Scrapped vehicle information rapid input system based on high-speed photographic apparatus image recognition technology

Abstract

The invention discloses a scrapped vehicle information rapid input system based on a high-speed photographic apparatus image recognition technology, and belongs to the field of image recognition. The system comprises an image acquisition and uploading module, an image identification and information extraction module, a data matching and verification module, an information automatic input and storage module and an accessory management and tracing module. The problems that existing scrapped vehicle information input depends on manual work, efficiency is low, errors are prone to occurring, and an effective verification and tracing mechanism is lacked are solved, an image collecting and uploading module and an image recognition and information extraction module integrate high-speed photographic instrument hardware and OCR recognition and an image preprocessing algorithm, and the accuracy of information input is improved. Batch acquisition of multiple types of certificate images and automatic extraction of key information are realized, automatic matching and consistency guarantee of data are realized through a data matching and verification module and an attachment management and tracing module, and the traceability of data management is also improved.

Description

Technical Field

[0001] This invention relates to the field of image recognition technology, specifically to a rapid data entry system for scrapped vehicles based on high-speed document scanner image recognition technology. Background Technology

[0002] In the scrap vehicle recycling business, traditional information entry methods rely on manually filling out paper forms or manually inputting data into electronic systems, which are inefficient, error-prone, and involve repetitive work. Especially in the vehicle information entry stage, information such as driver's licenses, ID cards, and bank cards needs to be entered multiple times, making the process cumbersome and prone to inaccuracies or omissions due to human error.

[0003] While some existing image recognition systems can assist in recognizing document information, most systems have limited recognition accuracy, do not support batch processing, and have low coupling with business systems, making it impossible to achieve automated matching and verification. Therefore, they do not meet the current needs. To address this, we propose a rapid data entry system for scrapped vehicles based on high-speed document scanner image recognition technology. Summary of the Invention

[0004] The purpose of this invention is to provide a rapid data entry system for scrapped vehicles based on high-speed document scanner image recognition technology. By introducing high-speed document scanner image acquisition and OCR recognition technology, the system achieves rapid, batch, and automated data entry, significantly improving data accuracy and operational efficiency. Furthermore, the system supports recognition of multiple types of documents, batch processing, intelligent matching, and verification, thus solving the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a rapid data entry system for scrapped vehicles based on high-speed document scanner image recognition technology, comprising:

[0006] The image acquisition and upload module is used to acquire images of various document types using a high-speed document scanner, and provides a unified image upload portal. It supports single or batch uploads and performs format verification and compression processing before image upload.

[0007] Among them, various document types include ID cards, vehicle registration certificates, and bank cards;

[0008] The image recognition and information extraction module is used to automatically identify key information such as license plate number, vehicle owner's name, ID number and vehicle model in the image using OCR technology, and to evaluate and enhance the image quality through image preprocessing algorithms and machine learning models.

[0009] The image preprocessing algorithms include grayscale conversion, binarization, and tilt correction.

[0010] The data matching and verification module is used to match the identified information with the existing vehicle, order and user data in the system. The data consistency is verified through the rule engine. If the matching fails or the information is abnormal, manual intervention or re-photographing is prompted.

[0011] The automatic information entry and storage module is used to automatically fill the corresponding form fields of the vehicle information table, order table and user table with the recognized information, and uses encrypted transmission and blockchain technology to store the data. It also supports data write-back and status synchronization.

[0012] The attachment management and traceability module is used to classify and store all uploaded attachments by type and associate them with vehicles, orders and users. It also provides attachment viewing, replacement, download and one-click package download functions, and records each attachment upload, modification and deletion behavior through operation log.

[0013] Furthermore, in the image recognition and information extraction module, the document scanner integrates a collaborative preprocessing module, which is used to perform dynamic exposure and white balance calibration, and automatically adjust hardware parameters according to different document materials, as well as to perform multi-frame fusion noise reduction. By continuously capturing multiple frames of images and performing pixel-level variance analysis and weighted fusion, character contrast is improved and interference is reduced.

[0014] Furthermore, the image recognition and information extraction module integrates a key field structured positioning enhancement mechanism, including a dynamic template matching unit for dynamically matching standard templates according to document type and performing deformation adaptation, and a multimodal field recognition fusion unit for performing multi-model recognition and cross-validation on mixed fields.

[0015] Furthermore, the dynamic template matching unit performs tilt correction and coordinate calibration on key fields of various document types, including vehicle registration certificates and vehicle licenses, through dynamic template matching and deformation adaptive mechanisms. The multimodal field recognition and fusion unit calls character recognition models, symbol recognition models and texture recognition models to perform multi-model recognition and cross-validation for the mixed field of VIN code and engine number.

[0016] Furthermore, the data matching and verification module further includes a VIN code verification and error correction enhancement unit, which is used to perform segmented semantic verification and fuzzy recognition error correction suggestions on the VIN code, and locate abnormal segments or provide similar character replacement suggestions when verification fails.

[0017] Furthermore, the data matching and verification module performs data consistency verification through a rule engine, including multi-dimensional consistency verification of the vehicle owner's identity, linkage verification of the certificate validity period, and multi-factor duplicate judgment of vehicle uniqueness and status, as well as multi-rule prediction of scrap status.

[0018] Furthermore, the data matching and verification module also includes a business process and information verification collaboration mechanism, which is used to dynamically control process nodes based on the verification results, including automatic advancement, suspension of manual review, or triggering of high-risk work orders.

[0019] Furthermore, the attachment management and traceability module further includes:

[0020] The intelligent classification and archiving unit is used to automatically classify attachments based on text and image features, segment and archive mixed scans independently, and support the judgment of the association and reuse of historical attachments. Based on the evaluation results of document validity period and image clarity, it prompts whether historical attachments can be reused.

[0021] The phased attachment inspection unit is used to dynamically adjust the attachment inspection rules according to the business stage, and to provide intelligent guidance and supplementary entry assistance when attachments are missing or of poor quality.

[0022] The multi-dimensional index and version management unit supports combined searches by attachment type, creation time, and operator, and retains all attachment versions and their operation records.

[0023] Compared with the prior art, the beneficial effects of the present invention are:

[0024] This invention integrates high-speed document scanner hardware with OCR recognition and image preprocessing algorithms to achieve batch acquisition of multiple types of document images and automatic extraction of key information. This avoids the inefficiency and error-prone nature of traditional manual data entry methods, thereby improving the accuracy and efficiency of information entry in the scrap vehicle recycling business. By setting up data matching and verification modules and attachment management and traceability modules, an automatic information verification, process collaborative control, and full-link attachment traceability mechanism is constructed. This not only enhances the coupling between the system and the business, achieving automated data matching and consistency assurance, but also improves the standardization and traceability of data management through intelligent classification archiving and version management. It provides an efficient and reliable integrated information entry solution for the scrap vehicle recycling business. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the structure of the scrapped vehicle information rapid entry system based on high-speed document scanner image recognition technology of the present invention. Detailed Implementation

[0026] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] To address the technical problems in the current vehicle scrapping operation, such as reliance on manual data entry, low efficiency, high error rates, and lack of effective verification and traceability mechanisms, please refer to [link / reference]. Figure 1 This embodiment provides the following technical solution:

[0028] A rapid information entry system for scrapped vehicles based on high-speed document scanner image recognition technology includes:

[0029] The image acquisition and upload module is used to acquire images of various document types using a high-speed document scanner, and provides a unified image upload portal. It supports single or batch uploads, with a maximum of 12 images uploaded at once, and performs format verification and compression processing before image upload.

[0030] Among them, various document types include ID cards, vehicle registration certificates, and bank cards;

[0031] The image recognition and information extraction module is used to automatically identify key information such as license plate number, vehicle owner's name, ID number and vehicle model in an image using OCR technology, and to evaluate and enhance the image quality through image preprocessing algorithms and machine learning models. The implementation process of the image preprocessing algorithms and machine learning models is existing technology in this field and is not an inventive solution of this application, and will not be described in detail here.

[0032] The image preprocessing algorithms include grayscale conversion, binarization, and tilt correction.

[0033] The data matching and verification module is used to match the identified information with the existing vehicle, order and user data in the system. It performs data consistency verification through the rule engine, such as the consistency verification of the name on the ID card and the bank card. If the matching fails or the information is abnormal, it prompts manual intervention or re-photographing.

[0034] The automatic information entry and storage module is used to automatically fill the corresponding form fields of the vehicle information table, order table and user table with the recognized information, and uses encrypted transmission and blockchain technology to store the data to ensure that the data cannot be tampered with. It also supports data write-back and status synchronization, such as updating the vehicle status to "recognized".

[0035] The attachment management and traceability module is used to classify and store all uploaded attachments by type and associate them with vehicles, orders and users. It also provides attachment viewing, replacement, downloading and one-click package download functions, and records each attachment upload, modification and deletion behavior through operation log to ensure that the operation is traceable.

[0036] In this embodiment, a table is used as an example of a specific implementation method. The table design closely revolves around the core logic of the "vehicle-certificate-person" association in the scrap vehicle recycling business, ensuring that the identified data can be stored in a structured manner and serve subsequent business processes.

[0037] Table 1. Vehicle Main Table (scrap_vehicle_main)

[0038] Column names name Data types Is this field required? describe VEHICLE_ID Vehicle key VARCHAR(32) yes Primary key, a unique identifier within the system. LICENSE_PLATE_NO license plate number VARCHAR(20) yes OCR recognition results, business unique key VIN Vehicle Identification Number VARCHAR(50) yes OCR recognition result, unique identifier ENGINE_NO Engine number VARCHAR(50) yes OCR recognition results VEHICLE_TYPE Vehicle type VARCHAR(50) yes The identified vehicle type, such as "small car". VEHICLE_BRAND Vehicle Brand VARCHAR(100) yes For example, "BYD" VEHICLE_MODEL Vehicle model VARCHAR(100) yes For example, "Qin EV" REGISTER_DATE Registration Date DATE yes Vehicle registration certificate recognition ISSUE_DATE Date of issuance DATE yes Vehicle registration certificate recognition OWNER_NAME Car owner's name VARCHAR(100) yes Cross-verification with ID card and bank card OWNER_ID_CARD Vehicle owner's ID number VARCHAR(20) yes Encrypted storage POWER_TYPE Power type VARCHAR(10) yes The system automatically determines the license plate number based on its digits: 7 digits - gasoline vehicles, 8 digits - electric vehicles. NET_WEIGHT curb weight DECIMAL(10,2) no Vehicle registration certificate recognition, used for valuation reference. STATUS Vehicle status VARCHAR(2) yes 01 - Pending Inspection, 02 - Inspected... Synchronized with business process status. CREATE_TIME Creation time DATETIME yes Data creation time UPDATE_TIME Update time DATETIME yes Last updated

[0039] As shown in Table 1, this table is the core table for vehicle information, storing standardized vehicle data that has been recognized and verified through image recognition.

[0040] Table 2 Vehicle Attachments

[0041] Column names name Data types Is this field required? describe ATTACHMENT_ID Attachment ID VARCHAR(32) yes primary key VEHICLE_ID Vehicle ID VARCHAR(32) yes Foreign key, related to the vehicle main table ATTACHMENT_TYPE Attachment type VARCHAR(20) yes 01-Vehicle registration certificate, 02-Owner's ID card, 03-Bank card, 04-Registration certificate, 05-Recycling certificate, 06-Cancellation certificate FILE_PATH File storage path VARCHAR(500) yes cloud storage path OCR_RAW_TEXT OCR Original Text TEXT no Store the complete original text recognized by OCR for tracing OCR_STATUS Identification status VARCHAR(2) yes 00 - Recognition in progress, 01 - Recognition successful, 02 - Recognition failed, 03 - Pending manual review UPLOAD_USER Uploader VARCHAR(32) yes Operator ID UPLOAD_TIME Upload time DATETIME yes IS_CURRENT Is it currently valid? CHAR(1) yes 1 - Yes, 0 - Previous version (e.g., when updating the vehicle registration certificate)

[0042] As shown in Table 2, this table manages all vehicle-related documents and images collected by the document scanner, and supports batch uploading and version management.

[0043] Table 3 Image Recognition Log (ocr_process_log)

[0044] Column names name Data types Is this field required? describe LOG_ID Log ID VARCHAR(32) yes primary key ATTACHMENT_ID Attachment ID VARCHAR(32) yes Foreign Key RECOGNIZED_FIELD Identify field name VARCHAR(50) yes For example, "license_plate_no", "vin" ORIGINAL_VALUE Identify raw values VARCHAR(500) yes OCR directly outputs the results. CORRECTED_VALUE Corrected value VARCHAR(500) no The value after fuzzy matching or manual correction CONFIDENCE_SCORE Confidence DECIMAL(5,4) yes The confidence level returned by the recognition engine is between 0 and 1. MATCH_STATUS Matching status VARCHAR(2) yes 01 - Automatic matching successful; 02 - Fuzzy matching recommendation; 03 - Matching failed. PROCESS_TIME Processing time DATETIME yes

[0045] As shown in Table 3, this table is used to record the detailed process of each image recognition, which facilitates problem tracing and algorithm optimization.

[0046] The technical effects of the above solution are as follows: The image acquisition and upload module, through a unified image upload portal, can acquire images of various document types such as ID cards, vehicle registration certificates, and bank cards, either individually or in batches. Before uploading, it performs format verification and compression processing, ensuring the standardization of image input and optimizing system resources. The image recognition and information extraction module utilizes OCR technology combining image preprocessing algorithms such as grayscale, binarization, and tilt correction. The system can automatically identify and extract key information such as license plate numbers, vehicle owner names, ID card numbers, and vehicle models. Simultaneously, by integrating machine learning models, it intelligently evaluates and enhances image quality, effectively improving the recognition success rate and data reliability in complex scenarios. The data matching and verification module uses a rule engine to match the recognized information with existing vehicle and order information in the system. The system automatically matches and verifies the consistency of user data, enabling timely manual intervention when anomalies are detected. This ensures data integrity and compliance with business rules. The automatic information entry and storage module further automatically populates verified information into relevant business forms, employing encrypted transmission and blockchain storage technology to ensure data security and immutability. It also updates business status in real time through data write-back and status synchronization. The attachment management and traceability module categorizes, manages, and facilitates the retrieval of all uploaded files, supporting operations such as viewing, replacing, downloading, and one-click packaging. Combined with full-process operation logs, it ensures the traceability of every attachment operation, thereby enhancing operational convenience while strengthening the system's security audit and risk control capabilities.

[0047] In the image recognition and information extraction module, the document scanner integrates a collaborative preprocessing module for performing dynamic exposure and white balance calibration, automatically adjusting hardware parameters according to different document materials, and performing multi-frame fusion noise reduction. This involves continuously capturing multiple frames of images and performing pixel-level variance analysis and weighted fusion to improve character contrast and reduce interference. Specifically:

[0048] Dynamic exposure and white balance calibration: Real-time reading of the document scanner's hardware parameters (such as current exposure value and white balance mode) and automatic parameter adjustment instructions for different document materials (vehicle registration certificate with reflective plastic seal, ID card with matte coating, and registration certificate with paper texture); In this embodiment, when capturing a vehicle registration certificate, "low exposure + cool-toned white balance" is triggered to suppress reflective plastic seal; when capturing an ID card, "medium exposure + neutral white balance" is enabled to preserve the skin tone and national emblem color details, solving the problems of "overexposure of reflective areas" or "blurred shadows" under traditional fixed parameters;

[0049] Multi-frame fusion denoising technology: Addressing the issue of static image capture by document scanners being susceptible to ambient light fluctuations, the algorithm triggers the scanner to continuously capture 3-5 frames. Abnormal frames (such as those with momentary shadows or dust occlusion) are removed through pixel-level variance analysis. The remaining valid frames are then weighted and fused. During the fusion process, emphasis is placed on preserving text edges and texture features, improving character grayscale contrast by over 20% and effectively reducing interference from paper wrinkles and stains on recognition.

[0050] The technical effects of the above solution are as follows: By using dynamic exposure and white balance calibration functions, the hardware parameters of the document scanner are read and intelligently adjusted in real time, thereby accurately addressing the challenges posed by the characteristics of different document materials. This fundamentally solves image quality problems such as overexposure due to reflections or blurring in dark areas caused by fixed parameters, providing a clear and accurate image foundation for subsequent OCR recognition. Simultaneously, multi-frame fusion denoising technology is introduced. By continuously capturing multiple frames and performing pixel-level variance analysis, abnormal frames caused by instantaneous shadows or dust occlusion can be intelligently removed. The remaining valid frames are then weighted and fused, focusing on text edges and texture features. This not only significantly improves the grayscale contrast of characters by more than 20%, greatly enhancing the recognizability of character areas, but also effectively overcomes common interferences such as ambient light fluctuations, paper wrinkles, and stains, significantly reducing the impact of noise on recognition accuracy. This overall improves the accuracy of information extraction and the stability of the system in complex environments.

[0051] The image recognition and information extraction module integrates a key field structured localization enhancement mechanism, including a dynamic template matching unit for dynamically matching standard templates according to document type and performing deformation adaptation, and a multimodal field recognition fusion unit for performing multi-model recognition and cross-validation of mixed fields.

[0052] The dynamic template matching unit, through dynamic template matching and deformation adaptive mechanisms, performs tilt correction and coordinate calibration on key fields of various document types, including vehicle registration certificates and licenses. Specifically:

[0053] A "document type-standard template library" (containing the layout features of different versions of vehicle registration certificates and vehicle licenses) is pre-established. After image acquisition, the actual placement angle and scaling ratio of the document are determined by edge contour detection and corner feature extraction (such as the anti-counterfeiting marks at the four corners of the vehicle registration certificate and the outline of the national emblem on the vehicle license and the vehicle license). Then, the coordinates of key fields are dynamically calibrated. In this embodiment, when the document is tilted within 15°, the rotation matrix is ​​automatically calculated to correct the field area to avoid "field omission recognition" caused by placement offset. When the document has slight wrinkles (deformation rate ≤ 5%), the field area is locally stretched and corrected by the elastic deformation matching algorithm to ensure complete character coverage.

[0054] The multimodal field recognition and fusion unit, for the mixed field of VIN code and engine number, calls character recognition model, symbol recognition model and texture recognition model to perform multi-model recognition and cross-validation, specifically:

[0055] For mixed fields such as VIN codes and engine numbers that combine characters, numbers, and special symbols, a multi-model voting mechanism is added: simultaneously calling a character recognition model (recognizing letters / numbers), a symbol recognition model (recognizing special characters such as "-" and "*" in the VIN code), and a texture recognition model (recognizing local texture features of worn characters), and cross-validating the recognition results of the three models; in this embodiment, the 10th digit of the VIN code is the year code (e.g., "L" represents 2020). If the character model recognizes "1" and the texture model matches the arc-shaped texture feature of "L", then the result of the texture model is taken as the standard, increasing the recognition accuracy from 92% of a single model to over 99.5%.

[0056] The technical effects of the above solution are as follows: The dynamic template matching unit can intelligently call the corresponding standard template according to different document types, and automatically correct the tilt and perspective distortion caused by the shooting angle using deformation adaptive technology. It can achieve accurate positioning and coordinate calibration of key areas of various types of documents such as vehicle registration certificates and vehicle licenses, laying a stable structural foundation for subsequent information extraction. At the same time, for mixed fields such as VIN codes and engine numbers composed of characters and special symbols, the system uses a multimodal field recognition fusion unit to collaboratively call multiple dedicated models such as character recognition, symbol recognition and texture recognition for parallel analysis and cross-validation. This effectively overcomes the limitations of a single model in the case of font wear, background interference or variable symbol shapes, thus forming complementary advantages at the model level, thereby greatly improving the recognition accuracy of complex mixed fields and the anti-interference ability of the overall system.

[0057] The data matching and verification module further includes a VIN code verification and error correction enhancement unit, which performs segmented semantic verification and fuzzy recognition error correction suggestions on the VIN code, and locates abnormal segments or provides similar character replacement suggestions when verification fails. In this embodiment, the 17-digit VIN code is split into "World Manufacturer Identifier (digits 1-3)", "Vehicle Description Part (digits 4-9)" and "Vehicle Indication Part (digits 10-17)" according to the ISO standard, and is matched with the built-in "Global Automaker Code Library" (e.g., "LZW" corresponds to Liuzhou Wuling, "WBA" corresponds to BMW) and "Vehicle Model Feature Code Rule Library" (e.g., the 4th digit "S" represents a sedan, "T" represents a truck). If the verification fails, the system not only prompts an error, but also locates the specific abnormal segment (e.g., "digits 1-3 did not match a valid automaker code, which may be an identification error"), reducing the difficulty of manual verification.

[0058] The data matching and verification module also includes a fuzzy recognition error correction mechanism, which generates error correction suggestions based on the character similarity matrix to assist operators in correcting the recognition results. In this embodiment, for the worn VIN codes collected by the document scanner (such as the characters "8" and "B", "0" and "O" are easily confused), the algorithm integrates a character similarity matrix: calculates the texture similarity between each character in the recognition result and similar characters (such as "8" and "B" have a similarity of 75%, and "8" and "0" have a similarity of 30%). If the verification fails and the similarity of a certain character is ≥60%, an error correction suggestion is automatically generated (such as "the 5th digit of the VIN code is identified as '8', and the similarity with 'B' is 75%, it is recommended to check whether it is 'B'"), further shortening the error correction time.

[0059] The technical effects of the above solution are as follows: By introducing a VIN code verification and error correction enhancement unit and a fuzzy recognition error correction mechanism, the VIN code is subjected to refined segmented semantic verification. This allows for the rapid location of specific abnormal segments that do not conform to the encoding rules or logic, rather than simply reporting errors. This greatly shortens the troubleshooting time and significantly improves the accuracy of data entry and the system's intelligent processing capabilities when dealing with complex recognition errors. When verification fails, the system does not stop at error messages but further activates a fuzzy recognition error correction mechanism based on a character similarity matrix. By analyzing common OCR recognition errors, it can intelligently generate high-probability similar character replacement suggestions and present them to the operator. This effectively transforms the tedious process that originally required manual full verification or re-photographing into an efficient and targeted auxiliary correction, thereby significantly reducing the operator's workload and the risk of human error.

[0060] In the data matching and verification module, data consistency verification is performed through a rule engine, including multi-dimensional consistency verification of vehicle owner identity, linkage verification of document validity period, and multi-factor duplicate judgment of vehicle uniqueness and status, and multi-rule prediction of scrap status.

[0061] In this embodiment, the vehicle owner's identity is verified using a multi-dimensional consistency check: going beyond the original single dimension of "name matching," it supplements the verification with ID number format and bank card opening bank association, specifically:

[0062] The ID card recognition results are verified for format legality using an 18-digit ID card number check code algorithm to eliminate basic issues such as "incorrect number of digits" and "incorrect check digit".

[0063] By linking to the bank interface, the system identifies the bank opening information based on the bank card number. If the "place of household registration" on the ID card is inconsistent with the "province and city where the bank is opened" on the bank card (e.g., the ID card is from "Beijing" but the bank card was opened in "Shenzhen, Guangdong"), a "low-risk alarm" is triggered (prompting "there is a difference in the region of identity information, it is recommended to confirm whether the car owner applied for the card in a different location"), rather than directly blocking the process, thus balancing risk control and business flexibility.

[0064] In this embodiment, the document validity period is linked for verification: a new "document validity period - business scenario" association rule is added, specifically as follows:

[0065] If the ID card is identified as "valid until May 2023", but the current business transaction date is January 2024, then the ID card is deemed "expired" and the process is blocked.

[0066] If the vehicle registration certificate states "inspection validity until December 2023", and the vehicle registration date is more than 10 years old (requiring annual inspection), then the system will prompt "the vehicle registration certificate has expired inspection validity; please confirm whether the vehicle meets the pre-scrapping conditions" to avoid invalid information entry.

[0067] In this embodiment, multi-factor duplicate determination of vehicle uniqueness and status and multi-rule prediction of scrap status are performed, specifically as follows:

[0068] Multi-factor duplicate detection algorithm: For special scenarios such as "license plate number change" and "VIN code wear and re-engraving", supplementary vehicle feature code auxiliary verification is added: the identified "brand and model", "body color" and "engine displacement" are combined and compared with the system's historical records; if "the license plate number is different but the VIN code is the same", it is directly determined to be duplicate; if "the license plate number and VIN code are different, but the brand and model, body color and engine displacement are completely the same and the registration date difference is ≤3 months", a "suspicious duplicate alarm" is triggered, prompting the operator to check the vehicle's appearance photos to prevent "cloned vehicles" and "fake vehicles" from entering the business process;

[0069] Multi-rule prediction model for scrap status: Breaking away from the original single rule of "registration date + scrap age," a multi-dimensional prediction model is constructed, including:

[0070] Integrated "mileage association rules" (if the "total mileage" on the vehicle registration certificate is ≥600,000 kilometers, even if the statutory scrapping age has not been reached, it will prompt "may meet the guided scrapping standards");

[0071] If the identified VIN code is marked as "severely damaged (unrepairable)" in the accident database, it will be directly determined that it "meets the scrapping standard".

[0072] For commercial vehicles, a mapping between "usage type - scrapping age" is added (e.g., commercial passenger vehicles are scrapped after 8 years, and non-commercial passenger vehicles are scrapped after 15 years) to achieve accurate prediction with "one policy for each vehicle".

[0073] The technical effects of the above solution are as follows: Multi-dimensional consistency verification of vehicle owner identity is achieved by cross-referencing names, numbers, and other information from multiple sources such as ID cards, vehicle registration certificates, and bank cards, effectively identifying and preventing the risk of identity theft or inconsistent information. Through the implementation of document validity period linkage verification, the logical relationship between the validity period of documents such as vehicle registration certificates and ID cards and the business processing time can be automatically detected, thus providing early warnings of expired documents and avoiding business process interruptions or compliance issues caused by document invalidation. Simultaneously, multi-factor duplicate determination of vehicle uniqueness is performed, combining VIN codes, engine numbers, and other identifiers to accurately identify and prevent the entry of duplicate vehicle information, ensuring the accuracy and uniqueness of database records. Furthermore, by performing multi-rule prediction of vehicle scrapping status, intelligently evaluating multiple factors such as vehicle age, traffic violation records, and accident history, vehicles meeting scrapping conditions can be screened in advance, providing reliable data decision support for subsequent processing. Based on the above series of data consistency verifications, the accuracy, security, and business compliance of the scrapped vehicle information entry process are significantly improved.

[0074] The data matching and verification module also includes a business process and information verification collaboration mechanism, which is used to dynamically control process nodes based on the verification results, including automatic advancement, suspension of manual review, or triggering of high-risk work orders.

[0075] In this embodiment, the verification result is deeply bound to the business process node to achieve "dynamic process guidance," specifically as follows:

[0076] If "the vehicle owner's identity is verified, the vehicle is not duplicated, and all attachments are complete", the process will automatically proceed to the "pending inspection" stage.

[0077] If "the vehicle owner identity verification has a low-risk alarm (such as regional differences in bank cards) + there are no abnormalities in the vehicle information", the process will be paused and a "manual review task" will be generated, which will be resumed after the operator confirms it.

[0078] If "vehicle duplicate entry + VIN code verification failure" occurs, the process will be frozen and a "high-risk work order" will be triggered. The order can only be unlocked after administrator approval to prevent unauthorized operations.

[0079] The technical effects of the above solution are as follows: the mechanism of directly linking verification results with process control not only greatly reduces the time delay of manual judgment and process switching, but also improves the overall operational efficiency. Through the deep collaboration mechanism between business processes and information verification, an intelligent, efficient and risk-controllable automated process management system is built. It can ensure the smooth and automatic operation of routine business while directing operational risks and potential errors to the most appropriate processing nodes, thereby significantly enhancing the reliability and risk defense capabilities of the entire system while improving the level of automation.

[0080] The attachment management and traceability module further includes:

[0081] The intelligent classification and archiving unit is used to automatically classify attachments based on text and image features, segment and archive mixed scans independently, and support the judgment of the association and reuse of historical attachments. Based on the evaluation results of document validity period and image clarity, it prompts whether historical attachments can be reused.

[0082] In this embodiment, attachments are automatically categorized across multiple scenarios: breaking through the original single logic of "collecting only by license plate number," intelligent categorization of attachment types is added, specifically:

[0083] Based on the dual recognition of "text features + image features", the attachments are classified into types (e.g., those containing the title text "Motor Vehicle Registration Certificate" + national emblem watermark are classified as "Registration Certificate", and those containing the red official seal of "Certificate of Recycling of Scrapped Vehicles" are classified as "Recycling Certificate").

[0084] For "mixed scanned documents" (such as a single page containing both ID card and vehicle registration certificate), a text block segmentation algorithm is used to identify the text boundaries of different documents, automatically crop them into independent attachments, and classify them separately, thus solving the pain point of "mixed attachments cannot be separated".

[0085] Historical attachment association and reuse mechanism: If there are historical attachments with the same license plate number in the system (such as scanned copies of ID cards retained from previous transactions), the algorithm determines whether they can be reused by comparing the validity period of the document and evaluating the image clarity. Specifically:

[0086] If the historical ID card is within its validity period and has a clarity score of ≥90 (out of 100, calculated based on edge sharpness), the system will prompt "The historical ID card attachment can be reused. Confirm?"

[0087] If the historical attachments have expired or are not clear enough, a "re-capture" prompt will be automatically triggered to reduce repeated scanning operations;

[0088] The phased attachment inspection unit is used to dynamically adjust the attachment inspection rules according to the business stage, and to provide intelligent guidance and supplementary entry assistance when attachments are missing or of poor quality.

[0089] In this embodiment, dynamic adaptation of business stage - attachment requirements: Based on business process nodes, the attachment inspection rules are dynamically adjusted to achieve "accurate verification by stage". Specifically:

[0090] When in the "information entry" stage, only check whether the "vehicle license + ID card" is complete;

[0091] When entering the "inspection confirmation" stage, automatically add the inspection requirements for "vehicle exterior photos + inspection reports";

[0092] When reaching the "cancellation and filing" stage, it is necessary to check the integrity of the "recovery certificate + cancellation certificate + tax payment certificate" to avoid "requiring attachments in advance" or "omitting key attachments";

[0093] The multi - dimensional index and version management unit is used to support combined retrieval by attachment type, creation time, and operator, and retain all attachment versions and their operation records;

[0094] In this embodiment, multi - dimensional index construction and fast retrieval: Based on the primary key of "license plate number", supplement the multi - dimensional index of attachment type - creation time - operator to support combined retrieval (such as "query the vehicle license attachment of 'Beijing A12345' entered by operator 'Zhang San' in March 2025"), and the retrieval response time is significantly shortened;

[0095] Attachment version management and traceability: For multiple supplementary recordings of the same attachment (such as the vehicle license being re - photographed 3 times), the algorithm automatically retains all versions and marks the "latest valid version". Specifically:

[0096] Each version is accompanied by "modification time + operator + reason for modification" (such as "October 15, 2025, 10:20, Zhang San re - photographed: The original attachment's national emblem was blurred");

[0097] When subsequent operations need to trace back, the modification records of all versions can be viewed to meet the requirements of compliance auditing.

[0098] The technical effects of the above solution are as follows: The intelligent classification and archiving unit automatically and accurately classifies attachments by utilizing text and image features, and can intelligently segment and independently archive mixed scanned documents, thereby improving the automation and standardization of attachment organization. Simultaneously, the integrated historical attachment association and reuse judgment function, by comprehensively evaluating document validity and image clarity, can intelligently suggest reusable historical attachments, effectively avoiding duplicate collection and improving business processing efficiency and user experience. The phased attachment inspection unit can dynamically adjust attachment inspection rules according to different stages of business processing, ensuring that attachment requirements always accurately match the business process. It also proactively provides intelligent guidance and supplementary entry assistance when attachments are missing or of poor quality, thus ensuring the integrity and quality of attachment materials from the source and reducing process interruptions caused by attachment issues. The multi-dimensional indexing and version management unit supports flexible combination retrieval based on attachment type, creation time, operator, and other conditions, greatly facilitating attachment location and auditing. By retaining all historical versions of attachments and their complete operation records, it ensures full traceability of every attachment upload, modification, or deletion, thereby comprehensively strengthening data security and operational compliance, providing a solid audit foundation for the system.

[0099] Working Principle: The system collects and preprocesses various document images through an image acquisition and uploading module. It also utilizes integrated hardware-algorithm collaborative preprocessing and multi-frame fusion denoising technology to improve image quality. The image recognition and information extraction module employs OCR technology combined with dynamic template matching and multimodal fusion mechanisms to accurately locate and identify key fields, effectively addressing deformed and mixed fields. The data matching and verification module uses a rule engine to perform multi-dimensional data consistency verification and VIN code error correction, and dynamically controls business processes based on verification results, ensuring data accuracy and process compliance. The automatic information entry and storage module automatically fills and encrypts verified data in the database, using blockchain technology to ensure data security and traceability. The attachment management and traceability module uses intelligent classification, phased checks, and multi-dimensional indexing to achieve standardized management and full lifecycle tracking of attachments. Based on this design, the system automates, improves accuracy, and enhances the intelligence of scrapped vehicle information entry, significantly improving entry efficiency and data quality while reducing manual intervention and error risks.

[0100] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0101] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention.

Claims

1. A rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology, characterized in that: include: The image acquisition and upload module is used to acquire images of various document types using a high-speed document scanner, and provides a unified image upload portal. It supports single or batch uploads and performs format verification and compression processing before image upload. Among them, various document types include ID cards, vehicle registration certificates, and bank cards; The image recognition and information extraction module is used to automatically identify key information such as license plate number, vehicle owner's name, ID number and vehicle model in the image using OCR technology, and to evaluate and enhance the image quality through image preprocessing algorithms and machine learning models. The image preprocessing algorithms include grayscale conversion, binarization, and tilt correction. The data matching and verification module is used to match the identified information with the existing vehicle, order and user data in the system. The data consistency is verified through the rule engine. If the matching fails or the information is abnormal, manual intervention or re-photographing is prompted. The automatic information entry and storage module is used to automatically fill the corresponding form fields of the vehicle information table, order table and user table with the recognized information, and uses encrypted transmission and blockchain technology to store the data. It also supports data write-back and status synchronization. The attachment management and traceability module is used to classify and store all uploaded attachments by type and associate them with vehicles, orders and users. It also provides attachment viewing, replacement, download and one-click package download functions, and records each attachment upload, modification and deletion behavior through operation log.

2. The rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology according to claim 1, characterized in that, In the image recognition and information extraction module, the document scanner integrates a collaborative preprocessing module, which is used to perform dynamic exposure and white balance calibration, automatically adjust hardware parameters according to different document materials, and perform multi-frame fusion noise reduction. By continuously capturing multiple frames of images and performing pixel-level variance analysis and weighted fusion, character contrast is improved and interference is reduced.

3. The rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology according to claim 1, characterized in that, The image recognition and information extraction module integrates a key field structured positioning enhancement mechanism, including a dynamic template matching unit for dynamically matching standard templates according to document type and performing deformation adaptive matching, and a multimodal field recognition fusion unit for performing multi-model recognition and cross-validation on mixed fields.

4. The rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology according to claim 2, characterized in that, The dynamic template matching unit performs tilt correction and coordinate calibration on key fields of various document types, including vehicle registration certificates and vehicle licenses, through dynamic template matching and deformation adaptive mechanisms. The multimodal field recognition and fusion unit calls character recognition models, symbol recognition models and texture recognition models to perform multi-model recognition and cross-validation for the mixed field of VIN code and engine number.

5. The rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology according to claim 1, characterized in that, The data matching and verification module further includes a VIN code verification and error correction enhancement unit, which is used to perform segmented semantic verification and fuzzy recognition error correction suggestions on the VIN code, and locate abnormal segments or provide similar character replacement suggestions when verification fails.

6. The rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology according to claim 1, characterized in that, The data matching and verification module uses a rule engine to perform data consistency verification, including multi-dimensional consistency verification of vehicle owner identity, linkage verification of document validity period, and multi-factor duplicate judgment of vehicle uniqueness and status, as well as multi-rule prediction of scrap status.

7. The rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology according to claim 1, characterized in that, The data matching and verification module also includes a business process and information verification collaboration mechanism, which is used to dynamically control process nodes based on the verification results, including automatic advancement, suspension of manual review, or triggering of high-risk work orders.

8. The rapid data entry system for scrapped vehicle information based on high-speed document scanner image recognition technology according to claim 1, characterized in that, The attachment management and traceability module further includes: The intelligent classification and archiving unit is used to automatically classify attachments based on text and image features, segment and archive mixed scans independently, and support the judgment of the association and reuse of historical attachments. Based on the evaluation results of document validity period and image clarity, it prompts whether historical attachments can be reused. The phased attachment inspection unit is used to dynamically adjust the attachment inspection rules according to the business stage, and to provide intelligent guidance and supplementary entry assistance when attachments are missing or of poor quality. The multi-dimensional index and version management unit supports combined searches by attachment type, creation time, and operator, and retains all attachment versions and their operation records.

Images