Method, System, Device and Medium for Verifying Itinerary Receipt of Air Ticket

Abstract

The present invention provides a method, system, device and medium for verifying an air ticket itinerary, which relates to the technical fields of fully automatic process robots, artificial neural networks, and air ticket verification technologies, and includes: using process robot technology and neural network technology to achieve full automation of the business process for verifying an air ticket itinerary, using intelligent process robots to reduce labor costs, avoid repeated operations, and improve verification efficiency. Business personnel can start the intelligent process robot program through desktop software interaction. This program can quickly locate and identify the four elements for verifying an air ticket based on the air ticket photo, automatically log in to the Ctrip Ticket Query and Verification website to confirm the authenticity of the air ticket, and record the verification result. The present invention realizes the full-automatic operation of the process, improves work efficiency, releases human resources, and ensures the timeliness and accuracy of the verification work.

Description

Technical Field

[0001] The present invention relates to the technical fields of fully automatic process robots, artificial neural networks, and verification technologies for airline tickets. Specifically, it relates to a method for verifying the itinerary of airline tickets based on neural network text recognition and intelligent process robots, and in particular, a method, system, device, and medium for verifying the itinerary of airline tickets. Background Art

[0002] With the development of the field of artificial intelligence, OCR (Optical Character Recognition) has more and more application scenarios in the banking industry. At the same time, the banking industry is also troubled by "fake tickets" caused by human errors or malicious operations. These tickets lead to false reimbursements and a waste of human resources for verification. Therefore, in order to ensure the correctness of airline tickets, solve the timeliness of verification, and at the same time reduce labor costs and improve verification efficiency, there is an urgent need for a method that can fully automatically verify airline tickets. Summary of the Invention

[0003] In view of the deficiencies in the prior art, the present invention provides a method, system, device, and medium for verifying the itinerary of airline tickets.

[0004] According to the method, system, device, and medium for verifying the itinerary of airline tickets provided by the present invention, the solutions are as follows:

[0005] In the first aspect, a method for verifying the itinerary of airline tickets is provided. The method includes:

[0006] Step S1: Start the main program of the process robot and initialize the parameters;

[0007] Step S2: The process robot sequentially accesses the airline ticket photos within the specified path;

[0008] Step S3: The process robot preliminarily determines whether the size of the airline ticket photo meets the set requirements. If the condition is met, execute Step S4; if not, execute Step S13;

[0009] Step S4: The process robot calls the text detection program to preprocess the airline ticket photos that meet the standards in Step S3 to determine the text area;

[0010] After Step S4, the process robot calls the text recognition program. The text recognition program uses a single-vision text recognition algorithm, introduces a pre-trained model based on the self-attention mechanism to mine context information in downstream tasks, and improves the text recognition ability; through this program, the recognition of the text area in Step S4 obtains the recognized text, credibility, and the coordinate values of its range;

[0011] Step S6: The process robot calls the element extraction program, which can determine the position of the elements in the original image based on the text coordinate values and the original image information obtained in step S5, so as to achieve element extraction;

[0012] Step S7: After step S6, the process robot screens the element recognition results;

[0013] Step S8: After step S7, the process robot logs in to the Xintianyou website and calls the text recognition program to recognize the arithmetic captcha on the website;

[0014] Step S9: After step S8, the process robot calls the text processing program, which can traverse the elements to determine whether they contain operators and convert the string into an arithmetic expression, so as to implement the processing of the four arithmetic operations in the text and return the calculation result;

[0015] Step S10: After step S9, the process robot automatically enters the Xintianyou website according to the element information obtained in step S7 and the captcha calculation result obtained in step S9, and starts the verification;

[0016] Step S11: After the process robot finishes the verification, it judges whether the verification result of the air ticket is true;

[0017] Step S12: The process robot locates and crops the screenshot obtained in step S11, retains the electronic air ticket information, and saves the cropping result to the specified path;

[0018] Step S13: The process robot records the relevant information including the file name, verification result, and failure reason in the specified file for the relevant situations including: unqualified air ticket photos, text recognition failure, and verification failure;

[0019] Step S14: The process robot judges whether there are still unprocessed air ticket photos. If so, it returns to step S2. If not, it executes step S15;

[0020] Step S15: After the process robot finishes its work, it sorts out the air ticket verification results, and the air ticket verification process ends.

[0021] Preferably, the text detection program in step S4 is based on the differentiable binarization algorithm and uses the principle of image segmentation to realize text detection in related multi-shape and multi-direction scenarios including bending, twisting, and tilting.

[0022] Preferably, in step S7, the process robot screens the element recognition results. If the passenger name, itinerary number, electronic ticket number, and total price have all been recognized and the credibility is greater than the threshold, it executes step S8. Otherwise, it executes step S13.

[0023] Preferably, after the process robot verification in step S1, if the verification result of the airline ticket is true, a full-screen screenshot of the page returned by TravelSky is taken, and information such as the ticket file name and verification result is recorded in a specified file, and then step S2 is entered; if the ticket verification result is false, step S3 is entered.

[0024] In a second aspect, an airline ticket itinerary verification system is provided, and the system includes:

[0025] Module M1: Start the main program of the process robot and initialize the parameters;

[0026] Module M2: The process robot sequentially accesses the airline ticket photos in the specified path;

[0027] Module M3: The process robot preliminarily judges whether the size of the airline ticket photo meets the set requirements. If the conditions are met, module M4 is executed; if not, module M13 is executed;

[0028] Module M4: The process robot calls the text detection program to preprocess the airline ticket photos that meet the standards in module M3 to determine the text area;

[0029] Module M5: After module M4, the process robot calls the text recognition program. The text recognition program uses a single-vision text recognition algorithm, introduces a pre-trained model based on the self-attention mechanism to mine context information in downstream tasks, and improves the text recognition ability; through the recognition of the text area in module M4 by this program, the recognized text, credibility, and the coordinate values of its range are obtained;

[0030] Module M6: The process robot calls the feature extraction program, which can determine the position of the feature in the original image according to the text coordinate values and the original image information obtained in module M5, so as to realize feature extraction;

[0031] Module M7: After module M6, the process robot screens the feature recognition results;

[0032] Module M8: After module M7, the process robot logs in to the TravelSky website and calls the text recognition program to recognize the arithmetic verification code on the website;

[0033] Module M9: After module M8, the process robot calls the text processing program, which can traverse the elements to judge whether they contain operators and convert the string into an arithmetic expression to realize the processing of the four arithmetic operations in the text, and return the calculation result;

[0034] Module M10: After module M9, the process robot automatically enters the TravelSky website according to the feature information obtained in module M7 and the verification code calculation result obtained in module M9, and starts the verification;

[0035] Module M11: After the process robot verification is completed, determine whether the verification result of the airline ticket is true;

[0036] Module M12: The process robot locates and crops the screenshot obtained in Module M11, retains the electronic airline ticket information, and saves the cropping result to the specified path;

[0037] Module M13: For relevant situations including: unqualified airline ticket photos, failed text recognition, and failed verification, the process robot records relevant information including file name, verification result, and failure reason in the specified file;

[0038] Module M14: The process robot determines whether there are still unprocessed airline ticket photos. If so, return to Module M2. If not, execute Module M15;

[0039] Module M15: After the process robot completes the work, organize the airline ticket verification results, and the airline ticket verification process ends.

[0040] Preferably, the text detection program in Module M4 is based on the differentiable binarization algorithm and uses the principle of image segmentation to realize text detection in related multi-shape and multi-direction scenarios including bending, twisting, and tilting.

[0041] Preferably, in Module M7, the process robot screens the element recognition results. If the passenger name, itinerary number, e-ticket number, and total price have all been recognized and the confidence level is greater than the threshold, execute Module M8. Otherwise, execute Module M13.

[0042] Preferably, after the process robot verification in Module M11 is completed, if the airline ticket verification result is true, take a full-screen screenshot of the page returned by the TravelSky and record information such as the ticket file name and verification result in the specified file, and enter Module M12. If the ticket verification result is false, enter Module M13.

[0043] In a third aspect, a device is provided, and the device includes:

[0044] One or more processors;

[0045] A storage device for storing one or more programs,

[0046] When the one or more programs are executed by the one or more processors, the one or more processors implement the steps in the method.

[0047] In a fourth aspect, a computer-readable storage medium storing a computer program is provided, and when the computer program is executed by a processor, the steps in the method are implemented.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] 1. The present invention can quickly locate and identify the four elements for verifying airline tickets based on airline ticket photos, query the authenticity of airline tickets, and ensure the timeliness and accuracy of verifying airline ticket itineraries;

[0050] 2. The present invention can achieve fully automated operation without interruption throughout the process, save labor costs and have high versatility. At the same time, it can avoid human operation errors and reduce supervision risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Other features, objects, and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments with reference to the following drawings:

[0052] Figure 1 It is a schematic diagram of the overall process of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The present invention will be described in detail below with reference to specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but do not limit the present invention in any form. It should be noted that those of ordinary skill in the art can make several changes and improvements without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0054] An embodiment of the present invention provides a method for verifying airline ticket itineraries, which automates the business process of verifying airline ticket itineraries by adopting process robot technology and neural network technology, saves labor costs and has high versatility. At the same time, it can avoid human operation errors and reduce supervision risks. Refer to Figure 1 As shown, the method includes:

[0055] Step S1: Start the main program of the process robot and initialize the parameters.

[0056] Step S2: The process robot sequentially accesses airline ticket photos within a specified path.

[0057] Step S3: The process robot preliminarily determines whether the size of the airline ticket photo is less than 2M and the photo format is: PNG, JPG, PDF. If the conditions are met, step S4 is executed; if not, step S13 is executed.

[0058] Step S4: The process robot calls a text detection program to preprocess the airline ticket photos that meet the standards in step S3 to determine the text area. This text detection program is based on the DifferentiableBinarization algorithm and uses the principle of image segmentation to achieve text detection in multi-shaped and multi-directional scenarios such as bending, twisting, and tilting.

[0059] Step S5: After step S4, the process robot calls the text recognition program, which uses the single visual text recognition algorithm (Single Visual Text Recognition) to effectively mine contextual information and improve text recognition capabilities on downstream tasks by introducing a pre-trained model (BERT: Bidirectional Encoder Representation from Transformers) based on the self-attention mechanism (Transformer). Through the program's recognition of the text area in step S4, the recognized text, credibility (ranging between 0 and 1), and the coordinate values of the range in which it is located are obtained. Among them, the single visual text recognition algorithm adopted in this embodiment can optimize the text recognition results according to the contextual content. By introducing a model with a self-attention mechanism structure, it can more effectively mine the contextual information of the text image, avoid typos, semantic contradictions, etc., and improve text recognition capabilities. If replaced with other algorithms, it will affect the recognition accuracy, such as typos, incoherent context, etc.

[0060] Step S6: The process robot calls the feature extraction program, which can determine the position of the feature in the original image based on the text coordinate value and original image information obtained in step S5, thereby realizing feature extraction.

[0061] Step S7: After step S6, the process robot screens the element recognition results. If the passenger name, itinerary number, electronic ticket number, and total price have all been recognized and the credibility is greater than the threshold, step S8 is executed; otherwise, step S13 is executed.

[0062] Step S8: After step S7, the process robot logs in to the Xintianyou website (www.tarvelsky.com.cn) and calls the text recognition program to identify the website's arithmetic verification code.

[0063] Step S9: After step S8, the process robot calls the text processing program, which can traverse the elements to determine whether they contain operators and convert the character string into an equation, implement the four arithmetic operations in the text, and return the calculation results.

[0064] Step S10: After step S9, the process robot automatically enters the element information obtained in step S7 and the verification code calculation result obtained in step S9 into the Xintianyou website and starts verification.

[0065] Step S11: After the process robot finishes the verification, if the verification result of the airline ticket is true, take a full-screen screenshot of the page returned by TravelSky, and record information such as the ticket file name and verification result in a specified file, then enter Step S12. If the ticket verification result is false, enter Step S13.

[0066] Step S12: The process robot locates and crops the screenshot obtained in Step S11, only retaining the electronic airline ticket information, and saves the cropping result to a specified path.

[0067] Step S13: For situations including: unqualified airline ticket photos, failed text recognition, and failed verification, etc., the process robot records information such as the file name, verification result, and failure reason in a specified file respectively.

[0068] Step S14: The process robot determines whether there are still unprocessed airline ticket photos. If so, return to Step S2. If not, execute Step S15.

[0069] Step S15: After the process robot finishes its work, organize the airline ticket verification results, and the airline ticket verification process ends.

[0070] The present invention also provides an airline ticket itinerary single verification system, which specifically includes:

[0071] Module M1: Start the main program of the process robot and initialize the parameters.

[0072] Module M2: The process robot sequentially accesses the airline ticket photos in a specified path.

[0073] Module M3: The process robot preliminarily determines whether the size of the airline ticket photo is less than 2M and the photo format is: PNG, JPG, PDF. If the conditions are met, execute Module M4. If not, execute Module M13.

[0074] Module M4: The process robot calls a text detection program to preprocess the airline ticket photos that meet the standards in Module M3, so as to determine the text area. This text detection program is based on the DifferentiableBinarization algorithm and uses the principle of image segmentation to achieve text detection in multi-shaped and multi-directional scenarios such as bending, twisting, and tilting.

[0075] Module M5: After module M4, the process robot calls the text recognition program. This text recognition program uses the Single Visual Text Recognition algorithm (Single Visual Text Recognition) to effectively mine contextual information and improve text recognition capabilities in downstream tasks by introducing a pre-trained model (BERT: Bidirectional Encoder Representation from Transformers) based on the self-attention mechanism (Transformer). Through this program, the text area in module M4 is recognized and the recognized text, credibility (ranging between 0 and 1), and the coordinate values of the range in which it is located are obtained.

[0076] Module M6: The process robot calls the feature extraction program, which can determine the position of the feature in the original image based on the text coordinate value and original image information obtained by module M5, thereby realizing feature extraction.

[0077] Module M7: After module M6, the process robot screens the element recognition results. If the passenger name, itinerary number, electronic ticket number, and total price have all been recognized and the credibility is greater than the threshold, module M8 is executed; otherwise, module M13 is executed.

[0078] Module M8: After module M7, the process robot logs in to the Xintianyou website (www.tarvelsky.com.cn) and calls the text recognition program to identify the website's arithmetic verification code.

[0079] Module M9: After module M8, the process robot calls the text processing program, which can traverse the elements to determine whether they contain operators and convert the string into an equation, implement the four arithmetic operations in the text, and return the calculation results.

[0080] Module M10: After module M9, the process robot automatically enters the factor information obtained in module M7 and the verification code calculation result obtained in module M9 into the Xintianyou website and starts verification.

[0081] Module M11: After the process robot completes the verification, if the airline ticket verification result is true, a full-screen screenshot of the page returned by Xintianyou is taken, and the ticket file name, verification result and other information are recorded in the designated file, and the process goes to step 12. If the ticket verification result is false, the process goes to module M13.

[0082] Module M12: The process robot locates and crops the screenshot obtained in module M11, retaining only the electronic version of the airline ticket information, and saves the cropping result to the specified path.

[0083] Module M13: The process robot records information such as the file name, verification result, and failure reason in a specified file for situations including unqualified airline ticket photos, failed text recognition, and failed verification, etc.

[0084] Module M14: The process robot determines whether there are still unprocessed airline ticket photos. If so, it returns to Module M2; if not, it executes Module M15.

[0085] Module M15: After the process robot finishes its work, it collates the verification results of airline tickets, and the airline ticket verification process ends.

[0086] The embodiments of the present invention provide a method, system, device, and medium for verifying the itinerary of airline tickets. By using process robot technology and neural network technology, it realizes the full automation of the business process for verifying the itinerary of airline tickets, avoiding situations such as false reimbursement caused by "fake tickets" due to human errors or malicious operations and wasting manpower for verification; ensuring the timeliness of airline ticket verification, and at the same time using intelligent process robots to reduce labor costs, avoid repeated operations, and improve verification efficiency. Business personnel can start the intelligent process robot program through desktop software interaction. This program can quickly locate and identify the four elements for verifying airline tickets (passenger name, itinerary number, electronic ticket number, total price) based on the airline ticket photo, automatically log in to the Ctrip ticket query and verification website to confirm the authenticity of the airline ticket, and record the verification result.

[0087] Those skilled in the art know that in addition to implementing the system and its various devices, modules, and units provided by the present invention in the form of pure computer-readable program code, the method steps can be logically programmed to enable the system and its various devices, modules, and units provided by the present invention to be implemented in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers, etc., to achieve the same functions. Therefore, the system and its various devices, modules, and units provided by the present invention can be regarded as a kind of hardware component, and the devices, modules, and units included therein for implementing various functions can also be regarded as the structure within the hardware component; the devices, modules, and units for implementing various functions can also be regarded as both software modules for implementing the method and the structure within the hardware component.

[0088] The specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the above specific implementation manners. Those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. Without conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. An aviation ticket itinerary single verification method, characterized in that, Including: Step S1: Start the main program of the process robot and initialize the parameters; Step S2: The process robot sequentially accesses the airline ticket photos within the specified path; Step S3: The process robot preliminarily judges whether the size of the airline ticket photo meets the set requirements. If the conditions are met, execute Step S4; otherwise, execute Step S13; Step S4: The process robot calls the text detection program to preprocess the airline ticket photos that meet the standards in Step S3 to determine the text area; Step S5: After Step S4, the process robot calls the text recognition program. The text recognition program uses a single-vision text recognition algorithm, introduces a pre-trained model based on the self-attention mechanism to mine context information in downstream tasks, and improves the text recognition ability; through the recognition of the text area in Step S4 by this program, the recognized text, credibility, and coordinate values of its location range are obtained; Step S6: The process robot calls the element extraction program, which can determine the position of the element in the original image according to the text coordinate values and the original image information obtained in Step S5, so as to achieve element extraction; Step S7: After Step S6, the process robot filters the element recognition results; Step S8: After Step S7, the process robot logs in to the TravelSky website and calls the text recognition program to recognize the arithmetic verification code on the website; Step S9: After Step S8, the process robot calls the text processing program, which can traverse the elements to judge whether they contain operators and convert the string into an arithmetic expression to implement the processing of the four arithmetic operations in the text and return the calculation result; Step S10: After Step S9, the process robot automatically enters the TravelSky website according to the element information obtained in Step S7 and the verification code calculation result obtained in Step S9 to start the verification; Step S11: After the process robot finishes the verification, judge whether the airline ticket verification result is true; Step S12: The process robot locates and crops the screenshot obtained in Step S11, retains the electronic airline ticket information, and saves the cropping result to the specified path; Step S13: The process robot records relevant information including the file name, verification result, and failure reason in a specified file for relevant situations including unqualified airline ticket photos, failed text recognition, and failed verification; Step S14: The process robot judges whether there are still unprocessed airline ticket photos. If so, return to Step S2; otherwise, execute Step S15; Step S15: After the process robot finishes its work, it sorts out the airline ticket verification results, and the airline ticket verification process ends.

2. The method for verifying an itinerary of an airline ticket according to claim 1, wherein The text detection program in Step S4 is based on the differentiable binarization algorithm and uses the principle of image segmentation to realize text detection in related multi-shape and multi-direction scenarios including bending, twisting, and tilting.

3. The air ticket itinerary verification method according to claim 1, characterized in that In Step S7, the process robot filters the element recognition results. If the passenger name, itinerary number, electronic ticket number, and total price have all been recognized and the credibility is greater than the threshold, execute Step S8; otherwise, execute Step S13.

4. The method for verifying an itinerary receipt of an airline ticket according to claim 1, wherein After the process robot verification in step S11 is completed, if the verification result of the airline ticket is true, take a full-screen screenshot of the page returned by TravelSky, and record information such as the ticket file name and verification result in a specified file, then enter step S12. If the ticket verification result is false, enter step S13.

5. An air ticket itinerary single verification system, characterized in that, Including: Module M1: Start the main program of the process robot and initialize the parameters; Module M2: The process robot sequentially accesses the airline ticket photos in the specified path; Module M3: The process robot preliminarily judges whether the size of the airline ticket photo meets the set requirements. If the condition is met, execute Module M4. If not, execute Module M13; Module M4: The process robot calls the text detection program to preprocess the airline ticket photos that meet the standards in Module M3 to determine the text area; Module M5: After Module M4, the process robot calls the text recognition program. The text recognition program uses a single-vision text recognition algorithm, introduces a pre-trained model based on the self-attention mechanism to mine context information in downstream tasks, and improves the text recognition ability; through the recognition of the text area in Module M4 by this program, the recognized text, credibility, and coordinate values of its location range are obtained; Module M6: The process robot calls the feature extraction program, which can determine the position of the feature in the original image according to the text coordinate values and the original image information obtained in Module M5, so as to realize feature extraction; Module M7: After Module M6, the process robot screens the feature recognition results; Module M8: After Module M7, the process robot logs in to the TravelSky website and calls the text recognition program to recognize the arithmetic verification code on the website; Module M9: After Module M8, the process robot calls the text processing program, which can traverse the elements to judge whether they contain operators and convert the string into an arithmetic expression to realize the processing of the four arithmetic operations in the text, and return the calculation result; Module M10: After Module M9, the process robot automatically enters the TravelSky website according to the feature information obtained in Module M7 and the verification code calculation result obtained in Module M9 to start the verification; Module M11: After the process robot verification is completed, judge whether the verification result of the airline ticket is true; Module M12: The process robot locates and crops the screenshot obtained in Module M11, retains the electronic version of the airline ticket information, and saves the cropping result to the specified path; Module M13: The process robot records relevant information including file name, verification result, and failure reason in a specified file for relevant situations including unqualified airline ticket photos, failed text recognition, and failed verification; Module M14: The process robot judges whether there are still unprocessed airline ticket photos. If so, return to Module M2. If not, execute Module M15; Module M15: After the process robot completes the work, sort out the verification results of the airline tickets, and the airline ticket verification process ends.

6. The air ticket itinerary verification system according to claim 5, characterized in that, The text detection program in Module M4 is based on the differentiable binarization algorithm and uses the principle of image segmentation to realize text detection in related multi-shape and multi-direction scenarios including bending, twisting, and tilting.

7. The air ticket itinerary verification system according to claim 5, wherein In the module M7, the process robot screens the element recognition results. If the passenger name, itinerary number, electronic ticket number, and total price have all been recognized and the confidence level is greater than the threshold, module M8 is executed; otherwise, module M13 is executed.

8. The air ticket itinerary verification system according to claim 5, characterized in that, After the process robot in the module M11 finishes verification, if the air ticket verification result is true, a full-screen screenshot of the page returned by TravelSky is taken, and information such as the air ticket file name and verification result is recorded in the specified file, and then module M12 is entered. If the air ticket verification result is false, module M13 is entered.

9. A device, characterized in that, The device includes: One or more processors; A storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the method according to any one of claims 1 to 4.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images