forms fall within the scope limited by the application equally.

[0051] In the present invention, terms such as "upper", "lower", "left", "right", "front", "back", "vertical", "horizontal", "side", "bottom", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are relational words determined only for the convenience of describing the structural relationships of the various parts or elements of the present invention, and do not specifically refer to any part or element in the present invention and should not be understood as limitations on the present invention.

[0052] In the present invention, terms such as "fixed connection", "connected", "connection", etc. should be understood in a broad sense, indicating that it can be fixedly connected, integrally connected or detachably connected; it can be directly connected or indirectly connected through an intermediate medium. Relevant scientific research or technical personnel in this field can determine the specific meanings of the above terms in the present invention according to specific circumstances, and they should not be understood as limiting the present invention.

[0053] Example:

[0054] like Figure 1 As shown, this embodiment provides a method for implementing a cross-border trade service platform, comprising the following steps: S1: collecting image samples;

[0055] S2: Establish an OCR model to convert text data in the collected image samples into electronic text and generate an electronic form;

[0056] S3: Generate RPA process based on electronic form.

[0057] The collected image samples in step S1 are specifically:

[0058] Collect multilingual documents in the cross-border transaction process, including handwritten documents and computer-generated documents, and save the handwritten documents and computer-generated documents in image format.

[0059] The step S2 is specifically:

[0060] S21: Select the PyTorch deep learning framework, build an OCR model based on convolutional neural network and recurrent neural network, and use the collected multilingual documents to pre-train the OCR model;

[0061] S22: Establishing recognition algorithm models for tables and lists;

[0062] S23: Establish multilingual model and language detection model;

[0063] S24: Establish recognition models for handwriting and blurred documents;

[0064] S25: Outputting an electronic form according to the established model.

[0065] The establishment of a recognition algorithm model for tables and lists is specifically as follows: establishing a TSRFormer algorithm model, predicting the segmentation line by direct regression, using a number of sampling points on each segmentation line to represent the segmentation line, and allowing the model to directly regress the coordinates of the sampling points on each segmentation line;

[0066] The establishment of a multilingual model is specifically as follows:

[0067] Use multilingual text data for pre-training to enable the model to learn rich language features; fine-tune on multilingual tasks to make the model more suitable for specific application scenarios; use the Transformer model to capture the correlation between different languages ​​through the self-attention mechanism;

[0068] The language detection model is established by: extracting features such as vocabulary frequency and grammatical structure from the text for training the language detection model; selecting a support vector machine to build the language detection model; using annotated multilingual data sets to train the model, and evaluating the performance of the model through a cross-validation method.

[0069] The establishment of a recognition model for handwriting and fuzzy documents includes the following steps:

[0070] Use MNIST to collect handwritten samples and fuzzy document sample data and annotate the data;

[0071] Use rotation, scaling, translation, and noise addition to enhance data to improve the generalization ability of the model;

[0072] For blurry documents, a denoising algorithm is used to improve image quality;

[0073] Convert grayscale images into binary images to reduce computational complexity;

[0074] Unify the image size and normalize the pixel values;

[0075] Use CNN to automatically extract high-level features;

[0076] Convert the trained model to ONNX format.

[0077] Test the OCR model using a standard test set, record the recognition accuracy in different languages ​​and document types, and analyze the model's performance under different lighting conditions, font sizes, font styles, etc.

[0078] Record the time it takes the OCR module to process different types of documents and analyze the factors that affect the processing speed. Compare the processing speed of different models under the same hardware conditions.

[0079] Evaluate the performance of the OCR model on blurry documents, handwritten documents, etc. Analyze the model's adaptability to images of different resolutions.

[0080] The specific process of generating RPA is as follows:

[0081] Analyze existing business processes and select repetitive and regular workflows for marking;

[0082] Use UiPath to edit the PRA process;

[0083] Perform optimization tests on the generated PRA process.

[0084] Record the execution time of the complete process from document scanning to data entry. Analyze the execution efficiency of different process templates;

[0085] Record the execution success rate of the RPA process, that is, the proportion of processes that are successfully completed. Analyze the reasons for failure and propose improvement measures;

[0086] Test the performance of the RPA process when encountering abnormal situations, such as network interruption, data loss, etc. Record the effectiveness of the exception handling mechanism.

[0087] like Figure 2 As shown, this embodiment also provides a system for implementing a cross-border trade service platform, including:

[0088] The data acquisition module is used to: acquire image samples;

[0089] The model building module is used to: build an OCR model, convert text data in the collected image samples into electronic text, and generate an electronic form;

[0090] The data output module is used to generate RPA processes based on electronic forms.

[0091] The contents of the above method embodiments are all applicable to the present system embodiments. The functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0092] The present invention also provides a device, comprising:

[0093] at least one processor;

[0094] At least one memory for storing at least one program;

[0095] When the at least one program is executed by at least one processor, the at least one processor implements the implementation method of the cross-border trade service platform.

[0096] An embodiment of the present invention further provides a storage medium, wherein the storage medium stores a program, and the program is executed by a processor to complete an implementation method of a cross-border trade service platform.

[0097] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the described embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.

Claims

1. A method for implementing a cross-border trade service platform, characterized in that: The following steps are involved: S1: collect image samples; S2: Establish an OCR model to convert text data in the collected image samples into electronic text and generate an electronic form; S3: Generate RPA process based on electronic form.

2. The method for implementing the cross-border trade service platform according to claim 1 is characterized in that: The acquisition of image samples in step S1 is specifically as follows: Collect multilingual documents in the cross-border transaction process, including handwritten documents and computer-generated documents, and save the handwritten documents and computer-generated documents in image format.

3. The implementation method of the cross-border trade service platform according to claim 1 is characterized in that: The step S2 is specifically: S21: Select the PyTorch deep learning framework, build an OCR model based on convolutional neural network and recurrent neural network, and use the collected multilingual documents to pre-train the OCR model; S22: Establishing recognition algorithm models for tables and lists; S23: Establish multilingual model and language detection model; S24: Establish recognition models for handwriting and blurred documents; S25: Outputting an electronic form according to the established model.

4. The method for implementing the cross-border trade service platform according to claim 3 is characterized in that: The establishment of a recognition algorithm model for tables and lists is specifically as follows: establishing a TSRFormer algorithm model, predicting the segmentation line by direct regression, using a number of sampling points on each segmentation line to represent the segmentation line, and allowing the model to directly regress the coordinates of the sampling points on each segmentation line; The establishment of a multilingual model is specifically as follows: Use multilingual text data for pre-training to enable the model to learn rich language features; fine-tune on multilingual tasks to make the model more suitable for specific application scenarios; use the Transformer model to capture the correlation between different languages ​​through the self-attention mechanism; The language detection model is established by: extracting features such as vocabulary frequency and grammatical structure from the text for training the language detection model; selecting a support vector machine to build the language detection model; using annotated multilingual data sets to train the model, and evaluating the performance of the model through a cross-validation method.

5. The method for implementing the cross-border trade service platform according to claim 3 is characterized in that: The establishment of a recognition model for handwriting and fuzzy documents includes the following steps: Use MNIST to collect handwritten samples and fuzzy document sample data and annotate the data; Use rotation, scaling, translation, and noise addition to enhance data to improve the generalization ability of the model; For blurry documents, a denoising algorithm is used to improve image quality; Convert grayscale images into binary images to reduce computational complexity; Unify the image size and normalize the pixel values; Use CNN to automatically extract high-level features; Convert the trained model to ONNX format.

6. The method for implementing the cross-border trade service platform according to claim 2 is characterized in that: The specific process of generating RPA is as follows: Analyze existing business processes and select repetitive and regular workflows for marking; Use UiPath to edit the PRA process; Perform optimization tests on the generated PRA process.

7. The implementation system of the cross-border trade service platform is characterized by: include: The data acquisition module is used to: acquire image samples; The model building module is used to: build an OCR model, convert text data in the collected image samples into electronic text, and generate an electronic form; The data output module is used to generate RPA processes based on electronic forms.

8. A device for implementing a cross-border trade service platform, characterized in that: The method comprises a processor and a memory for storing a computer program, wherein when the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by one or more processors, the steps of the method according to any one of claims 1 to 6 are implemented.

Images