Method for quickly searching door-to-door installation and maintenance service

Abstract

The invention relates to the technical field of mobile internet, and discloses a method for quickly searching door-to-door installation and maintenance services, which comprises the following steps of: acquiring visual data of a service target through an AR (Augmented Reality) interface of a user terminal, receiving an interaction label, and generating structured service demand data containing a three-dimensional scene reference model; matching the demand data with dynamic portraits of a plurality of service providers, wherein the dynamic portraits at least comprise real-time schedulability indexes calculated based on real-time states and positions and scenarized skill vectors generated based on historical data; the matching result is overlaid in an AR interface in a visual element form to be pushed, and the element associates and displays a reference quotation estimated based on the demand data; and according to user selection, synchronizing the demand data to a service provider terminal to support AR-assisted performance. According to the method, visual structured analysis of service requirements, multi-dimensional dynamic accurate matching of service providers and visual collaboration of service processes are realized, and the search efficiency, matching accuracy and user experience of door-to-door services are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of mobile internet technology, and more specifically to a method for quickly finding on-site installation and repair services. Background Technology

[0002] With the widespread adoption of mobile internet, online platforms for finding on-site installation and repair services (such as appliance repair, furniture installation, and drain cleaning) have become mainstream. Existing solutions primarily rely on the following models: The platform's search model, based on keywords and static information, involves users entering service keywords (such as "air conditioner repair" or "toilet installation") and a rough location into an application or website. The platform then matches these with static tags (such as skill categories) provided by the service providers and their fixed geographical locations, returning a list. Users must repeatedly describe the situation and send photos to multiple service providers via phone or online chat to obtain a rough quote and schedule an appointment. This method suffers from inefficient information transmission and inaccurate descriptions, preventing service providers from accurately assessing the problem before arrival. This leads to significant quote discrepancies, the need for secondary visits, or additional charges, resulting in a poor user experience.

[0003] Simple geolocation dispatching model: Some platforms attempt to introduce LBS (Location-Based Services) to prioritize pushing services from the nearest provider to the user. However, this only optimizes travel time and does not consider the service provider's real-time operational status (whether they are already on a task, the estimated completion time of the task), their skill level for the specific problem, or whether they have the appropriate spare parts. The result may be that the nearest service provider is not adept at handling that specific problem, or is already on other tasks and cannot actually respond quickly.

[0004] Information asymmetry and lack of visual assessment: Throughout the entire process, from requirement description and quotation to construction, there is a lack of an objective and visual benchmark. Users' written or verbal descriptions often differ from the actual on-site situation; service providers' quotations are based on vague descriptions and lack supporting evidence; users are also unable to make an intuitive and credible judgment on the service provider's suitability before receiving service. This information black box throughout the entire process is the main cause of trust deficiency and disputes.

[0005] In summary, the shortcomings of existing technologies can be summarized as follows: (1) The demand side lacks the ability to transform non-standardized on-site scenarios into structured and computable data, relying on inefficient manual descriptions; (2) The matching side relies on static and single dimensions (keywords, static locations), failing to integrate multi-dimensional dynamic information such as the real-time status, refined skills, and resource load of service providers; (3) The interaction side lacks transparent and intuitive visualization media, making the matching results, pricing basis, and service process opaque.

[0006] Based on this, the present invention proposes a method for quickly finding on-site installation and repair services to solve the above problems. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method for quickly finding on-site installation and repair services to solve the problems existing in the background art.

[0008] This invention provides the following technical solution: a method for quickly locating on-site installation and repair services, comprising the following steps: S1: Augmented Reality Acquisition and Structuring Steps for Demand Scenarios: Responding to user operations, the augmented reality (AR) interface of the smart terminal is activated; the user is guided to scan the service target through the AR interface, and interactive annotation information from the user in the AR screen is received; based on the visual data acquired by scanning and the interactive annotation information, structured service demand data is generated, the service demand data including a three-dimensional scene reference model constructed based on the visual data; S2: Service Provider Dynamic Matching Step: Based on the structured service demand data, perform matching calculations with the dynamic profiles of multiple service providers; wherein, the dynamic profile includes at least a real-time schedulability index calculated based on the service provider's real-time geographical location and current task status, and a scenario-based skill vector generated based on historical service data and associated with a specific service scenario; S3: Augmented Reality Visualization Push Step: The information of at least one recommended service provider obtained from the matching calculation is overlaid on the AR interface of the user terminal in the form of a visual element and displayed. The visual element is associated with a reference price estimated based on the service demand data. S4: Task Data Synchronization and Fulfillment Assistance Steps: In response to the user's instruction to select a recommended service provider, the structured service requirement data containing the 3D scene reference model is synchronized to the selected service provider's terminal; and AR assistance information is provided based on the 3D scene reference model during the service fulfillment process.

[0009] As a further aspect of the present invention: in step S1, generating structured service requirement data specifically includes: analyzing the visual data using a computer vision model, identifying the category and / or key components of the service target, and automatically generating the structured service requirement data containing the estimated operation complexity coefficient and / or the predicted list of required materials in combination with the interactive annotation information.

[0010] As a further aspect of the present invention: in step S2, the dynamic profile also includes the real-time resource load information of the service provider, which reflects the current spare parts inventory of the service provider; the matching calculation is further performed based on the degree of fit between the required material prediction list and the real-time resource load information.

[0011] As a further aspect of the present invention: in step S2, the matching calculation is also performed based on the similarity between the three-dimensional scene reference model or the features extracted from it and the scene model corresponding to the work orders completed in the past by the service provider.

[0012] As a further aspect of the present invention: in step S3, the visualization element is an interactive information board floating in the AR real-world scene corresponding to the virtual location of the service provider, and the reference price displayed on the interactive information board is a price range calculated based on the operation complexity coefficient and / or the required material prediction list.

[0013] As a further aspect of the present invention: in step S4, the provision of AR auxiliary information during service fulfillment includes: providing the service provider terminal with an access interface for the three-dimensional scene reference model, so that the service provider can access the fault points or installation locations pre-marked by the user in the AR view.

[0014] As a further aspect of the present invention: in step S4, the provision of AR auxiliary information during service fulfillment includes: providing the service provider terminal with an access interface for the three-dimensional scene reference model, so that the service provider can access the fault points or installation locations pre-marked by the user in the AR view.

[0015] A system for quickly locating on-site installation and repair services, for implementing the method as described in any one of the above statements, the system comprising: The user terminal module is equipped with an AR acquisition and interaction unit, which is used to execute steps S1 and S3. The service provider terminal module is used to receive synchronized task data and participate in contract fulfillment. A service matching engine is used to maintain dynamic profiles of service providers and perform the matching calculation in step S2. The data synchronization and service management module is used to perform task data synchronization and process coordination in step S4.

[0016] A computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the method as described in any one of the preceding descriptions.

[0017] The technical effects and advantages of this invention are as follows: This invention guides users to use an AR interface for intuitive scene scanning and annotation, transforming traditional vague and subjective text or voice descriptions into structured data containing three-dimensional spatial information and visual features. This method effectively eliminates information ambiguity between users and service providers, enabling service providers to accurately understand the problem, environmental conditions, and required resources before arriving on site. This directly reduces repeated communication to confirm the problem and lowers the probability of invalid on-site visits or secondary services due to misjudgment of information. This invention differs from traditional matching based on static labels and coarse distance. It integrates multi-dimensional dynamic factors such as service provider real-time schedulability indicators, scenario-based skill vectors, real-time resource load, and historical scenario matching degree. This matching mechanism not only considers "who can do it" but also evaluates "who is most suitable to do it now" and "who is best able to do this specific problem well." As a result, the system can achieve better task allocation from a global perspective, improve the overall operational efficiency of the service network, and match users with the service provider that has the fastest comprehensive response, the most professional expertise, and the most thorough preparation. From the search stage, users can intuitively view key information of recommended service providers and analysis-based reference quotes in AR real-world scenes, making decision-making clearer and more transparent. In the fulfillment stage, the AR-assisted function provides both parties with a collaborative tool based on the same visualized scene model, enabling users to perceive service progress and service providers to accurately locate problems. The full-process visualized interaction significantly enhances users' sense of control and trust in the service process. By receiving a structured requirement data package (including a 3D scene model), the service provider of this invention can accurately assess the service task before departure, thereby preparing tools and spare parts in a targeted manner. This improves the success rate of solving problems on the first visit, reduces round-trip costs and time waste caused by insufficient preparation, and optimizes the individual's workflow and resource management. The method of this invention continuously generates high-quality structured service scenario data (problem-scenario-solution association data); this data can be used to continuously optimize the identification and matching algorithm, provide a case library for service provider skills training, and even provide data support for predicting regional and seasonal service demand trends, thereby enhancing the intelligence level and long-term competitiveness of the entire service ecosystem. Attached Figure Description

[0018] The invention will now be further described with reference to the accompanying drawings.

[0019] Figure 1 This is a flowchart of a method for quickly finding on-site installation and repair services according to the present invention; Figure 2 This is a system block diagram of a method for quickly finding on-site installation and repair services according to the present invention. Detailed Implementation

[0020] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. These embodiments are only used to explain the present invention and are not intended to limit the scope of protection of the present invention.

[0021] Please see Figure 1 As shown, a method for quickly finding on-site installation and repair services includes the following steps: S1: Augmented Reality Acquisition and Structured Process for Requirement Scenarios: Respond to user operation and launch the augmented reality (AR) interface of the smart terminal; guide the user to scan the service target through the AR interface and receive the user's interactive annotation information in the AR screen; based on the visual data and interactive annotation information obtained from the scan, generate structured service requirement data, which includes a 3D scene reference model built based on the visual data. In one specific embodiment, after the user opens the App and selects "Publish Request," the camera is activated, and the AR data acquisition interface is entered. The interface provides guidance prompts (such as "Please scan around the device"). The user scans the target object (such as a faulty refrigerator) according to the prompts. During the scanning process, the App uses SLAM (Simultaneous Localization and Mapping) technology, such as using APIs provided by ARKit or ARCore, to build environmental point clouds in real time and locate the target, generating a preliminary 3D scene reference model. At the same time, the user can select specific parts on the screen (such as the "refrigerator door seal") with their finger and select preset problem labels (such as "not sealed" or "damaged") or make a voice description. The App packages and uploads the collected visual data stream (keyframe images), the coordinates and label information marked by the user, and the speech-to-text results to the cloud. After receiving the data, the cloud first uses a pre-trained convolutional neural network (CNN) model (such as ResNet) to analyze the keyframe images, identify the object category as "refrigerator," and detect the brand logo. Simultaneously, another dedicated component detection model identifies key components such as "door seal" and "compressor." Combining this with the user's selected tag "door seal - not sealed," the system automatically structures this information, generating a structured service request data in JSON format. This data includes at least: {"Category": "Refrigerator," "Suspected Faulty Component": "Door Seal," "Problem Description": "Not Sealed Properly," "3D Scene Model ID": "xxx"}. This 3D model can be stored in the cloud in a lightweight format (such as .glb or .usdz) and associated with the request form. S2: Service Provider Dynamic Matching Steps: Based on structured service demand data, perform matching calculations with the dynamic profiles of multiple service providers; among which, the dynamic profiles include at least real-time schedulability indicators calculated based on the service provider's real-time geographical location and current task status, as well as scenario-based skill vectors generated based on historical service data and associated with specific service scenarios; In one specific embodiment, a dynamic profile database of service providers is maintained in the cloud; when a new request (such as the refrigerator repair mentioned above) is generated, the matching engine is activated; Real-time schedulability metric calculation: The engine queries all service providers with the tag "home appliance repair"; for each service provider, it reads their current status (e.g., "working" or "idle"), the GPS coordinates of the current work order, and the estimated completion time of the current work order; through a route planning API (e.g., integrated with the Gaode Map API), it calculates the travel time from the current location or the estimated location after the completion of the current work order to the user's address, and combines this with the status to generate an estimated response time (e.g., service provider A, can depart in 30 minutes, travel time 20 minutes, total 50 minutes). Contextualized skill vector matching: Each service provider's skill vector is trained from its historical work orders; for example, if service provider B has a high number of successful repairs of "refrigerator door seals", short processing time, and high positive feedback rate in its historical work orders, then its vector will have a high score in the "refrigerator-door seal-repair" dimension; the engine calculates the cosine similarity between the features of the current demand ("refrigerator", "door seal", "sealing problem") and each dimension of the service provider's skill vector to obtain the skill matching score; Comprehensive matching calculation: The matching engine uses a weighted scoring algorithm: Comprehensive score = a * (1 / estimated response time) + b * skill matching score + c * historical positive review rate + ... (a, b, c are adjustable weights); all potential service providers are sorted according to the score; The following section provides a detailed explanation of the quantitative construction method for "scenario-based skill vectors" and the optimization strategy for weight coefficients in the comprehensive matching algorithm: Construction and quantification of scenario-based skill vectors: A scenario-based skill vector is a multi-dimensional vector that structurally and digitally represents a service provider's historical service capabilities; its construction process includes two main stages: feature engineering and vectorization. Feature engineering phase: Structured Tag Extraction: The system extracts pre-defined multi-level structured tags from the service provider's successfully completed historical work orders. These tags include at least: "Service Category" (e.g., home appliances, furniture, pipes), "Specific Object" (e.g., refrigerator, air conditioner, wardrobe), "Problem / Operation Type" (e.g., repair, installation, unclogging), and "Specific Component / Fault Point" (e.g., door seal, compressor, water pipe interface). For example, a work order can be tagged as [home appliances, refrigerator, repair, door seal]; Unstructured text analysis: Use natural language processing (NLP) techniques to extract keywords and analyze topics from the text descriptions, user reviews, and service summaries in work orders; for example, use TF-IDF (term frequency-inverse document frequency) algorithms or pre-trained word embedding models (such as Word2Vec, BERT) to extract technical keywords such as "poor cooling", "refrigerant leak", and "replace sealing ring". Performance Indicator Association: Associate the quantitative performance indicators of each work order, such as task completion time (compared with the average time of similar tasks), user rating, first-time resolution rate, rework rate, etc., with the above tags. Vectorization representation stage: Dimension definition: Each dimension of the vector corresponds to a normalized "capability feature"; these features are composed of the above labels and indicators; for example, one dimension could be the proficiency in "refrigerator-door seal-repair", and another dimension could be the "efficiency in handling problems related to poor cooling". Weight calculation (i.e., the values ​​of each dimension of the vector): For an individual service provider, the value of its vector in a specific dimension (such as the "refrigerator-door seal-repair" dimension) is calculated in the following way: Frequency weight: Count the number of work orders corresponding to this dimension that have been completed in the past and normalize them (e.g., divide by the total number of work orders of this service provider, or the maximum value of this dimension across the entire platform). Quality weight: Calculate the average user rating and average efficiency coefficient (e.g., standard time / actual time) of all work orders under this dimension. Comprehensive assignment: The final value of this dimension is set as a weighted sum of frequency weight and quality weight. For example: Dimension value = α * normalized frequency + β * average score + γ * average efficiency coefficient, where α, β, and γ are adjustable internal coefficients used to balance quantity and quality. Vector storage and updating: Each service provider's scenario-based skill vector is stored in the database as a high-dimensional array; whenever a service provider completes a new work order and receives closed-loop feedback, the system automatically updates the corresponding dimension values ​​of its skill vectors incrementally based on the new work order's tags and performance data, realizing the dynamic evolution of capabilities; Training and optimization of matching algorithm weights: The weight coefficients of each factor in the comprehensive matching calculation model (such as a, b, c mentioned above) are not fixed, but are initialized and continuously optimized through machine learning methods based on historical data to ensure that the matching results can maximize the overall service network efficiency (such as average response time, overall user satisfaction, and service provider capacity utilization). Determining the initial weights (offline training): Training data preparation: Collect historical completed work order data with clear results over a period of time as the training set; each data point includes: the demand characteristics at that time (which can be converted into a structured vector), the dynamic profile data of all candidate service providers at that time (schedulable time, skill vectors at that time, etc.), and the final service result labels (such as: actual completion time, user five-star rating, whether rework occurred). Optimization objective definition: Set the optimization objective function; for example, the objective could be "maximize the probability of predicting successful service" or "minimize the combined loss of prediction completion time and user satisfaction"; Model training: Supervised learning algorithms (such as gradient descent and genetic algorithms) are used to train the weight coefficients in the matching model; the system tries different weight combinations to simulate the matching decisions of historical work orders and compares the simulation results with actual excellent results (such as work orders that are highly rated by users and completed efficiently); through continuous iteration, a set of weight coefficients that makes the simulated decisions closest to the historical excellent results is found and used as the initial weights of the model. Dynamic adjustment of weights (online learning): Feedback loop establishment: The system records each matching decision (i.e., order dispatch) and its subsequent complete fulfillment results (including user rating, service provider feedback, actual time consumption, whether secondary service is required, etc.) as a new training sample; Periodic retraining: The system periodically (e.g., weekly or monthly) uses newly added sample data to fine-tune or retrain the weight coefficients, enabling the matching model to adapt to changes in service provider capabilities, seasonal demand fluctuations, and adjustments to business strategies. A / B testing verification: Within a controllable range, the system can deploy multiple sets of different weight configurations to conduct A / B testing. By comparing the performance of key indicators (such as order conversion rate and overall positive review rate) under different configurations, the optimal combination of weight parameters can be scientifically selected. S3: Augmented Reality Visual Push Steps: Display at least one recommended service provider information obtained from the matching calculation as a visual element overlaid on the AR interface of the user terminal, and display the reference price estimated based on the service demand data in association with the visual element. In one specific embodiment, after the matching and sorting are completed, the cloud sends the information of the top 3-5 recommended service providers (including avatar, nickname, comprehensive score, estimated response time, skill matching details) and the reference price range (such as "150-300 yuan") estimated based on the category, faulty parts and publicly available industry price data model to the user's App; The AR interface of the user app is reactivated; next to the 3D model of the refrigerator that was just scanned and created, a virtual logo of the top-ranked service provider is displayed as a floating information panel (UI Panel); the user can rotate the phone to view it from different angles; clicking on the information panel expands to display details, including the reference price range "150-300 yuan", the estimated on-site arrival time "about 50 minutes later", and the matching reason "proficient in handling refrigerator sealing problems"; S4: Task Data Synchronization and Fulfillment Assistance Steps: In response to the user's instruction to select a recommended service provider, the structured service requirement data containing the 3D scene reference model is synchronized to the selected service provider's terminal; and AR assistance information based on the 3D scene reference model is provided during the service fulfillment process. In one specific embodiment, after the user clicks to select and confirm the reservation of the service provider, a formal order is created in the cloud, and the complete structured service requirement data package (including the 3D scene model access link) generated by S1 is pushed to the service provider's App; After accepting an order, the service provider can click "View Site" in the "To-Do Orders" section of their app to directly download and load the 3D scene reference model of the refrigerator and preview the "door seal" problem points previously marked by the user in AR mode. Before setting off, they can prepare the corresponding replacement parts based on the prediction of the "door seal". When on-site repair, they can open the AR view again, align the virtual markings with the actual refrigerator, and perform precise repairs. Specifically, in step S1, generating structured service requirement data includes: using a computer vision model to analyze visual data, identifying the category and / or key components of the service target, and combining interactive annotation information to automatically generate structured service requirement data containing the estimated operation complexity coefficient and / or the predicted list of required materials. In one specific embodiment, the specific applications of the computer vision model include: a category recognition model deployed in the cloud uses a ResNet-50 model pre-trained on the ImageNet dataset and fine-tuned on a self-built "home appliance product image dataset"; a component detection model uses the YOLOv5 architecture and is trained on a dataset labeled with various home appliance components (such as air conditioner heat sinks, water pipe interfaces, and wardrobe connecting plates); the system combines the "category: refrigerator, confidence level 98%" identified by the model with the "door seal" labeled by the user, and automatically generates an estimated operation complexity coefficient (set to medium "3" based on the steps required to replace the door seal) and a predicted list of required materials by querying a predefined "repair knowledge graph" (which stores common faulty components, corresponding common tools and materials required for the problem). Specifically, in step S2, the dynamic profile also includes the service provider's real-time resource load information, which reflects the service provider's current spare parts inventory; the matching calculation is further based on the degree of fit between the required material prediction list and the real-time resource load information. In one specific embodiment, the real-time resource load information in the service provider's dynamic profile is manually updated by the service provider through its App or automatically reported through IoT tags (such as toolkit chips identified by Bluetooth). For example, after completing the previous order, service provider B changes the quantity of "Door Seal (Model F-001)" from 1 to 0 in the "My Toolbox" list of its App. When the matching engine processes the refrigerator repair request, it checks whether the "New Door Seal" in the predicted material list exists in the service provider's real-time resource load list. If it does, a resource fit bonus is added to the matching calculation, thereby prioritizing the recommendation of service providers with stock and reducing the need for secondary visits due to missing parts. Specifically, in step S2, the matching calculation is also based on the similarity between the 3D scene reference model or the features extracted from it and the scene model corresponding to the work orders completed by the service provider in the past. In one specific embodiment, the system also calculates the historical scene matching degree; the cloud database stores the 3D scene models (after anonymization) corresponding to the service provider's historical successful work orders; when the 3D scene model of the new requirement is generated, the engine extracts its geometric features (such as feature vectors extracted using the PointNet network) and texture features; then, it compares the similarity of the model with the feature vectors of the service provider's historical work order scenes (such as calculating the Euclidean distance); if service provider C has repaired similar refrigerator models many times, and its historical scene model is highly similar to the current requirement model, then this item has a high score; this score participates in the comprehensive weighted calculation in S102, so that service providers with experience in handling highly similar on-site environments are given priority recommendation; Specifically, in step S3, the visualization element is an interactive information board floating in the AR real-world scene corresponding to the virtual location of the service provider. The reference price displayed on the interactive information board is a price range calculated based on the operation complexity coefficient and / or the expected list of required materials. In one specific embodiment, the visualization element in step S3—the interactive information board—is implemented as a preset 3DUI object in Unity3D or Apple RealityKit engine; the object is bound to the spatial coordinates of the three-dimensional scene reference model; its display content includes: service provider avatar (circular sprite), name text, a dynamically generated price range text (content derived from "150-300 yuan" sent from the cloud), and a "details" button; when the "details" button is clicked, the information board will expand a secondary panel, listing key matching reasons in text form, such as: "1. Closest to you; 2. Completed 5 similar repairs this month with a 100% positive feedback rate; 3. Currently has relevant spare parts"; the price range is calculated based on the operation complexity coefficient (3) multiplied by a basic labor cost (e.g., 50 yuan), plus the sum of the average market price of the materials in the predicted material list, generated through a preset algorithm model; Specifically, in step S4, supporting the provision of AR-assisted information during service fulfillment includes: providing an access interface for a 3D scene reference model to the service provider's terminal, enabling the service provider to access the fault points or installation locations pre-marked by the user in the AR view; In one specific embodiment, during the fulfillment assistance phase of S4, the AR assistance function provided by the service provider terminal is as follows: the service provider's app provides an "AR navigation" button on the order details page; after clicking, the camera is activated and a lightweight 3D scene model downloaded from the cloud is loaded; the app aligns the virtual model with the real world through visual inertial odometry (VIO); at this time, the "door seal" position marked by the user during the demand collection phase will be superimposed on the actual refrigerator door seal position as a highlighted 3D arrow or aperture, guiding the service provider to quickly locate the problem point without requiring the user to point it out again; Specifically, in step S4, supporting the provision of AR-assisted information during service fulfillment includes: providing an access interface for a 3D scene reference model to the service provider's terminal, enabling the service provider to access the fault points or installation locations pre-marked by the user in the AR view; In one specific embodiment, during the service fulfillment process, the system pushes AR status prompts to the user's terminal; when the service provider starts the repair, they click "Start replacing door seal" in their App; this status is pushed to the user's App in real time via WebSocket; the user opens the "View Progress" AR interface in the App and points it at the refrigerator being repaired; a semi-transparent progress bar animation and status text will be superimposed on the refrigerator door seal position on the screen: "In service: Replacing door seal (70% complete)"; this is achieved by associating the progress status data with the coordinates of specific components in the 3D scene model, thereby improving the transparency of the service process; Please see Figure 1As shown, a system for quickly locating on-site installation and repair services is provided to implement any of the methods described above. The system includes: The user terminal module is equipped with an AR acquisition and interaction unit, which is used to execute steps S1 and S3. The service provider terminal module is used to receive synchronized task data and participate in contract fulfillment. The service matching engine is used to maintain dynamic profiles of service providers and perform matching calculations in step S2. The data synchronization and service management module is used to perform task data synchronization and process coordination in step S4. In one specific embodiment, the hardware and software modules are configured as follows: User terminal module: physically, it is a smartphone; software includes: AR Acquisition and Interaction Unit: Calls the iOS ARKit framework to realize scene scanning, model building, screen annotation and voice input; AR Visualization Rendering Unit: Using the SceneKit engine, it is responsible for receiving and rendering virtual information boards from service providers sent from the cloud; Service provider terminal module: The physical form is a smartphone or tablet; the software includes units such as task reception, resource status reporting, AR preview and navigation; Cloud-based matching engine: Deployed on cloud servers (such as Alibaba Cloud ECS); developed using Java / Python, including a weighted scoring algorithm module, a real-time path calculation module (calling map API), and a skill vector calculation module; Data synchronization and service management module: Also deployed in the cloud, serving as the core of business logic, using a microservice architecture; responsible for user / service provider data management, order state machine transitions, 3D model file storage and distribution (using object storage services such as OSS), and real-time message push between user and service provider terminals (using message queue MQTT service). The modules exchange data through the defined RESTful API interface and work together to complete the closed loop from requirement release to service completion; A computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the method as described above. In one specific embodiment, a computer-readable storage medium, such as an SSD or read-only memory (ROM) of a server, stores a computer program (including client-side App code, service provider-side App code, and cloud service program code). When the processor of a user's mobile phone or service provider's mobile phone loads and runs the client-side or service provider-side program, the corresponding steps of the above method can be executed. When the processor of the cloud server loads and runs the server-side program, the corresponding steps of the above method (such as matching calculation, data synchronization, etc.) can be executed. The program code includes all the key algorithm calls, API interface requests, and data processing logic described in the foregoing embodiments.

[0022] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for quickly locating on-site installation and repair services, characterized in that, Includes the following steps: S1: Augmented Reality Acquisition and Structured Process for Required Scenarios: Respond to user operation and launch the augmented reality (AR) interface of the smart terminal; guide the user to scan the service target through the AR interface and receive the user's interactive annotation information in the AR screen; Based on the visual data obtained from the scan and the interactive annotation information, structured service requirement data is generated, and the service requirement data includes a three-dimensional scene reference model constructed based on the visual data. S2: Service Provider Dynamic Matching Step: Based on the structured service demand data, perform matching calculations with the dynamic profiles of multiple service providers; wherein, the dynamic profile includes at least a real-time schedulability index calculated based on the service provider's real-time geographical location and current task status, and a scenario-based skill vector generated based on historical service data and associated with a specific service scenario; S3: Augmented Reality Visualization Push Step: The information of at least one recommended service provider obtained from the matching calculation is overlaid on the AR interface of the user terminal in the form of a visual element and displayed. The visual element is associated with a reference price estimated based on the service demand data. S4: Task Data Synchronization and Fulfillment Assistance Steps: In response to the user's instruction to select a recommended service provider, the structured service requirement data containing the 3D scene reference model is synchronized to the selected service provider's terminal; and AR assistance information is provided based on the 3D scene reference model during the service fulfillment process.

2. The method according to claim 1, characterized in that, In step S1, generating structured service requirement data specifically includes: analyzing the visual data using a computer vision model to identify the category and / or key components of the service target, and automatically generating the structured service requirement data containing the estimated operation complexity coefficient and / or the predicted list of required materials, in conjunction with the interactive annotation information.

3. The method according to claim 1, characterized in that, In step S2, the dynamic profile also includes the service provider's real-time resource load information, which reflects the service provider's current spare parts inventory. The matching calculation is further performed based on the degree of fit between the required material prediction list and the real-time resource load information.

4. The method according to claim 1, characterized in that, In step S2, the matching calculation is also performed based on the similarity between the three-dimensional scene reference model or the features extracted from it and the scene model corresponding to the work orders completed in the past by the service provider.

5. The method according to claim 1, characterized in that, In step S3, the visualization element is an interactive information board floating in the AR real-world scene corresponding to the virtual location of the service provider. The reference price displayed on the interactive information board is a price range calculated based on the operation complexity coefficient and / or the required material prediction list.

6. The method according to claim 1, characterized in that, In step S4, the provision of AR-assisted information during service fulfillment includes: providing the service provider's terminal with an access interface for the three-dimensional scene reference model, enabling the service provider to view the fault points or installation locations pre-marked by the user in the AR view.

7. The method according to claim 1, characterized in that, Step S4 further includes: during the service fulfillment process, pushing AR status prompt information related to the service progress to the user terminal, wherein the status prompt information is associated with a specific location in the three-dimensional scene reference model.

8. A system for quickly locating on-site installation and repair services, characterized in that, The system for implementing the method as described in any one of claims 1-7 comprises: The user terminal module is equipped with an AR acquisition and interaction unit, which is used to execute steps S1 and S3. The service provider terminal module is used to receive synchronized task data and participate in contract fulfillment. A service matching engine is used to maintain dynamic profiles of service providers and perform the matching calculation in step S2. The data synchronization and service management module is used to perform task data synchronization and process coordination in step S4.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-7.

Images