Vehicle loss assessment method and device

Abstract

The embodiment of the invention provides a vehicle loss assessment method and device, the vehicle loss assessment method is applied to a scheduling node of a vehicle loss assessment system, the vehicle loss assessment system further comprises a loss assessment node and a loss checking node, and the method comprises the following steps: responding to a vehicle loss assessment task of a loss reporting vehicle, and calling the loss assessment node based on loss reporting information of the vehicle loss assessment task, obtaining damage identification information of the damaged vehicle; under the condition that the damage identification confidence coefficient of the damage identification information meets a first path scheduling constraint, a loss checking node is called based on the damage identification information, damage checking information is obtained, and the first path scheduling constraint is a confidence coefficient rule which needs to be met by scheduling from the loss assessment node to the loss checking node in the vehicle loss assessment system; and based on the damage verification information, generating a target damage assessment result of the damaged vehicle. According to the scheme, the task path is dynamically decided based on the information confidence, so that the high-confidence task can automatically enter a rapid verification process, and the loss assessment efficiency is remarkably improved on the premise of ensuring the processing quality.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of computer technology, and in particular to a method and apparatus for assessing vehicle damage. Background Technology

[0002] In vehicle insurance claims, the ability to quickly and accurately assess the damage to an accident vehicle and determine the repair costs is a core element that determines the efficiency of claims processing, customer satisfaction, and cost control for insurance companies.

[0003] Existing automated vehicle insurance claims assessment systems typically employ a pre-defined, linear, and fixed workflow. The typical process involves the system receiving a claim report and photos of vehicle damage, then using a claims assessment model for preliminary analysis. The system directly transmits the list of identified damaged parts and estimated repair costs to a rules-based calculation engine, which automatically calculates the compensation amount according to the claims rules. However, existing methods lack a mechanism for autonomous coordination and arbitration when the identification results of a single model are inherently uncertain, or when conflicts arise between the outputs of different modules (such as damage identification, repair pricing, and rule-based claims verification). This often necessitates interrupting the process and transferring it to manual processing. The low level of automation results in low claims assessment efficiency. Therefore, the industry urgently needs a claims assessment method that can intelligently schedule and coordinate multiple capabilities, and automatically handle uncertainties and conflicts. Summary of the Invention

[0004] In view of this, embodiments of this specification provide a vehicle damage assessment method. One or more embodiments of this specification also relate to a vehicle damage assessment device, a computing device, a computer-readable storage medium, and a computer program product, to address the technical deficiencies existing in the prior art.

[0005] According to a first aspect of the embodiments of this specification, a vehicle damage assessment method is provided, applied to a scheduling node of a vehicle damage assessment system, the vehicle damage assessment system further including a damage assessment node and a damage verification node, the method comprising: In response to the vehicle damage assessment task for a reported damaged vehicle, the damage assessment node is invoked based on the damage assessment information of the vehicle damage assessment task to obtain the damage assessment information of the reported damaged vehicle. If the damage assessment confidence level of the damage assessment information satisfies the first path scheduling constraint, the damage assessment node is called based on the damage assessment information to obtain the damage assessment information. The first path scheduling constraint is the confidence level rule that needs to be satisfied for the vehicle damage assessment system to schedule from the damage assessment node to the damage assessment node. Based on the damage assessment information, the target damage assessment result for the reported vehicle is generated.

[0006] According to a second aspect of the embodiments of this specification, a vehicle damage assessment device is provided, applied to a scheduling node of a vehicle damage assessment system, the vehicle damage assessment system further including a damage assessment node and a damage verification node, the device comprising: The first acquisition module is configured to respond to the vehicle damage assessment task of the reported vehicle, and call the damage assessment node based on the damage information of the vehicle damage assessment task to obtain the damage assessment information of the reported vehicle. The second acquisition module is configured to call the damage assessment node based on the damage assessment information to obtain damage assessment information when the damage assessment confidence of the damage assessment information meets the first path scheduling constraint. The first path scheduling constraint is the confidence rule that needs to be met for the vehicle damage assessment node to be scheduled to the damage assessment node in the vehicle damage assessment system. The generation module is configured to generate target damage assessment results for the reported vehicles based on damage assessment information.

[0007] According to a third aspect of the embodiments of this specification, a computing device is provided, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the above-described vehicle damage assessment method.

[0008] According to a fourth aspect of the embodiments of this specification, a computer-readable storage medium is provided that stores computer-executable instructions that, when executed by a processor, implement the steps of the vehicle damage assessment method described above.

[0009] According to a fifth aspect of the embodiments of this specification, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the steps of the vehicle damage assessment method described above.

[0010] This specification provides a vehicle damage assessment method in one embodiment, applied to a scheduling node of a vehicle damage assessment system. The vehicle damage assessment system further includes a damage assessment node and a damage verification node. The method includes: responding to a vehicle damage assessment task for a reported damaged vehicle, calling the damage assessment node based on the damage reporting information of the vehicle damage assessment task to obtain damage assessment information of the reported damaged vehicle; if the damage assessment confidence of the damage assessment information satisfies a first path scheduling constraint, calling the damage verification node based on the damage assessment information to obtain damage verification information, wherein the first path scheduling constraint is a confidence rule that must be satisfied for scheduling from the damage assessment node to the damage verification node in the vehicle damage assessment system; and generating a target damage assessment result for the reported damaged vehicle based on the damage verification information.

[0011] This embodiment of the specification embeds a confidence-based path decision-making capability into the scheduling node, enabling it to automatically determine and call the verification node based on the confidence level of the damage assessment task. For high-confidence tasks that meet the confidence level rules, the system can automatically and reliably transfer them to a professional verification node for subsequent processing, realizing a fully automated and efficient closed loop for such tasks, significantly improving the overall damage assessment efficiency and processing capacity. Attached Figure Description

[0012] Figure 1 This is a flowchart illustrating a vehicle damage assessment method provided in one embodiment of this specification; Figure 2 This is a flowchart illustrating the processing procedure of a vehicle damage assessment method provided in one embodiment of this specification; Figure 3 This is a schematic diagram of the structure of a vehicle damage assessment device provided in one embodiment of this specification; Figure 4 This is a structural block diagram of a computing device provided in one embodiment of this specification. Detailed Implementation

[0013] Many specific details are set forth in the following description to provide a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar extensions without departing from the spirit of this specification. Therefore, this specification is not limited to the specific implementations disclosed below.

[0014] The terminology used in one or more embodiments of this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of the one or more embodiments of this specification. The singular forms “a” and “the” as used in one or more embodiments of this specification and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0015] It should be understood that although the terms first, second, etc., may be used to describe various information in one or more embodiments of this specification, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, first may also be referred to as second without departing from the scope of one or more embodiments of this specification, and similarly, second may also be referred to as first. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to a determination."

[0016] Furthermore, it should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in one or more embodiments of this specification are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0017] This specification provides a method for assessing vehicle damage. It also relates to a vehicle damage assessment device, a computing device, a computer-readable storage medium, and a computer program product, which will be described in detail in the following embodiments.

[0018] See Figure 1 , Figure 1 A flowchart of a vehicle damage assessment method according to an embodiment of this specification is shown. The method is applied to a scheduling node of a vehicle damage assessment system, which also includes a damage assessment node and a damage verification node. The method specifically includes the following steps 102-106.

[0019] Step 102: In response to the vehicle damage assessment task for the reported vehicle, call the damage assessment node based on the damage information of the vehicle damage assessment task to obtain the damage assessment information of the reported vehicle.

[0020] A vehicle damage assessment task refers to the collection of core data entities and processing instructions that drive the vehicle damage assessment system to complete a single vehicle damage report. It is a standardized digital work object that encapsulates all the context information and initial data required to initiate and drive this automated damage assessment process.

[0021] In some embodiments, the vehicle damage assessment system further includes a customer node, which is used to receive and structure the damage report information submitted by the customer, generate a standardized vehicle damage assessment task based on the damage report information, and send the vehicle damage assessment task to the scheduling node; the scheduling node responds to the received vehicle damage assessment task and executes the vehicle damage assessment method.

[0022] Damage reporting information refers to the core dataset within the "Vehicle Damage Assessment Task" data package, which describes the basic facts of the accident, the condition of the vehicles involved, and supporting evidence. It serves as the direct input and factual basis for the scheduling node to initiate the damage assessment process and invoke the damage assessment node (Damage Agent) for damage identification and analysis. Damage reporting information typically includes the following: Basic accident information includes: time of occurrence, GPS coordinates / location description, accident type (e.g., rear-end collision, scrape, single-vehicle collision), and brief descriptions of the parties involved. This information is used to provide spatiotemporal and contextual information, and to correlate with historical data and assess the likelihood of an accident.

[0023] Vehicle identification information, including license plate number, vehicle identification number (VIN), brand, model, type (e.g., 2023 model), and vehicle color, is the basis for linking insurance policies, checking parts inventory, and determining vehicle repair plans.

[0024] Damage evidence materials: Image / video list: high-definition media files from multiple angles of each damaged area, along with brief descriptions (e.g., "scratched left front door," "dented right rear bumper"); Scanned copies / PDFs: may include photos of handwritten repair slips, old damage markings, etc. Damage evidence materials are the core content of visual analysis at the damage assessment stage, and their completeness and clarity directly affect the confidence level of the damage assessment.

[0025] Accordingly, the damage assessment node is a multimodal large-scale intelligent agent. The large model deployed at the damage assessment node simultaneously processes textual information (vehicle model, accident description) and visual information (damage photos), establishing semantic relationships between the two. For example, it can understand the textual description of "front collision" and focus on areas such as the front bumper, headlights, and hood in the image.

[0026] The damage assessment node receives the damage report information from the scheduling node and generates damage assessment information under the guidance of prompts.

[0027] Damage assessment information is a structured list of descriptions of the physical damage to a vehicle, generated by the damage assessment department after analyzing damage evidence (such as images) in the reported damage information. Its core function is to objectively record "what is broken" and "how badly it is broken." Damage assessment information mainly includes the following: List of damaged parts: names of objectively identified damaged parts (e.g., left front headlight, right rear door, front bumper cover).

[0028] Damage description: An objective description of the damage (e.g., cracks, dents, scratches, deformation, missing parts).

[0029] Damage extent / repair quantity estimation: Quantification of the damage scale (e.g., dent area 10cm x 5cm, scratch length 15cm). This provides a basis for subsequent calculation of labor hours and materials.

[0030] Damage assessment confidence level: The degree of confidence that the damage assessment node has in each of the above identification results.

[0031] Step 104: If the damage assessment confidence level of the damage assessment information satisfies the first path scheduling constraint, the damage assessment node is called based on the damage assessment information to obtain the damage assessment information. The first path scheduling constraint is the confidence level rule that needs to be satisfied for the vehicle damage assessment system to schedule from the damage assessment node to the damage assessment node.

[0032] The first path scheduling constraint is an automated decision-making rule based on confidence level. It defines the reliability threshold that must be met for a path to proceed from the damage assessment node to the damage verification node.

[0033] Specifically, the scheduling node receives damage assessment information output by the damage assessment node, which includes a detailed list of one or more damage items and their corresponding damage assessment confidence levels. The scheduling node matches the built-in "first path scheduling constraint" rules with this confidence level data and decides whether to call the damage assessment node based on the matching result.

[0034] It should be noted that the first path scheduling constraint can be flexibly configured according to actual needs. For example: Rule A (Strict Global Constraint): Call the damage assessment node = (confidence of all individual damages > 85%). This rule ensures that only cases with extremely high overall identification quality can enter automated review, minimizing damage assessment based on erroneous facts, but may be slightly conservative.

[0035] Rule B (Key Item Constraint): Call the damage assessment node = (Overall confidence level > 80%) AND (Confidence level of all key safety components > 75%). This rule is more intelligent; while pursuing overall efficiency, it imposes higher reliability requirements on the identification results of core components involving safety (such as headlights and braking systems) or high value. Even if the identification confidence level of scratches on the car door is average, as long as the identification of key components is reliable, the case can still be automatically processed.

[0036] In one optional implementation of this embodiment, after calling the damage assessment node based on the damage assessment task's reporting information to obtain the damage assessment information of the reported vehicle, the method further includes: If the damage assessment confidence level of the damage assessment information does not meet the first path scheduling constraint, an additional verification node is called to verify the damage assessment information. If the verification passes, a damage verification node is called based on the damage assessment information to obtain the damage verification information.

[0037] The supplementary verification node is an auxiliary intelligent agent or service module used to quickly review, verify, or supplement low-confidence damage assessment results. After completing its processing, the supplementary verification node determines that the problems in the damage assessment information have been resolved or are insufficient to hinder subsequent processes, making a positive judgment. The supplementary verification node can verify the damage assessment information from multiple perspectives.

[0038] In some embodiments, an additional verification node can verify the internal consistency of damage assessment information.

[0039] Specifically, examine whether there are logical inconsistencies within the damage assessment information. For example: Physical collision: The list includes both "severe dent of the left front door" and "intact glass of the left front door", which is extremely unlikely in a serious collision.

[0040] Geometric conflict: The claimed "scratching length of 50cm" exceeds the visible length of the identified part itself.

[0041] Type conflict: The same part is marked with both "crack" and "dent", and these two damage types are usually mutually exclusive.

[0042] If internal consistency checks reveal fundamental logical contradictions (such as the aforementioned "type conflict"), the result can be directly determined as "check failed," without the need for further checks. This is because self-contradictory results lack credibility as a basis for any decision.

[0043] In some embodiments, the additional verification node may also perform a consistency verification between the damage assessment information and prior knowledge after the content consistency verification, in order to check whether the damaged parts given in the damage assessment information actually exist on the vehicle model.

[0044] If the prior knowledge consistency check finds that the damaged part does not exist in this vehicle model, it can be determined as "verification failed".

[0045] In some embodiments, the additional verification node may also invoke a dedicated fine-tuning model for re-evaluation of specific damage items with low confidence levels after content consistency verification or consistency verification with prior knowledge.

[0046] If a specialized model supports the judgment of a previously low-confidence item with high confidence, it is considered strong evidence for the judgment of that item, significantly improving the credibility of the overall information.

[0047] If a specialized model refutes the original judgment with high confidence, it is considered a crucial counter-evidence, which usually results in "verification failure".

[0048] Once the additional verification node completes its verification process (which may include one or more of the above verifications), if no fundamental contradictions sufficient to negate the damage identification information are found, it is determined that the verification has passed.

[0049] In the embodiments described in this specification, a two-layer filtering mechanism of "damage assessment confidence screening" plus "additional rule verification" is used to achieve refined classification of cases. Only those cases that have both technical uncertainty (low confidence) and fail verification are ultimately submitted to human review. This avoids pushing all low-confidence cases to human review indiscriminately, enabling the precise allocation of human resources to truly complex and high-value disputed cases, thereby reducing overall operating costs.

[0050] This embodiment provides a human-machine collaborative damage assessment optimization mechanism. When the damage assessment confidence level of the damage assessment information does not meet the first path scheduling constraint, the scheduling node will automatically trigger the manual intervention process, whereby professional damage assessors will review and process the case.

[0051] Step 106: Based on the damage assessment information, generate the target damage assessment result for the reported vehicle.

[0052] The target damage assessment result is a structured damage assessment conclusion document output by the vehicle damage assessment system, possessing final operational validity. It typically contains the following structured information: Case identifiers include: task number, license plate number, and policy number.

[0053] Vehicle and accident information: basic facts such as vehicle model, time and location of the accident.

[0054] Damage confirmation: The final list of damaged parts and repair / replacement plan after identification by the damage assessment node and approval by the damage verification node.

[0055] Final compensation conclusion: The amount of compensation (or the conclusion of rejection of compensation) determined after loss verification and scheduling node adjudication.

[0056] In one possible implementation of the embodiments of this specification, the vehicle damage assessment system further includes a repair verification node; After calling the damage assessment node based on the vehicle damage assessment task's reporting information and obtaining the damage assessment information of the reported vehicle, it also includes: If the damage assessment confidence level of the damage assessment information satisfies the second path scheduling constraint, the repair verification node is called based on the damage assessment information to obtain the repair verification information of the reported vehicle. The second path scheduling constraint is the confidence level rule that needs to be satisfied for the vehicle damage assessment node to be scheduled to the repair verification node in the vehicle damage assessment system. Accordingly, based on the damage assessment information, the target damage assessment result for the reported vehicle is generated, including: By combining the damage assessment confidence level of the damage assessment information and the repair assessment confidence level of the repair assessment information, the target damage assessment result for the reported vehicle is generated.

[0057] The repair verification node is an intelligent agent or service module that connects to external repair market data or an internal labor and parts database. Its core responsibility is to query and return the market-based repair price (including parts costs, labor costs, and auxiliary material costs) and corresponding repair suggestions for the current region and specific vehicle model, based on the damaged parts and repair plans listed in the damage assessment information. It represents an independent quote from either the "market side" or the "repair service side."

[0058] Maintenance verification information is the output after processing at the maintenance verification node. It typically includes: Reference repair price: Estimated market repair cost for each damaged component.

[0059] Suggested repair recommendations: Repair methods (such as "repair" or "replace") based on repair techniques and market practices.

[0060] Repair certification confidence level: The reliability assessment of the prices and recommendations provided by the repair certification node. This confidence level may be derived from factors such as the authority of the data source, the timeliness of the price, and the vehicle model compatibility.

[0061] The second path scheduling constraint is similar to the first path scheduling constraint; it is a preset confidence threshold rule. It is specifically used to determine whether damage assessment information can be scheduled from the damage assessment node to the maintenance verification node. Its threshold may be the same as or different from the first constraint, reflecting the system's differentiated quality requirements for different subsequent paths.

[0062] After obtaining damage assessment information, the scheduling node will simultaneously (or according to the strategy sequence) evaluate the scheduling constraints of the two paths.

[0063] If the first path scheduling constraint is met, the damage assessment node is invoked to obtain damage assessment information.

[0064] If the second path scheduling constraint is met, the maintenance verification node is invoked to obtain maintenance verification information.

[0065] These two paths are executed in parallel in most cases to improve overall processing efficiency.

[0066] After receiving damage assessment information from the damage assessment node and repair assessment information from the repair assessment node, the scheduling node comprehensively considers the damage assessment confidence level of the damage assessment information and the repair assessment confidence level of the repair assessment information to generate the target damage assessment result for the reported vehicle.

[0067] In the embodiments described in this specification, by introducing market price information independent of the assessment node, the system can perform market conformity verification on the theoretical compensation amount calculated by the assessment node. This can effectively identify and correct pricing anomalies (too high or too low) caused by model bias, making the final compensation amount closer to the actual repair cost, and improving the objectivity of the damage assessment results and customer acceptance.

[0068] In one optional implementation of this embodiment, the damage assessment information includes the set repair price and / or set repair recommendations for the damaged parts of the reported vehicle, and the repair assessment information for the reported vehicle includes the reference repair price and / or reference repair recommendations for the damaged parts of the reported vehicle. By combining the damage assessment confidence level of the damage assessment information and the repair assessment confidence level of the repair assessment information, a target damage assessment result for the reported vehicle is generated, including: Based on the damage assessment confidence level of the damage assessment information, the repair assessment confidence level of the repair assessment information, and the set repair price and reference repair price of the damaged component, a repair price assessment result is generated. Based on whether the initial repair recommendations and the reference repair recommendations for the damaged parts are consistent, a repair recommendation damage assessment result is generated; Based on the repair price assessment results and / or repair recommendation assessment results, a target damage assessment result for the damaged vehicle is generated.

[0069] The repair price is determined by the assessment node through analysis of historical similar case compensation data and the specific details of the current damage (parts, extent), and inferred by its embedded intelligent model (such as a trained machine learning model) as the estimated reasonable compensation amount for this reported damage.

[0070] Setting repair recommendations is a process by which the damage assessment node analyzes repair records of similar historical cases, mainstream repair scheme data under different damage conditions, and the characteristics of the components themselves (such as safety and repairability), and infers recommended repair schemes (such as "replacement" or "repair") for each damaged component in this report through an intelligent model.

[0071] The reference repair price is an estimate of repair costs obtained by the repair verification point through querying real-time market databases, partner repair shop quotation systems, or industry price standards, reflecting the current market conditions. It represents the "actual cost" from a market perspective.

[0072] Reference repair recommendations are common or recommended repair solutions proposed by the repair verification nodes based on market-standard repair processes, parts availability, and repair shop practices. They represent "practical solutions" from a market perspective.

[0073] The repair price assessment result is the final compensation amount determined by the dispatch node after comprehensively considering the set repair price, the reference repair price, and the confidence level of both, through certain integration and adjudication rules.

[0074] The repair recommendation damage assessment result is the final repair plan for each damaged component determined by the scheduling node after comparing the set repair recommendations and the reference repair recommendations, through consistency judgment or conflict resolution (e.g., "left front headlight: replacement").

[0075] In the embodiments described in this specification, the scheduling node obtains opinions from the damage assessment node and the maintenance verification node in parallel or sequentially, and each opinion is accompanied by a self-assessed confidence level.

[0076] In terms of price: The scheduling node takes two prices and their confidence levels as inputs, and uses built-in decision logic (such as weighted average, prioritizing the one with higher confidence, and strategy selection based on difference thresholds) to generate a repair price assessment result that is both compliant and fair to the market.

[0077] At the solution level: The scheduling node first compares the two repair suggestions. If they match, they are adopted directly; if they do not match, a pre-set conflict resolution mechanism is triggered (e.g., adopting a safer solution, a more cost-effective solution, or transferring it to a higher-level decision), thus forming a repair suggestion damage assessment result.

[0078] Results Integration and Output: Finally, the scheduling node integrates the "price result" and "solution result", and may combine other case information to package and generate a complete, structured, and actionable target loss assessment result.

[0079] By applying the embodiments in this specification, and through cross-validation and fusion of the "reference price / suggestion" of the repair assessment node and the "set price / suggestion" of the damage assessment node, the system can overcome the limitations of a single information source. This ensures that the final damage assessment result not only inherits solutions that have been validated in historical practice but also aligns with current market realities, thereby significantly improving the fairness and acceptability of the compensation plan and making it easier for customers and repair companies to accept.

[0080] In one optional implementation of this embodiment, a repair price assessment result is generated based on the damage assessment confidence level of the damage assessment information, the repair assessment confidence level of the repair assessment information, and the set repair price and reference repair price of the damaged component, including: If the damage assessment confidence of the damage assessment information satisfies the third path scheduling constraint, and the repair assessment confidence of the repair assessment information of the reported vehicle satisfies the fourth path scheduling constraint, determine the degree of repair difference between the set repair price and the reference repair price of the damaged parts of the reported vehicle. Repair price assessment results are generated based on the degree of repair differences.

[0081] The third-path scheduling constraint, set by the Orchestrator node in the decision-making process, is a confidence threshold rule specifically used to determine whether damage assessment information (from the damage assessment node) is qualified to enter the "price difference comparison and fusion" stage. Only when the damage assessment confidence level meets this constraint (e.g., confidence level ≥ X%), will its included "set repair price" be adopted as one of the reliable inputs.

[0082] The fourth path scheduling constraint is a confidence threshold rule parallel to the third constraint, specifically used to determine whether maintenance verification information (from the maintenance verification node) is qualified to enter the "price difference comparison and fusion" stage. Only when the maintenance verification confidence level meets this constraint (e.g., confidence level ≥ Y%) will the "reference maintenance price" it contains be adopted as another reliable input.

[0083] These two constraints together constitute a "double high-confidence gating" mechanism. It ensures that subsequent difference comparisons and price integration are based on the high credibility of both data sources, thereby avoiding decision-making biases or risks caused by using low-quality data from one source.

[0084] The degree of repair difference is quantified by setting an objective indicator of the deviation between the repair price and the reference repair price. It can be an absolute difference (reference price - set price, such as 300 yuan) or a relative difference (absolute difference / set price × 100%, such as 20%). The specific calculation method can be configured according to the business scenario. The core is to objectively reflect the deviation between the two types of prices.

[0085] In this scheme, the scheduling node first extracts the damage assessment confidence level of the damage assessment agent and the repair assessment confidence level of the repair shop agent, and compares them with the scheduling constraints of the third and fourth paths, respectively. Only when both types of confidence levels meet the corresponding constraints (e.g., damage assessment confidence level ≥ 75% and repair assessment confidence level ≥ 70%) will the subsequent price fusion process begin; if either confidence level does not meet the constraint, it indicates that the reliability of the output data of the corresponding node is insufficient, and it will be directly transferred to manual review or further verified by an additional verification node to avoid fusion errors caused by low-reliability data.

[0086] By applying the embodiments in this manual and setting two independent confidence thresholds, the system can effectively filter out high-quality "experience prices" and "market prices" for comparison, fundamentally eliminating decision-making errors that may arise from using low-reliability data and ensuring the stability of the damage assessment price. Furthermore, the degree of difference in repair costs can reflect the risk of the conclusion to some extent. Significant price differences often indicate potential problems (such as model bias, abnormal market prices, or fraud risk). This mechanism calculates and makes decisions based on this difference, thereby automatically identifying anomalies and triggering corresponding processing paths (such as conservative pricing or submission to manual intervention), achieving proactive risk control.

[0087] In one possible implementation of this embodiment, generating a repair price assessment result based on the degree of repair difference includes: If the degree of difference in repair is less than or equal to the first threshold, the repair price assessment result is determined based on the reference repair price. If the degree of repair difference is greater than the first threshold and less than or equal to the second threshold, then the repair price assessment result is determined based on the set repair price. If the degree of difference in repair exceeds the second threshold, an abnormal alarm will be issued.

[0088] In this plan, the degree of repair discrepancy is compared with the first and second thresholds to determine the corresponding discrepancy level. The tiered processing rules are then implemented as follows: Scenario 1 (Repair Difference ≤ First Threshold): The difference is within a reasonable market fluctuation range, indicating that the reference repair price (actual price at the repair shop) conforms to market rules and has not deviated from historical experience. In this case, the repair price assessment result is determined based on the reference repair price (e.g., reference price 1580 yuan, set price 1500 yuan, difference 5% ≤ 10%, then the assessment result is 1580 yuan), which fully adapts to the actual repair cost and avoids repair shops refusing to repair or customers paying extra fees due to strictly adhering to the theoretical price.

[0089] Scenario 2 (First threshold < Repair difference ≤ Second threshold): The difference exceeds normal fluctuations but is still manageable, possibly due to regional price differences or differences in repair shop quality (e.g., 4S shops vs. ordinary repair shops). To avoid the insurance company bearing excessive premiums, the repair price assessment result is determined based on a set repair price (e.g., reference price 1875 yuan, set price 1500 yuan, difference 25% ≤ 30%, then the assessment result is 1500 yuan). Simultaneously, the dispatcher records this difference data and feeds it back to the rule optimization module, providing a basis for subsequent updates to the damage assessment model and adjustments to the vehicle model's standard price (e.g., verifying whether the price of parts for this vehicle model has increased).

[0090] Scenario 3 (Repair Difference > Second Threshold): The difference exceeds the risk control boundary, posing a high risk of anomalies (e.g., repair shops deliberately inflating prices, incorrect part name matching leading to distorted pricing, poor performance of the damage assessment model). In this case, the system triggers an anomaly alarm, and the scheduling node transfers the task to manual review, pushing complete task information (difference level, two price categories, repair shop qualifications, accident photos, etc.) to help reviewers quickly pinpoint the root cause of the problem (e.g., verifying whether the repair shop's price is inflated, reconfirming the accuracy of part name matching). After verification, the final repair price assessment result is determined.

[0091] By applying the embodiments in this specification, a structured conflict resolution mechanism is provided for output conflicts between different nodes through predefined and explicit arbitration rules. This mechanism adjudicates based on objectively quantified difference data, avoiding a state of decision-making stalemate or inability to reach consensus.

[0092] In one possible implementation of this embodiment, a repair price assessment result is generated based on the damage assessment confidence level of the damage assessment information, the repair assessment confidence level of the repair assessment information, and the set repair price and reference repair price of the damaged component. This includes: If the damage assessment confidence level of the damage assessment information satisfies the third path scheduling constraint, and the repair assessment confidence level of the repair assessment information of the reported vehicle satisfies the fourth path scheduling constraint, the target repair price is determined from the set repair price and the reference repair price based on the confidence level of the set repair price and the confidence level of the reference repair price, and the repair price assessment result is determined based on the target repair price.

[0093] In this scheme, assuming both sources meet quality standards, the confidence levels of the set repair price and the reference repair price are directly compared. The system will select the price with the higher confidence level and determine it as the target repair price for this damage assessment. The final target repair price will then be adopted as the repair price assessment result for this task.

[0094] This specification provides a direct and efficient conflict resolution strategy, simplifying the decision-making process into a confidence-based source selection problem. The underlying logic is: given that both parties possess a certain level of credibility, the answer provided by the agent with greater certainty in its judgment should be trusted. This method features a short decision path, low computational overhead, and is suitable for scenarios requiring high processing efficiency and where the confidence index can effectively distinguish the reliability of information sources.

[0095] In one possible implementation of this embodiment, a repair suggestion damage assessment result is generated based on whether the set repair suggestion and the reference repair suggestion of the damaged component are consistent, including: Determine whether the recommended repair for the damaged component is consistent with the recommended repair. If the set repair recommendations for the damaged component are inconsistent with the reference repair recommendations, a repair recommendation assessment result for the damaged component will be generated based on the set conflict resolution rules.

[0096] Conflict resolution rules are set to automatically resolve conflicts between the set maintenance recommendations and the reference maintenance recommendations.

[0097] For example, confidence level is used as the basis for conflict resolution. Repair recommendations for nodes with higher confidence levels are given priority.

[0098] Alternatively, data from similar historical cases (same vehicle model, same damaged parts, similar scenarios) can be retrieved, mainstream repair recommendations can be compiled (e.g., 80% of similar cases use "repair"), and the final recommendation can be determined based on the majority decision principle.

[0099] The embodiments in this specification coordinate the outputs of the damage assessment node and the maintenance assessment node through the scheduling node, and realize intelligent collaboration between "experience-oriented" and "practice-oriented" approaches through conflict resolution rules, thereby balancing the contradictions between the two.

[0100] In one optional embodiment of this invention, the vehicle damage assessment system further includes a compliance node; before generating the target damage assessment result for the reported vehicle by combining the damage assessment confidence level of the damage assessment information and the repair assessment confidence level of the repair assessment information, the system further includes: Call the compliance node to perform risk detection based on at least one of the following information: damage report information, damage assessment information, damage verification information, and repair verification information, and obtain risk warning information; Combining the damage assessment confidence level of the damage assessment information and the repair assessment confidence level of the repair assessment information, the target damage assessment result for the reported vehicle is generated, which also includes: The target damage assessment result is determined by combining the damage assessment confidence level of the damage assessment information, the maintenance assessment confidence level of the maintenance assessment information, and the risk warning information.

[0101] Compliance Node: This is a risk control and compliance management module independent of the loss assessment, loss verification, and repair shop nodes. It is equipped with a compliance rule library or risk detection model. Its core responsibilities are to receive data from the entire process, perform risk detection, and generate risk warning information.

[0102] Risk warning information is a structured risk assessment conclusion output by the compliance node. It includes the risk level (such as high, medium, low), and may also include the risk type, risk score, specific risk point description, and supporting evidence.

[0103] In some embodiments, compliance nodes embed compliance rules, such as: the rule base presets "the number of reports of the same damaged part of the same vehicle within 12 months ≤ 2 times", and if the limit is exceeded, "the risk of duplicate claims" is triggered.

[0104] In other embodiments, compliance nodes are equipped with risk monitoring models. These risk detection models are typically large language models, trained on historical data of unusual cases, and possess semantic understanding and pattern recognition capabilities. For example, by comparing the lighting, angles, and damage features of reported damage photos with those of historical actual accidents, they can identify "false reports" (such as using photos of other people's accidents); by analyzing the discrepancies between reference repair prices and quotes from similar historical cases (same vehicle model, damaged parts, region), they can identify "over-quoting."

[0105] After the compliance node completes the inspection, it generates standardized risk warning information. The scheduling node adds risk warning information as a decision-making dimension based on the integration of damage assessment confidence and maintenance assessment confidence, forming a three-dimensional fusion logic of "dual confidence + risk".

[0106] For example, one decision rule is as follows: The risk level is "low" (e.g., slight rule adaptation deviation, which does not affect the reasonableness of compensation): The target loss assessment result is generated based on the original dual confidence level fusion result, and a risk warning is marked in the result (e.g., "There is a slight rule adaptation deviation, which does not affect compensation") for subsequent traceability.

[0107] For risks classified as "medium" (e.g., abnormal pricing, slight suspicion of duplicate claims): the compensation ratio will be appropriately reduced based on the dual-confidence fusion price (e.g., compensation at 90% of the fusion price) or manual review will be triggered to verify the risk before determining the final result; if the manual review confirms that the risk is not established (e.g., the abnormal pricing is due to parts upgrades), compensation will be paid at the original fusion price.

[0108] For cases with a risk level of "high" (such as false reports, forged liability determination letters, or severely excessive pricing): the automated loss assessment process will be suspended, an emergency manual review will be triggered, and complete case information (risk warnings, full-chain data, and testing evidence) will be pushed simultaneously. Final compensation information will not be generated until the reviewers have verified that the information is correct, in order to avoid significant losses for the insurance company.

[0109] In addition, compliance nodes synchronously record data throughout the entire process, including input and output parameters of each node, risk detection process, risk warning information, decision results of scheduling nodes, timestamps, etc., and adopt a chain-like log structure (which can be connected to blockchain technology to ensure immutability), providing complete evidence for regulatory audits and dispute tracing.

[0110] In the embodiments described in this specification, the compliance node acts as an independent intelligent agent, performing risk detection from a global perspective, unaffected by business processing flows. This allows for a more objective and comprehensive identification of complex risk patterns across nodes and processes, forming effective internal checks and balances. Simultaneously, a dual-dimensional intelligent decision-making system of "business + risk control" is constructed: system decisions no longer rely solely on the output confidence levels of business nodes, but instead use independent risk assessments as key decision factors. This ensures that the final loss assessment considers both business rationality and risk control compliance requirements, achieving a balance between efficiency and security.

[0111] In some embodiments, the scheduling node also maintains a task queue and a status table to enable dynamic task scheduling.

[0112] The task queue is used to store pending vehicle damage assessment tasks. The task queue is usually sorted according to a certain priority strategy (such as first-in-first-out, or according to the urgency of the case) to ensure that high-priority tasks are processed in a timely manner.

[0113] The status table is used to record the current processing status of each task, the steps completed, the results of each step (such as damage identification information, damage verification information, etc.), and the information of the agent (i.e., node) currently responsible for processing the task.

[0114] The status table can be a database table or a key-value pair in distributed storage. Each task has a unique identifier, which can be used to query and update the task status.

[0115] The scheduling node dynamically assigns tasks to appropriate agents based on the status and priority of tasks in the task queue, as well as the current load and availability of each agent.

[0116] The scheduling process is dynamic because the status of tasks changes (e.g., from being completed at the damage assessment node to waiting for processing at the damage verification node), and the system load may also change.

[0117] Specifically, step one: task reception and queuing. When a new vehicle damage assessment task is generated, the scheduling node puts it into the task queue and creates a new record for the task in the status table, with the initial status being "pending assignment".

[0118] Step 2: Task Allocation. The scheduling node monitors the task queue, retrieves tasks according to priority, and assigns them to the corresponding agents based on the steps the task needs to perform (such as loss assessment, loss verification, etc.).

[0119] For example, if a task currently requires damage assessment, the scheduling node will find an available damage assessment node and assign the task to it. After assignment, the status of the task in the status table is updated to "Damage Assessment in Progress," and the identifier of the assigned agent is recorded.

[0120] Step 3: State Monitoring and Updates. The scheduling node monitors the processing results returned by each agent. When an agent completes a task step, it returns the result to the scheduling node and updates the state table.

[0121] For example, after the damage assessment node completes the damage assessment, it returns the damage assessment information. The scheduling node updates the status of the task in the status table to "damage assessment completed" and stores the damage assessment information.

[0122] Step Four: Subsequent Step Scheduling. The scheduling node determines the next operation based on the current task status and preset procedures. For example, after damage assessment is completed, the scheduling constraints are determined based on the confidence level of the damage assessment information. If satisfied, a verification node is assigned. If the scheduling constraints are not satisfied, an additional verification node may be triggered, or the task may be transferred to manual processing. These decision-making logics are executed by the scheduling node according to preset rules.

[0123] Step 5: Task Completion. After the task has gone through all the necessary steps (which may include damage assessment, damage verification, repair verification, compliance checks, etc.), the scheduling node generates the final target damage assessment result, updates the task status to "completed", and removes it from the task queue.

[0124] With this design, the scheduling node can efficiently and reliably coordinate multiple agents to complete complex vehicle damage assessment processes, while ensuring the system's scalability and fault tolerance.

[0125] The following is in conjunction with the appendix Figure 2 Taking the application of the vehicle damage assessment method provided in this manual in automatic insurance claims as an example, the vehicle damage assessment method will be further explained. Among other things, Figure 2 The present specification shows a process flowchart of a vehicle damage assessment method according to an embodiment, which specifically includes the following steps 202-212.

[0126] Step 202: In response to the vehicle damage assessment task for the reported vehicle, call the damage assessment node based on the damage information of the vehicle damage assessment task to obtain the damage assessment information of the reported vehicle.

[0127] Step 204: Determine whether the damage assessment confidence level of the damage assessment information satisfies the first path scheduling constraint.

[0128] If the conditions are met, the process proceeds to step 206 for continued automated processing.

[0129] If the conditions are not met, the process proceeds to step 204B, where an additional verification mechanism is initiated.

[0130] Step 204B: Call the additional verification node to verify the damage assessment information.

[0131] If the verification passes, it means the problem is resolved or has little impact, and the process can continue to step 206.

[0132] If the verification fails, the system determines that it cannot be reliably automated at present, and directly transfers the process to manual review, thus terminating the automated process.

[0133] Step 206: Based on the damage assessment information, call the damage verification node and the repair verification node in parallel or sequentially to obtain damage verification information and repair verification information. The damage verification information includes the set repair price and set repair suggestions for the damaged parts of the reported vehicle, and the repair verification information includes the reference repair price and reference repair suggestions for the damaged parts of the reported vehicle.

[0134] Step 208: Based on the damage assessment confidence level of the damage assessment information, the repair assessment confidence level of the repair assessment information, and the set repair price and reference repair price of the damaged component, generate a repair price assessment result; based on whether the set repair suggestion and reference repair suggestion of the damaged component are consistent, generate a repair suggestion assessment result.

[0135] Step 210: Call the compliance node to perform risk detection based on at least one of the following information: damage report information, damage assessment information, damage verification information, and repair verification information, and obtain risk warning information.

[0136] Step 212: Combine the damage assessment confidence level of the damage assessment information, the maintenance assessment confidence level of the maintenance assessment information, and the risk warning information to determine the target damage assessment result.

[0137] Corresponding to the above method embodiments, this specification also provides embodiments of a vehicle damage assessment device. Figure 3 This specification shows a schematic diagram of a vehicle damage assessment device according to one embodiment. This device is applied to a scheduling node of a vehicle damage assessment system. The vehicle damage assessment system also includes an assessment node and a verification node, such as... Figure 3As shown, the device includes: The first acquisition module 302 is configured to respond to the vehicle damage assessment task of the reported vehicle, call the damage assessment node based on the damage information of the vehicle damage assessment task, and obtain the damage assessment information of the reported vehicle. The second acquisition module 304 is configured to, when the damage assessment confidence of the damage assessment information satisfies the first path scheduling constraint, call the verification node based on the damage assessment information to obtain damage verification information. The first path scheduling constraint is the confidence rule that needs to be satisfied for the vehicle damage assessment system to schedule from the damage assessment node to the verification node. The generation module 306 is configured to generate the target damage assessment result for the reported vehicle based on the damage assessment information.

[0138] Optionally, the vehicle damage assessment system also includes a repair verification node; the aforementioned device also includes a third acquisition module configured to: If the damage assessment confidence level of the damage assessment information satisfies the second path scheduling constraint, the repair verification node is called based on the damage assessment information to obtain the repair verification information of the reported vehicle. The second path scheduling constraint is the confidence level rule that needs to be satisfied for the vehicle damage assessment node to be scheduled to the repair verification node in the vehicle damage assessment system. Accordingly, generation module 306 is further configured as follows: By combining the damage assessment confidence level of the damage assessment information and the repair assessment confidence level of the repair assessment information, the target damage assessment result for the reported vehicle is generated.

[0139] Optionally, the damage assessment information includes the set repair price and / or set repair recommendations for the damaged parts of the reported vehicle, and the repair assessment information for the reported vehicle includes the reference repair price and / or reference repair recommendations for the damaged parts of the reported vehicle. Generation module 306 is further configured as follows: Based on the damage assessment confidence level of the damage assessment information, the repair assessment confidence level of the repair assessment information, and the set repair price and reference repair price of the damaged component, a repair price assessment result is generated. Based on whether the initial repair recommendations and the reference repair recommendations for the damaged parts are consistent, a repair recommendation damage assessment result is generated; Based on the repair price assessment results and / or repair recommendation assessment results, a target damage assessment result for the damaged vehicle is generated.

[0140] Optionally, the generation module 306 is further configured as follows: If the damage assessment confidence of the damage assessment information satisfies the third path scheduling constraint, and the repair assessment confidence of the repair assessment information of the reported vehicle satisfies the fourth path scheduling constraint, determine the degree of repair difference between the set repair price and the reference repair price of the damaged parts of the reported vehicle. Repair price assessment results are generated based on the degree of repair differences.

[0141] Optionally, the generation module 306 is further configured as follows: If the degree of difference in repair is less than or equal to the first threshold, the repair price assessment result is determined based on the reference repair price. If the degree of repair difference is greater than the first threshold and less than or equal to the second threshold, then the repair price assessment result is determined based on the set repair price. If the degree of difference in repair exceeds the second threshold, an abnormal alarm will be issued.

[0142] Optionally, the generation module 306 is further configured as follows: If the damage assessment confidence level of the damage assessment information satisfies the third path scheduling constraint, and the repair assessment confidence level of the repair assessment information of the reported vehicle satisfies the fourth path scheduling constraint, the target repair price is determined from the set repair price and the reference repair price based on the confidence level of the set repair price and the confidence level of the reference repair price, and the repair price assessment result is determined based on the target repair price.

[0143] Optionally, the generation module 306 is further configured as follows: Determine whether the recommended repair for the damaged component is consistent with the recommended repair. If the set repair recommendations for the damaged component are inconsistent with the reference repair recommendations, a repair recommendation assessment result for the damaged component will be generated based on the set conflict resolution rules.

[0144] Optionally, the vehicle damage assessment system also includes a compliance node; the device also includes a fourth acquisition module, configured to: Call the compliance node to perform risk detection based on at least one of the following information: damage report information, damage assessment information, damage verification information, and repair verification information, and obtain risk warning information; Generation module 306 is further configured as follows: The target damage assessment result is determined by combining the damage assessment confidence level of the damage assessment information, the maintenance assessment confidence level of the maintenance assessment information, and the risk warning information.

[0145] Optionally, the device also includes a verification module configured to: If the damage assessment confidence level of the damage assessment information does not meet the first path scheduling constraint, an additional verification node is called to verify the damage assessment information. If the verification passes, a damage verification node is called based on the damage assessment information to obtain the damage verification information.

[0146] The above is a schematic scheme of a vehicle damage assessment device according to this embodiment. It should be noted that the technical solution of this vehicle damage assessment device and the technical solution of the above-described vehicle damage assessment method belong to the same concept. For details not described in detail in the technical solution of the vehicle damage assessment device, please refer to the description of the technical solution of the above-described vehicle damage assessment method.

[0147] Figure 4 A structural block diagram of a computing device 400 according to one embodiment of this specification is shown. The components of the computing device 400 include, but are not limited to, a memory 410 and a processor 420. The processor 420 is connected to the memory 410 via a bus 430, and a database 450 is used to store data.

[0148] The computing device 400 also includes an access device 440, which enables the computing device 400 to communicate via one or more networks 460. Examples of these networks include Public Switched Telephone Network (PSTN), Local Area Network (LAN), Wide Area Network (WAN), Personal Area Network (PAN), or combinations of communication networks such as the Internet. The access device 440 may include one or more of any type of wired or wireless network interface (e.g., a network interface card (NIC)), such as an IEEE 802.11 Wireless Local Area Network (WLAN) wireless interface, a Wi-MAX (Worldwide Interoperability for Microwave Access) interface, an Ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a Bluetooth interface, or a Near Field Communication (NFC) interface.

[0149] In one embodiment of this specification, the aforementioned components of the computing device 400 and Figure 4 Other components, not shown, can also be connected to each other, for example, via a bus. It should be understood that... Figure 4 The block diagram of the computing device shown is for illustrative purposes only and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0150] The computing device 400 can be any type of stationary or mobile computing device, including mobile computers or mobile computing devices (e.g., tablet computers, personal digital assistants, laptop computers, notebook computers, netbooks, etc.), mobile phones (e.g., smartphones), wearable computing devices (e.g., smartwatches, smart glasses, etc.) or other types of mobile devices, or stationary computing devices such as desktop computers or personal computers (PCs). The computing device 400 can also be a mobile or stationary server.

[0151] The processor 420 is used to execute the following computer-executable instructions, which, when executed by the processor, implement the steps of the above-described vehicle damage assessment method.

[0152] The above is a schematic diagram of a computing device according to this embodiment. It should be noted that the technical solution of this computing device and the technical solution of the vehicle damage assessment method described above belong to the same concept. For details not described in detail in the technical solution of the computing device, please refer to the description of the technical solution of the vehicle damage assessment method described above.

[0153] An embodiment of this specification also provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the vehicle damage assessment method described above.

[0154] The above is an illustrative scheme of a computer-readable storage medium according to this embodiment. It should be noted that the technical solution of this storage medium and the technical solution of the vehicle damage assessment method described above belong to the same concept. For details not described in detail in the technical solution of the storage medium, please refer to the description of the technical solution of the vehicle damage assessment method described above.

[0155] An embodiment of this specification also provides a computer program, wherein when the computer program is executed in a computer, it causes the computer to perform the steps of the above-described vehicle damage assessment method.

[0156] The above is an illustrative example of a computer program according to this embodiment. It should be noted that the technical solution of this computer program and the technical solution of the aforementioned vehicle damage assessment method belong to the same concept. Details not described in detail in the computer program's technical solution can be found in the description of the technical solution of the aforementioned vehicle damage assessment method.

[0157] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require the specific or sequential order shown to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0158] The computer instructions include computer program code, which may be in the form of source code, object code, executable file, or certain intermediate forms. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording media, USB flash drive, portable hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium may be appropriately added or removed according to the requirements of patent practice. For example, in some regions, according to patent practice, computer-readable media may not include electrical carrier signals and telecommunication signals.

[0159] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the embodiments in this specification are not limited to the described order of actions, because according to the embodiments in this specification, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in this specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the embodiments in this specification.

[0160] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0161] The preferred embodiments disclosed above are merely illustrative of this specification. The optional embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. Clearly, many modifications and variations can be made based on the embodiments described herein. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the embodiments, thereby enabling those skilled in the art to better understand and utilize this specification. This specification is limited only by the claims and their full scope and equivalents.

Claims

1. A method of vehicle damage assessment, characterized by, The application discloses a scheduling node applied to a vehicle loss assessment system, and the vehicle loss assessment system further comprises a loss assessment node and a loss verification node. In response to a vehicle loss assessment task of a damaged vehicle, the loss assessment node is called based on damage information of the vehicle loss assessment task to obtain damage identification information of the damaged vehicle. When damage identification confidence of the damage identification information meets a first path scheduling constraint, the loss verification node is called based on the damage identification information to obtain damage verification information, wherein the first path scheduling constraint is a confidence rule required to be met in the vehicle loss assessment system from the loss assessment node to the loss verification node. Based on the damage verification information, a target loss assessment result of the damaged vehicle is generated.

2. The method of claim 1, wherein, The vehicle loss assessment system further comprises a repair verification node. After the loss assessment node is called based on the damage information of the vehicle loss assessment task to obtain the damage identification information of the damaged vehicle, the following steps are further included. When damage identification confidence of the damage identification information meets a second path scheduling constraint, the repair verification node is called based on the damage identification information to obtain repair verification information of the damaged vehicle, wherein the second path scheduling constraint is a confidence rule required to be met in the vehicle loss assessment system from the loss assessment node to the repair verification node. Correspondingly, the generation of the target loss assessment result of the damaged vehicle based on the damage verification information includes: The target loss assessment result of the damaged vehicle is generated in combination with damage verification confidence of the damage verification information and repair verification confidence of the repair verification information.

3. The method of claim 2, wherein, The damage verification information includes a set repair price and / or a set repair suggestion of a damaged component of the damaged vehicle, and the repair verification information of the damaged vehicle includes a reference repair price and / or a reference repair suggestion of the damaged component of the damaged vehicle. The generation of the target loss assessment result of the damaged vehicle in combination with the damage verification confidence of the damage verification information and the repair verification confidence of the repair verification information includes: Based on the damage verification confidence of the damage verification information, the repair verification confidence of the repair verification information and the set repair price and the reference repair price of the damaged component, a repair price loss assessment result is generated. Based on whether the set repair suggestion and the reference repair suggestion of the damaged component are consistent, a repair suggestion loss assessment result is generated. Based on the repair price loss assessment result and / or the repair suggestion loss assessment result, the target loss assessment result of the damaged vehicle is generated.

4. The method of claim 3, wherein, The generation of the repair price loss assessment result based on the damage verification confidence of the damage verification information, the repair verification confidence of the repair verification information and the set repair price and the reference repair price of the damaged component includes: When the damage verification confidence of the damage verification information meets a third path scheduling constraint and the repair verification confidence of the repair verification information of the damaged vehicle meets a fourth path scheduling constraint, a repair difference degree of the set repair price and the reference repair price of the damaged component of the damaged vehicle is determined. Based on the repair difference degree, the repair price loss assessment result is generated.

5. The method of claim 4, wherein, The generating the repair price loss assessment result based on the repair difference degree comprises: if the repair difference degree is less than or equal to a first threshold value, determining the repair price loss assessment result based on the reference repair price; if the repair difference degree is greater than the first threshold value and less than or equal to a second threshold value, determining the repair price loss assessment result based on the set repair price; if the repair difference degree is greater than the second threshold value, performing an abnormality alarm.

6. The method of claim 3, wherein, The generating the repair price loss assessment result based on the damage appraisal confidence of the damage appraisal information, the repair appraisal confidence of the repair appraisal information, and the set repair price and the reference repair price of the damaged component comprises: in a case where the damage appraisal confidence of the damage appraisal information satisfies a third path scheduling constraint and the repair appraisal confidence of the repair appraisal information of the damaged vehicle satisfies a fourth path scheduling constraint, determining a target repair price from the set repair price and the reference repair price based on the confidence of the set repair price and the confidence of the reference repair price, and determining the repair price loss assessment result based on the target repair price.

7. The method of claim 3, wherein, The generating the repair suggestion loss assessment result based on whether the set repair suggestion and the reference repair suggestion of the damaged component are consistent comprises: determining whether the set repair suggestion and the reference repair suggestion of the damaged component are consistent; if the set repair suggestion and the reference repair suggestion of the damaged component are inconsistent, generating the repair suggestion loss assessment result of the damaged component based on a set conflict resolution rule.

8. The method of claim 2, wherein, The vehicle loss assessment system further comprises a compliance node; before the generating the target loss assessment result of the damaged vehicle based on the damage appraisal confidence of the damage appraisal information and the repair appraisal confidence of the repair appraisal information, the vehicle loss assessment system further comprises: calling the compliance node to perform risk detection based on at least one of the damage information, the damage identification information, the damage appraisal information, and the repair appraisal information, and obtaining risk prompt information; the generating the target loss assessment result of the damaged vehicle based on the damage appraisal confidence of the damage appraisal information and the repair appraisal confidence of the repair appraisal information further comprises: determining the target loss assessment result based on the damage appraisal confidence of the damage appraisal information, the repair appraisal confidence of the repair appraisal information, and the risk prompt information.

9. The method according to any one of claims 1 to 8, characterized in that, After the calling the loss assessment node based on the damage information of the vehicle loss assessment task to obtain the damage identification information of the damaged vehicle, the vehicle loss assessment system further comprises: in a case where the damage identification confidence of the damage identification information does not satisfy the first path scheduling constraint, calling an additional verification node to verify the damage identification information, and in a case where the verification is passed, calling the damage appraisal node based on the damage identification information to obtain the damage appraisal information.

10. A vehicle damage assessment device characterized by comprising: The scheduling node applied to a vehicle loss assessment system, the vehicle loss assessment system further comprises a loss assessment node and a damage appraisal node, and the device comprises: The first obtaining module is configured to, in response to a vehicle damage assessment task of a damaged vehicle, call the damage assessment node based on damage information of the vehicle damage assessment task, and obtain damage identification information of the damaged vehicle; The second obtaining module is configured to, in a case where a damage identification confidence of the damage identification information meets a first path scheduling constraint, call the damage determination node based on the damage identification information, and obtain damage determination information, wherein the first path scheduling constraint is a confidence rule required to be met by the vehicle damage assessment system when the damage assessment node is scheduled to the damage determination node. The generating module is configured to generate a target damage assessment result of the damaged vehicle based on the damage determination information.

11. A computing device, comprising: Comprising: a memory and a processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions, and the computer executable instructions, when executed by the processor, realize the steps of the vehicle damage assessment method in any one of claims 1-9.

12. A computer-readable storage medium, characterized in that, The memory has stored computer executable instructions, and the computer executable instructions, when executed by the processor, realize the steps of the vehicle damage assessment method in any one of claims 1-9.

13. A computer program product, characterised in that, Comprising computer program / instructions, which, when executed by the processor, realize the steps of the vehicle damage assessment method in any one of claims 1-9.

Images