An automatic invoice review and intelligent settlement system for automobile roadside assistance

Abstract

The present invention discloses an automatic invoice review and intelligent settlement system for automobile roadside assistance, which relates to the technical field of intelligent settlement systems. The correction module integrates a machine learning model to analyze historical bills, automatically corrects unreasonable settlement amounts and generates a correction suggestion report. The prompt module establishes an intelligent priority matrix and dynamically adjusts the reminder frequency based on parameters such as the supplier's historical settlement scale and years of cooperation. The settlement module uses fingerprint recognition technology of the invoicing device to identify abnormal invoicing terminals, automatically triggers the payment process after verification, establishes a supplier credit profile system, and supports dynamic settlement plans. The settlement system establishes an intelligent priority matrix to achieve "high-frequency reminders for important suppliers and low-frequency intervention for low-risk projects", ensuring the timeliness of the overall process, effectively compensating for the shortcomings of traditional rule engines in dealing with complex scenarios, and improving the system's responsiveness to abnormal behavior and autonomous decision-making capabilities.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent settlement systems, and in particular to an automatic invoice review and intelligent settlement system for automobile roadside assistance. Background Art

[0002] With the continuous growth of vehicle ownership and the increasingly complex road traffic environment, the probability of vehicle breakdowns or accidents while driving has also increased accordingly. As a result, the demand for roadside assistance services has increased year by year. To improve customer satisfaction and service response efficiency, more and more insurance companies, automobile manufacturers, and third-party assistance service agencies are providing 24 / 7 roadside assistance services, including towing, battery jump-starting, tire changing, emergency fuel delivery, and other services.

[0003] As road rescue services become increasingly popular, the problems of invoice management and fee settlement are becoming increasingly prominent. The settlement system, combined with OCR recognition technology, business rule engine, big data comparison and smart contracts, realizes the automation and intelligent processing of the entire process of rescue invoices from uploading, review to settlement, improving operational efficiency, reducing management costs and enhancing system security.

[0004] The existing technology has the following defects:

[0005] 1. The current industry-wide use of rule-based engines for expense auditing makes it difficult to cope with volatile, unstructured anomalies. For example, if the frequency of rescues in a remote area increases abnormally during a specific period, traditional systems cannot proactively detect or issue warnings, which can easily lead to a large number of false settlements flowing into the financial system. The system lacks integrated analysis of multi-source data such as historical bills, geographic locations, and behavioral patterns, resulting in weak anomaly identification capabilities and a high risk of financial losses.

[0006] 2. Most existing systems communicate with suppliers about invoice verification matters only through simple to-do items and manual email reminders. Dynamic priority models have not been established, and there is a lack of intelligent reminders and emergency event handling mechanisms. The importance of suppliers is disconnected from communication strategies, which can easily lead to delayed reminders for high-priority businesses, affecting the overall process timeliness.

[0007] Based on this, the present invention proposes an automatic invoice review and intelligent settlement system for automobile roadside assistance, establishes an intelligent priority matrix to achieve "high-frequency reminders for important suppliers and low-frequency intervention for low-risk projects", ensures the timeliness of the overall process, effectively makes up for the shortcomings of traditional rule engines in dealing with complex scenarios, and improves the system's responsiveness to abnormal behaviors and autonomous decision-making capabilities. Summary of the Invention

[0008] The purpose of the present invention is to provide an automatic invoice review and intelligent settlement system for automobile roadside assistance to overcome the shortcomings of the background technology.

[0009] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: an automatic invoice review and intelligent settlement system for automobile roadside assistance, comprising a correction module, a prompt module, a storage module, an anti-counterfeiting identification module, and a settlement module;

[0010] Correction module: Integrates machine learning models to analyze historical data, alerts users of abnormally high settlement items, automatically corrects unreasonable settlement amounts, and generates correction suggestion reports;

[0011] Reminder module: establishes an intelligent priority matrix to dynamically adjust reminder frequency;

[0012] Storage module: uses alliance chain technology to store data in a distributed manner;

[0013] Anti-counterfeiting identification module: When receiving a complaint request, historical data is associated. After the supplier uploads the invoice, the module automatically associates the settlement statement and extracts the information into the database. OCR + 3D watermark recognition technology is used to prevent false printing of invoices.

[0014] Settlement module: Combines the fingerprint recognition technology of the invoicing device to identify abnormal invoicing terminals, and automatically triggers the payment process after verification.

[0015] Preferably, the settlement module collects the unique identifier of the invoicing device through the invoicing device fingerprint recognition technology each time the supplier issues an invoice, and the device fingerprint is used to identify each invoicing device;

[0016] The collected device fingerprint is compared with the records in the historical device fingerprint library to determine whether the device is a trusted device. If an abnormal device fingerprint in the historical record is matched, it is marked as an abnormal terminal and triggers manual review or further verification. If an abnormal invoicing terminal is detected, the invoicing operation of the terminal will be automatically stopped and the administrator will be notified for verification.

[0017] Preferably, after passing the device fingerprint verification, the settlement module will further verify the content of the invoice uploaded by the supplier, compare the supplier's past settlement records with the content of the current invoice, and determine whether the invoice complies with the regular settlement model. If the invoice amount deviates significantly from the historical data or exceeds the preset range, an early warning will be issued and manual review will be required;

[0018] If the device fingerprint verification is successful and the invoice information is legitimate, the payment process will be automatically triggered. The payment process will determine whether to settle in real time based on the supplier's credit rating, settlement amount, and invoice validity factors;

[0019] Automatically generate and update the supplier's credit profile based on the supplier's historical settlement behavior, invoice authenticity, and performance data.

[0020] Preferably, the supplier credit profile includes the following information:

[0021] Credit score: A comprehensive score based on the supplier's settlement history and invoice compliance indicators;

[0022] Settlement behavior: including timeliness of settlement and accuracy of invoices;

[0023] Contract Performance: The timeliness and quality of the supplier's contract performance;

[0024] After each settlement, the credit profile is automatically updated based on the new settlement information.

[0025] Preferably, the anti-counterfeiting identification module receives a settlement appeal request initiated by a supplier, obtains the corresponding historical rescue work order picture, GPS trajectory data, verified settlement order and bill correction report from the alliance chain storage module through the appeal order number or associated bill ID, automatically creates an identification task order, and pushes it to the anti-counterfeiting identification queue;

[0026] Suppliers upload invoices for verification, and the multimodal OCR engine extracts fields including invoice number, invoice code, invoice date, amount, tax rate, invoicing unit, purchaser information, item details, and total amount, while also recording the field location coordinates.

[0027] The identified field information is automatically matched with the associated billing database. If a field is missing, the amount is inconsistent, or the invoicing date does not fit the time range, the system marks it as a preliminary exception.

[0028] Perform light and shadow depth analysis on the security watermark area in the scanned image to check for legal angle reflection features and micro-print information, and use the watermark signature matching model to determine whether the ticket is an original machine-printed ticket;

[0029] Use image feature point matching algorithm to compare with the historical seal library of the issuing unit to detect counterfeit situations including seal offset, seal overlap, and font blur; combine with the fingerprint of the device to which the seal belongs to further verify the authenticity.

[0030] Preferably, the anti-counterfeiting identification module uses image segmentation technology to extract the seal area in the invoice, and uses SIFT or SURF algorithm to extract feature points of the seal area. The extracted feature points include: key point position, scale, direction, and descriptor information for subsequent feature matching;

[0031] Load the real seal image registered by the invoicing unit from the historical seal library, compare the feature points of the uploaded invoice seal with the feature points of the historical seal image, and determine the matching degree through the feature point matching algorithm;

[0032] If the match is qualified, the offset between the seal in the invoice image and the preset standard position is compared. If the offset is greater than the offset threshold, it is marked as abnormal;

[0033] An image overlap algorithm is used to determine whether the seal overlaps with other elements, and an image blur measurement algorithm is used to evaluate whether the text in the seal is blurred.

[0034] Preferably, the storage module deploys a multi-node alliance chain network including: platform operator nodes, third-party audit nodes and core supplier nodes;

[0035] Use the identity authentication module to manage the access rights of each participant and set up smart contract templates to standardize the storage rules for bill verification, confirmation, modification and settlement status;

[0036] Each bill verification process will automatically generate a verification transaction object. The verification transaction object generates a data fingerprint through the hash summary. The data fingerprint is packaged together with the original data and submitted to the alliance chain network for storage.

[0037] Each node participates in the consensus to confirm the validity of the verification transaction. If the consensus is passed, the data is packaged into a block to form an on-chain record.

[0038] Preferably, the prompt module obtains basic data and historical cooperation information of each supplier from the database, including the total amount of supplier's cumulative settlement, cooperation years, settlement abnormality rate and response time, to form a profile of each supplier;

[0039] All the acquired parameters are normalized and the comprehensive score is calculated. All suppliers are divided into multiple levels according to the comprehensive score, including A, B, C, and D, to form a priority matrix.

[0040] Preferably, the correction module automatically collects historical billing data and corresponding GPS rescue trajectory data, the historical bill includes rescue time, service type, settlement amount, supplier information, and service area, and the GPS rescue trajectory data includes the positioning points before and after the vehicle rescue, the driving path, and the rescue start and end time information;

[0041] The machine learning model embedded in the correction module trains on preprocessed data to extract the normal rescue frequency, rescue type, and amount distribution patterns in each region and time period, identify the characteristic boundaries of normal settlement behavior, and build a baseline behavior pattern library;

[0042] When new billing data enters the review process, it is compared with baseline behavior patterns to identify abnormally high-incidence settlement items.

[0043] Preferably, the correction module constructs a feature vector:

[0044] , where is the feature vector of the i-th rescue, is the duration of the i-th rescue, is the distance of the i-th rescue, is the amount of the rescue in Article i;

[0045] After performing K-Means clustering on all feature vectors, all clusters are obtained to minimize the total intra-class distance sum of squares: , where is the total intra-class distance sum of squares, is the number of cluster categories, is the sample set contained in the k-th cluster, is the centroid of the k-th cluster;

[0046] After obtaining a new bill, the new bill is divided into the corresponding cluster according to its feature vector. The deviation value is obtained by subtracting the amount corresponding to the cluster centroid from the new bill amount. If the deviation value is greater than the deviation threshold, the new bill is considered abnormal.

[0047] Re-adjust the new bill amount, the expression is: , where The adjusted amount for the new bill. is the amount corresponding to the centroid of the cluster, is the standard deviation of the cluster amount, is the regulating factor, and .

[0048] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0049] The present invention integrates a machine learning model through the correction module to analyze historical bills and GPS rescue trajectory data, warns of abnormal and high-incidence settlement items, automatically corrects unreasonable settlement amounts and generates a correction suggestion report. The prompt module establishes an intelligent priority matrix, dynamically adjusts the reminder frequency based on parameters such as the supplier's historical settlement scale and years of cooperation, and automatically triggers a telephone robot outbound call in an emergency. The settlement module combines the fingerprint recognition technology of the billing device to identify abnormal billing terminals, automatically triggers the payment process after verification, establishes a supplier credit profile system, and supports dynamic settlement plans. The settlement system establishes an intelligent priority matrix to achieve "high-frequency reminders for important suppliers and low-frequency intervention for low-risk projects", ensuring the timeliness of the overall process, effectively making up for the shortcomings of traditional rule engines in dealing with complex scenarios, and improving the system's responsiveness to abnormal behavior and autonomous decision-making capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments described in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0051] Figure 1 This is a system architecture diagram of the present invention.

[0052] Figure 2 This is a system timing diagram of the present invention.

[0053] Figure 3 This is a system framework diagram of the present invention. DETAILED DESCRIPTION

[0054] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0055] Example 1: Please refer to Figure 1-Figure 3 As shown, the invoice automatic review and intelligent settlement system for automobile roadside assistance described in this embodiment includes a correction module, a prompt module, a storage module, an anti-counterfeiting identification module, and a settlement module;

[0056] Correction module: Integrates machine learning models to analyze historical bills and GPS rescue trajectory data, alerting users of abnormally high-incidence settlement items (such as a surge in rescue frequency in remote areas), automatically correcting unreasonable settlement amounts, and generating a correction recommendation report. The corrected settlement amount and recommendation report are then sent to the storage module.

[0057] Reminder module: Establishes an intelligent priority matrix and dynamically adjusts reminder frequency based on parameters such as the supplier's historical settlement scale and years of cooperation. In emergencies, it automatically triggers a robot outbound call, and the priority matrix is ​​sent to the storage module.

[0058] Storage module: uses alliance chain technology to distribute the verification process and confirmation results to ensure that the operation process is auditable and traceable. The distributed storage results are sent to the anti-counterfeiting identification module and the settlement module;

[0059] Anti-counterfeiting identification module: When receiving a complaint request, it automatically associates historical rescue work order images, GPS tracks, and other original data. After the supplier uploads the invoice, it automatically associates the settlement statement and extracts key information into the database. It uses OCR + 3D watermark recognition technology to prevent false printing of invoices, and the invoice anti-counterfeiting identification results are sent to the settlement module;

[0060] Settlement module: Combines invoicing device fingerprint recognition technology to identify abnormal invoicing terminals, automatically triggers the payment process after verification, establishes a supplier credit profile system, and supports dynamic settlement plans - high-quality suppliers can settle basic payments in real time, and suppliers who do not meet the standards will have partial payment delayed.

[0061] This application integrates a machine learning model through the correction module to analyze historical bills and GPS rescue trajectory data, warns of abnormal and high-incidence settlement items, automatically corrects unreasonable settlement amounts and generates a correction suggestion report. The prompt module establishes an intelligent priority matrix, dynamically adjusts the reminder frequency based on parameters such as the supplier's historical settlement scale and years of cooperation, and automatically triggers a telephone robot outbound call in an emergency. The settlement module combines the fingerprint recognition technology of the billing device to identify abnormal billing terminals, automatically triggers the payment process after verification, establishes a supplier credit profile system, and supports dynamic settlement plans. The settlement system establishes an intelligent priority matrix to achieve "high-frequency reminders for important suppliers and low-frequency intervention for low-risk projects", ensures the timeliness of the overall process, effectively makes up for the shortcomings of the traditional rule engine in dealing with complex scenarios, and improves the system's responsiveness to abnormal behavior and autonomous decision-making capabilities.

[0062] The specific workflow of the settlement system is as follows:

[0063] The settlement system integrates machine learning models to analyze historical bills and GPS rescue trajectory data, issue warnings for abnormally high-incidence settlement items (such as a surge in the frequency of rescue operations in remote areas), automatically correct unreasonable settlement amounts and generate correction suggestion reports, establish an intelligent priority matrix, dynamically adjust the reminder frequency based on parameters such as the supplier's historical settlement scale and years of cooperation, automatically trigger outbound calls from telephone robots in emergencies, and use alliance chain technology to distribute and store the verification process and confirmation results to ensure that the operation process is auditable and traceable. When receiving a complaint request, it automatically associates historical rescue work order images, GPS trajectories and other original data. After the supplier uploads the invoice, it automatically associates the settlement statement and extracts key information into the database. It uses OCR+3D watermark recognition technology to prevent false printing of invoices, combines the fingerprint recognition technology of the invoicing device to identify abnormal invoicing terminals, and automatically triggers the payment process after verification. It establishes a supplier credit portrait system and supports dynamic settlement plans - high-quality suppliers can settle basic payments in real time, and suppliers that do not meet the standards will have partial amount settlement delayed.

[0064] Example 2: The correction module integrates a machine learning model to analyze historical bills and GPS rescue trajectory data, warns of abnormally high settlement items (such as a surge in the frequency of rescue in remote areas), automatically corrects unreasonable settlement amounts and generates a correction suggestion report. The corrected settlement amount and suggestion report are sent to the storage module.

[0065] The correction module automatically collects historical billing data and corresponding GPS rescue trajectory data from the system. Historical billing includes, but is not limited to, rescue time, service type, settlement amount, supplier information, and service area. GPS rescue trajectory data includes information such as the vehicle's location before and after the rescue, the driving route, and the rescue start and end times. The collected data undergoes preprocessing operations such as format standardization, missing value completion, and outlier removal to establish a unified data input format.

[0066] The machine learning model embedded in the module (such as the cluster analysis model or the time series anomaly detection model) is trained on the preprocessed data to extract the normal rescue frequency, rescue type and amount distribution pattern in each region and time period, identify the characteristic boundaries of normal settlement behavior, and thus build a benchmark behavior pattern library.

[0067] Correction module to construct feature vector: , where is the feature vector of the i-th rescue, is the duration of the i-th rescue, is the distance of the i-th rescue, is the amount of the rescue in Article i.

[0068] After performing K-Means clustering on all feature vectors, all clusters are obtained to minimize the total intra-class distance sum of squares: , where is the total intra-class distance sum of squares, is the number of cluster categories, is the sample set contained in the k-th cluster, is the centroid of the kth cluster (i.e., the mean of the eigenvectors of the samples in this cluster).

[0069] After obtaining a new bill, the new bill is divided into the corresponding cluster according to its feature vector. The deviation value is obtained by subtracting the amount corresponding to the cluster centroid from the new bill amount. If the deviation value is greater than the deviation threshold, the new bill is determined to be abnormal and the new bill amount is readjusted. The expression is: , where The adjusted amount for the new bill. is the amount corresponding to the centroid of the cluster, is the standard deviation of the cluster amount, is the regulating factor, and .

[0070] When new billing data enters the review process, the system uses the model to compare it to baseline behavior patterns and identify abnormally high-incidence billing items. If the rescue frequency in a geographic area during a specific time period is significantly higher than the historical average, or the rescue costs deviate significantly from the historical range, it will be marked as a "billing anomaly warning area" and the anomaly parameters (such as the frequency surge ratio and anomaly type) will be recorded.

[0071] The reminder module establishes an intelligent priority matrix and dynamically adjusts the reminder frequency based on parameters such as the supplier's historical settlement scale and years of cooperation. In emergencies, an outbound call is automatically triggered by a telephone robot, and the priority matrix is ​​sent to the storage module.

[0072] The prompt module obtains the basic data and historical cooperation information of each supplier from the database, including: total settlement amount: the total amount of the supplier's cumulative settlement, cooperation years: the number of years the supplier has cooperated with the platform, settlement anomaly rate: the proportion of settlement anomalies (such as corrections, delays, and rejections) that occurred in the past year, response time: the average bill processing response time. These parameters are input into the model as feature vectors to form a profile of each supplier.

[0073] Normalize all the acquired parameters and calculate the comprehensive score based on the set weights: , where is the comprehensive score of the i-th supplier, 、 、 、 is the weight (adjustable), The total amount of accumulated settlements with suppliers. For the cooperation period, is the settlement abnormality rate, For response time.

[0074] All suppliers are divided into multiple levels based on their comprehensive scores, including A, B, C, and D, to form a priority matrix. The priority matrix can be shown in Table 1:

[0075] Table 1

[0076] grade Score range Reminder Strategy A-level S≥80 Normal rhythm reminder (once a day) Class B 60≤S<80 Increase frequency (twice daily) C-level S<60 Strong reminder (once every 4 hours) D-Class Exception Marker Emergency notification + outbound call

[0077] The system writes each supplier's current level, reminder frequency, trigger conditions and other information into the "priority matrix structure table". The system runs regularly and executes corresponding reminder tasks according to the priority matrix: for levels B and C, the system sends reminder notifications through App messages, SMS or email; for level D, if it detects "bill rejection confirmation" or "no response for more than X hours", the telephone robot will be called immediately for voice outbound calls.

[0078] In this application, the weights (adjustable) are manually set by business experts or project teams based on their experience. Example weight settings (adjustable) are shown in Table 2:

[0079] Table 2

[0080] Parameter items illustrate Weight (Example) Total settlement amount Reflect service activity 0.30 Years of cooperation Indicates cooperation stability 0.20 Response time Reflect timely service 0.25 Settlement abnormality rate Indicates bill credibility 0.25

[0081] The storage module uses alliance chain technology to distribute the verification process and confirmation results to ensure that the operation process audit is traceable, and the distributed storage results are sent to the anti-counterfeiting identification module and the settlement module.

[0082] The storage module deploys a multi-node consortium chain network (such as based on Hyperledger Fabric or FISCOBCOS), including platform operator nodes, third-party audit nodes and core supplier nodes, and uses a unified identity authentication module (such as CA certificate) to manage the access rights of each participant; set up smart contract templates for standardized bill verification, confirmation, modification and settlement status storage rules.

[0083] For each bill verification process, the system will automatically generate a "verification transaction object", which includes: bill number, verification personnel, verification time, verification content, abnormal mark type, correction suggestion summary, operation IP address and device ID (for later auditing). The above information is generated through hash summary to generate a unique "data fingerprint", which is packaged together with the original data and submitted to the alliance chain network for storage.

[0084] Each node participates in a consensus (e.g., PBFT or RAFT) to confirm the validity of the verification transaction. Once consensus is reached, the system packages the data into a block, forming an on-chain record. Each verification transaction stores: original summary information, verification operation details, transaction status (e.g., "pending confirmation" or "correction proposal sent"), operation timestamp, and signature. Each block contains the hash value of the previous block, forming an immutable data chain.

[0085] When the anti-counterfeiting identification module receives a complaint request, it automatically associates historical rescue work order images, GPS tracks and other original data. After the supplier uploads the invoice, it automatically associates the settlement statement and extracts key information into the database. It uses OCR+3D watermark recognition technology to prevent false printing of tickets, and the invoice anti-counterfeiting identification results are sent to the settlement module.

[0086] The anti-counterfeiting identification module receives a settlement complaint request initiated by the supplier; through the complaint number or associated bill ID, it obtains the corresponding historical rescue work order images (such as vehicle rescue site photos, repair reports, etc.), GPS trajectory data (including the start and end locations and times of the rescue), verified settlement orders and bill correction reports from the alliance chain storage module, automatically creates an identification task order, and pushes it to the anti-counterfeiting identification queue.

[0087] Suppliers upload invoices for verification (supporting formats such as PDF and images). The anti-counterfeiting recognition module automatically performs the following processing:

[0088] 1) OCR text recognition processing: Use a multimodal OCR engine (supporting printed and handwritten text) to extract the following key fields: invoice number, invoice code, invoice date, amount, tax rate, invoicing unit, purchaser information, item details, and total amount. At the same time, the field position coordinates are recorded for subsequent format verification.

[0089] 2) Structured information storage: The identified field information is automatically matched with the associated billing database. If a field is missing, the amount is inconsistent, or the invoicing date does not fit the time range, the system will mark it as a "preliminary exception."

[0090] The anti-counterfeiting identification module performs anti-counterfeiting identification, including:

[0091] 1) 3D watermark feature verification: Perform light and shadow depth analysis on the security watermark area in the scanned image to check whether there are legal angle reflection features and micro-print information, and use the watermark signature matching model to determine whether the ticket is an original machine-printed ticket.

[0092] 2) Invoice seal image comparison: Use image feature point matching algorithms (such as SIFT / SURF) to compare with the historical seal library of the issuing unit to detect forgeries such as seal offset, seal overlap, and blurred fonts; combine with the fingerprint of the device to which the seal belongs to further verify authenticity.

[0093] The anti-counterfeiting identification module uses image segmentation technology (such as the Sobel operator) to extract the seal area in the invoice, and uses the SIFT (Scale-Invariant Feature Transform) or SURF (Speeded Robust Features) algorithm to extract the feature points of the seal area. The extracted feature points include: key point location, scale, direction, descriptor and other information for subsequent feature matching.

[0094] Load the real seal image registered by the invoicing unit from the historical seal library, compare the feature points of the uploaded invoice seal with the feature points of the historical seal image, and determine the matching degree through a feature point matching algorithm (such as brute force matching or FLANN matcher). The comparison result will return the number of matched feature points and the matching degree. A low matching degree indicates that the seal may be forged.

[0095] If the match is qualified, the offset of the seal in the invoice image and the preset standard position is compared. If the offset is greater than the offset threshold, it is marked as "abnormal". The image overlap algorithm is used to determine whether the seal overlaps with other elements (such as text, graphics, etc.). If it overlaps, it indicates that there may be forgery. The image blur measurement algorithm (such as Laplacian transform) is used to evaluate whether the text in the seal is blurred. If it is blurred, it may be a forged seal.

[0096] Let's assume a company uploads an invoice that includes a stamp, and that stamp might be forged. We'll explain in detail, step by step, how to use the above techniques to detect forgeries.

[0097] Input: Invoice image file.

[0098] The Sobel operator is used for edge detection. First, the invoice image is converted to grayscale. Then, the Sobel operator is applied to detect edges in the image, specifically the edges of the seal area. By calculating the gradient of pixels in the image, the Sobel operator is able to highlight the outline of the seal area.

[0099] Seal area extraction: The edge information detected by the Sobel operator can be used to extract the seal area through image segmentation algorithm.

[0100] Within the extracted seal region, key feature points are extracted using the SIFT (Scale-Invariant Feature Transform) or SURF (Speeded Up Robust Features) algorithms. These feature points include location, scale, orientation, and descriptors. These feature points are highly robust to seals of varying angles, scales, and lighting, ensuring that matching can be performed even when the seal changes.

[0101] Assume that there are real seal images of the company in the historical seal database. Load the historical images into the system and extract their feature points.

[0102] Use the Brute-Force Matcher or FLANN (Fast Library for Approximate Nearest Neighbors) matcher to compare the seal feature points in the invoice with the feature points of the historical seal image.

[0103] After detecting the seal area, the system compares the seal's position in the image with a preset standard position (e.g., the upper right corner of an invoice). The system calculates the deviation between the seal's position in the image and the standard position. If the deviation exceeds a set threshold (e.g., 5mm), it is marked as "abnormal."

[0104] Use image overlap algorithms (such as image stitching technology) to determine whether the seal overlaps with other elements (such as text or graphics). If so, it may be a forgery because a normal seal will not overlap with other elements.

[0105] Using an image blur measurement algorithm such as the Laplacian transform, check whether the text in the stamp area is blurry. If the text is blurry, it is usually because the stamp was forged using low-quality equipment.

[0106] The final verification result of the invoice seal is obtained by combining various detection results (feature matching, offset, overlap, and fuzziness). If there are any anomalies (such as large seal offset, overlap, or fuzziness), the invoice is marked as "forged." Otherwise, the invoice is considered valid and legal.

[0107] The device obtains the invoicing device number or signature (such as the anti-counterfeiting invoice code) shown on the invoice. This information is then combined with the "invoicing device fingerprint database" in the settlement module to verify that the device is registered. The device is also identified as frequently appearing in unusual invoices (e.g., a single device issuing unusually high invoice amounts within a single day). The identification and analysis results are then integrated to generate an anti-counterfeiting identification report, which includes an invoice key information extraction table, anti-counterfeiting verification status (passed / suspicious / forged), and details of suspicious items (such as amount discrepancies, watermark failures, or seal anomalies). The identification results are automatically sent to the settlement module via an interface for subsequent payment processing. If the identification result is "suspicious" or "forged," the system automatically suspends the corresponding settlement process and issues an alert. If the device presents a risk, it is immediately marked as a "suspicious terminal" and forwarded to the credit profiling module.

[0108] The settlement module uses fingerprint recognition technology of invoicing devices to identify abnormal invoicing terminals, automatically triggers the payment process after verification, establishes a supplier credit profile system, and supports dynamic settlement plans - high-quality suppliers can settle basic payments in real time, and suppliers who do not meet the standards will have partial payment delayed.

[0109] Each time a supplier issues an invoice, the settlement module uses invoicing device fingerprint recognition technology to collect the unique identifier of the invoicing device (for example, device hardware fingerprint, MAC address, operating system information, etc.). This device fingerprint is used to identify each invoicing device to ensure that invoices are issued using authentic and compliant devices. The collected device fingerprint is compared with the records in the historical device fingerprint library to determine whether the device is a trusted device. If an abnormal device fingerprint in the historical record is matched (for example, an unregistered device, an illegal device, etc.), it is marked as an "abnormal terminal" and triggers manual review or further verification. If an abnormal invoicing terminal is detected, the system automatically stops the invoicing operation of the terminal and notifies the administrator for verification to prevent the risk of counterfeit invoices.

[0110] After verifying the device's fingerprint, the system further verifies the invoice content uploaded by the supplier, ensuring that the invoice amount, tax ID, invoice number, and other information are consistent with pre-set standards to prevent settlement anomalies caused by human error or forgery. The system compares the supplier's past settlement records with the current invoice content to determine whether the invoice complies with standard settlement patterns. If the invoice amount significantly deviates from historical data or exceeds a pre-set range, the system will issue an alert and request manual review.

[0111] If the device fingerprint verification passes and the invoice information is valid, the system automatically triggers the payment process. The payment process determines whether to settle in real time based on factors such as the supplier's credit rating, the settlement amount, and the validity of the invoice. Multiple payment methods are supported (e.g., bank transfer, third-party payment platforms), and the most appropriate payment method is automatically selected based on the supplier's credit rating.

[0112] Automatically generate and update the supplier's credit profile based on the supplier's historical settlement behavior, invoice authenticity, contract performance and other data. The supplier credit profile includes the following information:

[0113] Credit score: A comprehensive score based on the supplier's settlement history, invoice compliance and other indicators.

[0114] Settlement behavior: including the timeliness of settlement, accuracy of invoices, etc.

[0115] Contract Performance: The timeliness and quality of the supplier's performance of the contract.

[0116] Credit Profile Update: After each settlement, the system automatically updates the credit profile based on the new settlement information. If a supplier exhibits unusual behavior (such as late invoice submission or frequent forgery of invoices), the system will lower their credit score, potentially impacting their subsequent settlement conditions.

[0117] High-quality suppliers (i.e., suppliers with high credit scores, compliant invoices, and good contract performance) enjoy real-time settlement, meaning that settlement payments are paid immediately after verification, improving their capital turnover efficiency. Substandard suppliers (i.e., suppliers with low credit scores, invoice issues, or poor contract performance) will have partial settlement deferred. Specifically, a portion of the settlement amount (such as the base payment) may need to be delayed until the supplier improves its credit behavior. Settlement conditions are dynamically adjusted based on the supplier's credit profile and historical performance. For example, high-quality suppliers are automatically marked as "priority settlement," while substandard suppliers will need to further improve their credit rating to enjoy real-time settlement.

[0118] This application ensures the security, transparency, and compliance of the settlement process through methods such as fingerprint recognition of billing devices, invoice data verification, and the construction of supplier credit profiles. The dynamic settlement solution automatically adjusts settlement conditions based on the supplier's credit status, providing real-time settlement services for high-quality suppliers while implementing delayed settlement for those that fail to meet standards, thereby mitigating financial risks and promoting standardized supplier behavior.

[0119] Throughout this specification, references to terms such as "one embodiment," "example," or "specific example" indicate that the specific features, structures, materials, or characteristics described in conjunction with that embodiment or example are included in at least one embodiment or example of the present invention. In this specification, schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0120] The preferred embodiments of the present invention disclosed above are intended only to help illustrate the present invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the present invention to specific embodiments. Obviously, many modifications and variations are possible based on the contents of this specification. These embodiments are selected and described in detail in this specification to better explain the principles and practical applications of the present invention, thereby enabling those skilled in the art to better understand and utilize the present invention. The present invention is limited only by the claims and their full scope and equivalents.

Claims

1. An automatic invoice verification and intelligent settlement system for automobile roadside assistance, characterized by: It includes correction module, prompt module, storage module, anti-counterfeiting identification module and settlement module; Correction module: Integrates machine learning models to analyze historical data, alerts users of abnormally high settlement items, automatically corrects unreasonable settlement amounts, and generates correction suggestion reports; Reminder module: establishes an intelligent priority matrix to dynamically adjust reminder frequency; Storage module: uses alliance chain technology to store data in a distributed manner; Anti-counterfeiting identification module: When receiving a complaint request, historical data is associated. After the supplier uploads the invoice, the module automatically associates the settlement statement and extracts the information into the database. OCR + 3D watermark recognition technology is used to prevent false printing of invoices. Settlement module: Combines invoicing device fingerprint recognition technology to identify abnormal invoicing terminals and automatically triggers the payment process after verification; The anti-counterfeiting identification module receives a settlement complaint request initiated by the supplier, and obtains the corresponding historical rescue work order image, GPS trajectory data, and verified settlement order and bill correction report from the alliance chain storage module through the complaint order number or associated bill ID, automatically creates an identification task order, and pushes it to the anti-counterfeiting identification queue; Suppliers upload invoices that need to be verified, and the multimodal OCR engine extracts fields and records the field location coordinates. The identified field information is automatically matched with the associated billing database. If a field is missing, the amount is inconsistent, or the invoicing date does not fit the time range, the system marks it as a preliminary exception. Perform light and shadow depth analysis on the security watermark area in the scanned image to check for legal angle reflection features and micro-print information, and use the watermark signature matching model to determine whether the ticket is an original machine-printed ticket; Use image feature point matching algorithm to compare with the historical seal library of the issuing unit to detect forgeries, and combine the fingerprint of the device to which the seal belongs to further verify the authenticity.

2. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 1 is characterized by: The settlement module collects the unique identifier of the invoicing device through the invoicing device fingerprint recognition technology each time the supplier issues an invoice. The device fingerprint is used to identify each invoicing device. The collected device fingerprint is compared with the records in the historical device fingerprint library to determine whether the device is a trusted device. If an abnormal device fingerprint in the historical record is matched, it is marked as an abnormal terminal and triggers manual review or further verification. If an abnormal invoicing terminal is detected, the invoicing operation of the terminal will be automatically stopped and the administrator will be notified for verification.

3. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 2, characterized in that: After passing the device fingerprint verification, the settlement module will further verify the content of the invoice uploaded by the supplier, compare the supplier's past settlement records with the content of the current invoice, and determine whether the invoice complies with the regular settlement model. If the invoice amount deviates significantly from historical data or exceeds the preset range, an alert will be issued and manual review will be required; If the device fingerprint verification is successful and the invoice information is legitimate, the payment process will be automatically triggered. The payment process will determine whether to settle in real time based on the supplier's credit rating, settlement amount, and invoice validity factors; Automatically generate and update the supplier's credit profile based on the supplier's historical settlement behavior, invoice authenticity, and performance data.

4. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 3 is characterized by: The supplier credit profile includes the following information: Credit score: A comprehensive score based on the supplier's settlement history and invoice compliance indicators; Settlement behavior: including timeliness of settlement and accuracy of invoices; Contract Performance: The timeliness and quality of the supplier's contract performance; After each settlement, the credit profile is automatically updated based on the new settlement information.

5. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 4 is characterized by: The anti-counterfeiting identification module uses a multimodal OCR engine to extract fields, including invoice number, invoice code, invoicing date, amount, tax rate, invoicing unit, purchaser information, item details, and total amount; Detect forgeries, including seal offset, seal overlap, and blurred fonts.

6. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 5, characterized in that: The anti-counterfeiting recognition module uses image segmentation technology to extract the seal area in the invoice, and uses SIFT or SURF algorithm to extract feature points of the seal area. The extracted feature points include: key point position, scale, direction, and descriptor information for subsequent feature matching; Load the real seal image registered by the invoicing unit from the historical seal library, compare the feature points of the uploaded invoice seal with the feature points of the historical seal image, and determine the matching degree through the feature point matching algorithm; If the match is qualified, the offset between the seal in the invoice image and the preset standard position is compared. If the offset is greater than the offset threshold, it is marked as abnormal; An image overlap algorithm is used to determine whether the seal overlaps with other elements, and an image blur measurement algorithm is used to evaluate whether the text in the seal is blurred.

7. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 6, characterized in that: The storage module deploys a multi-node alliance chain network including: platform operator nodes, third-party audit nodes and core supplier nodes; Use the identity authentication module to manage the access rights of each participant and set up smart contract templates to standardize the storage rules for bill verification, confirmation, modification and settlement status; Each bill verification process will automatically generate a verification transaction object. The verification transaction object generates a data fingerprint through the hash summary. The data fingerprint is packaged together with the original data and submitted to the alliance chain network for storage. Each node participates in the consensus to confirm the validity of the verification transaction. If the consensus is passed, the data is packaged into a block to form an on-chain record.

8. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 7, characterized in that: The prompt module obtains basic data and historical cooperation information of each supplier from the database, including the total amount of supplier's cumulative settlement, years of cooperation, settlement abnormality rate and response time, to form a profile of each supplier; All the acquired parameters are normalized and the comprehensive score is calculated. All suppliers are divided into multiple levels according to the comprehensive score, including A, B, C, and D, to form a priority matrix.

9. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 8, characterized in that: The correction module automatically collects historical billing data and corresponding GPS rescue trajectory data. The historical bill includes rescue time, service type, settlement amount, supplier information, and service area. The GPS rescue trajectory data includes the positioning points before and after the vehicle rescue, the driving path, and the rescue start and end time information; The machine learning model embedded in the correction module trains on preprocessed data to extract the normal rescue frequency, rescue type, and amount distribution patterns in each region and time period, identify the characteristic boundaries of normal settlement behavior, and build a baseline behavior pattern library; When new billing data enters the review process, it is compared with baseline behavior patterns to identify abnormally high-incidence settlement items.

10. The automatic invoice verification and intelligent settlement system for automobile roadside assistance according to claim 9, characterized in that: The correction module constructs the feature vector: , where is the feature vector of the i-th rescue, is the duration of the i-th rescue, is the distance of the i-th rescue, is the amount of the rescue in Article i; After performing K-Means clustering on all feature vectors, all clusters are obtained to minimize the total intra-class distance sum of squares: , where is the total intra-class distance sum of squares, is the number of cluster categories, is the sample set contained in the k-th cluster, is the centroid of the k-th cluster; After obtaining a new bill, the new bill is divided into the corresponding cluster according to its feature vector. The deviation value is obtained by subtracting the amount corresponding to the cluster centroid from the new bill amount. If the deviation value is greater than the deviation threshold, the new bill is considered abnormal. Re-adjust the new bill amount, the expression is: , where The adjusted amount for the new bill. is the amount corresponding to the centroid of the cluster, is the standard deviation of the cluster amount, is the regulating factor, and .

Images