A clamp intelligent detection system and detection method with real-time three-dimensional visualization monitoring

Abstract

The application discloses a kind of fixture intelligent detection system and detection method with real-time three-dimensional visualization monitoring, the system includes: design data analysis and task generation module, for extracting measurement control point from three-dimensional design model and automatically generating structured detection task;Mobile guidance and data acquisition module, through industrial terminal equipped with special client, with three-dimensional visualization interface, sequentially measured by operator is forcibly guided, and automatically bind the measured data returned by measuring equipment;Real-time monitoring and decision analysis module is deployed in server end, for receiving data and driving Web three-dimensional model real-time update measuring point state, process monitoring and early warning are carried out, and report containing deviation chromatogram and statistical analysis is automatically generated.The method realizes the mistake of detection process, real-time closed loop of data flow, intuitive and forward-looking of quality decision, significantly improves the reliability, efficiency and intelligent level of large complex fixture detection.

Description

Technical Field

[0001] This invention relates to the field of industrial digital inspection and quality control technology, specifically to an intelligent fixture inspection system and method with real-time three-dimensional visualization monitoring. Background Technology

[0002] In modern high-end equipment manufacturing fields, such as automobile body welding, aerospace component assembly, and ship section construction, large, precision modular tooling fixtures (hereinafter referred to as "fixtures") are key process equipment for ensuring product manufacturing accuracy and consistency. The precision of the fixture itself directly determines the dimensional quality of the assembled product. Therefore, regular and comprehensive precision testing and maintenance of these fixtures is an indispensable part of the manufacturing quality control system.

[0003] Currently, the common process for on-site accuracy inspection of large and complex fixtures in the industry is as follows: operators use portable coordinate measuring machines (such as articulated arm measuring machines, laser trackers, etc.) to go deep into the production site and perform contact-based measurements on hundreds or even thousands of designated measurement control points on the massive fixture. However, this traditional human-machine collaborative inspection mode is increasingly revealing the following urgent technical bottlenecks and systemic defects in the context of digital transformation and intelligent manufacturing's pursuit of high efficiency and high reliability:

[0004] (1) The testing process is highly dependent on human experience and condition, resulting in a high risk of missed or incorrect measurements. Existing testing operations lack intelligent process guidance and error prevention mechanisms. Operators typically need to manually identify and locate each point to be measured on complex two-dimensional engineering drawings or static electronic PDF files on intricately structured fixtures. This process is highly susceptible to systematic missed measurements due to visual fatigue, distraction, or misunderstanding of the drawings. At the same time, the measurement sequence lacks mandatory constraints, and operators may skip measurements based on on-site convenience or memory, further increasing the risk of missed measurements and making it impossible to verify the completeness of the task in real time during the measurement process.

[0005] (2) Data acquisition is isolated, and processing and analysis are severely delayed. In the current process, the raw coordinate data acquired by the measuring equipment is usually recorded manually in paper forms or temporarily stored in the measuring equipment controller. The comparison and analysis with theoretical values ​​need to be performed by engineers on their office computers using independent software after the measurement is completed. This "acquisition-recording-offline analysis" model causes a break in the data chain, resulting in low efficiency and making it easy to introduce human errors during multiple transcription and processing. More importantly, quality management personnel cannot obtain real-time data and progress during the inspection process and can only passively wait for the final analysis report. Once the report shows that the fixture is out of tolerance, it has often caused assembly deviation and rework of a batch of products, and quality loss has already occurred, making it impossible to achieve "in-process intervention" and real-time decision-making.

[0006] (3) The testing task planning is disconnected from authoritative data in product design, resulting in repetitive, inefficient, and potentially biased planning. During the product design phase, engineers have precisely defined the set of measurement control points (MCPs) for the fixture in the 3D computer-aided design (such as CATIA, NX, etc.) model, which includes authoritative information such as the theoretical coordinates, normals, and tolerances of the points. However, existing general-purpose data acquisition devices or testing management software typically cannot directly read and utilize this structured design data. The formulation of the testing plan requires testing engineers to manually reinterpret the 2D drawings or 3D models and re-identify and input the measurement point information. This is not only repetitive and wasteful of manpower, but may also introduce misunderstandings during the interpretation process, leading to discrepancies between the testing task and the design intent.

[0007] (4) The data analysis and result presentation methods are simplistic, report generation consumes a lot of manpower, and value mining is insufficient. Currently, most measurement results are presented in the form of numerical tables, lacking intuitive spatial quality distribution visualization (such as deviation chromatograms) that are directly related to the three-dimensional entity of the fixture. This makes it difficult to quickly locate problem areas and understand the spatial correlation of deviations. At the same time, generating a professional inspection report that includes statistical process control (SPC) charts and historical trend comparisons requires engineers to spend a lot of time on data organization, chart creation, and document editing, which is inefficient and difficult to standardize. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention provides an intelligent fixture inspection system and method with real-time three-dimensional visualization monitoring. This system overcomes the deficiencies of existing technologies, achieves error prevention and omission prevention in the inspection process, real-time closed-loop data flow, and intuitive and forward-looking quality decision-making, significantly improving the reliability, efficiency, and intelligence level of inspection of large and complex fixtures.

[0009] To achieve the above objectives, the present invention provides the following technical solution:

[0010] A fixture intelligent inspection system with real-time 3D visualization monitoring includes:

[0011] The design data parsing and task generation module is used to parse the 3D design model file of the fixture, extract the set of measurement control points defined therein and the theoretical data of each measurement point, and map the set of measurement control points to a lightweight 3D model for visualization display, so as to automatically generate a structured inspection task containing the measurement point sequence.

[0012] The mobile guidance and data acquisition module includes a mobile terminal, which is used to load and execute structured detection tasks on site, force the operator to perform measurements according to the measurement point sequence through a three-dimensional visualization interface, receive measured data from the measuring equipment, and bind the measured data with the corresponding theoretical data to form a detection record;

[0013] The real-time monitoring and decision analysis module, deployed on a server or in the cloud, communicates with the mobile guidance and data acquisition module to receive the detection records in real time, dynamically update the status of each measurement control point in a web-based 3D visualization view, perform process monitoring and analysis, and automatically generate a detection report based on the complete detection records.

[0014] Preferably, the design data parsing and task generation module is specifically configured as follows:

[0015] Read and parse the 3D model file from the computer-aided design system, and extract the measurement control point information therein. The information includes at least the point identifier, theoretical 3D coordinates, normal vector and tolerance.

[0016] The extracted measurement control point information is mapped to a lightweight 3D model for visualization rendering;

[0017] Based on the spatial distribution of the measurement control points, a recommended measurement sequence is generated through a path optimization algorithm, forming a machine-readable inspection plan that includes the measurement point sequence, theoretical data, and the relationship with the lightweight 3D model.

[0018] Preferably, the mobile terminal includes:

[0019] Industrial-grade mobile terminal with protective features suitable for industrial environments and workpiece recognition function;

[0020] The detection guidance client software supports 3D model interaction, real-time updates of measurement point status, and has an offline data storage and synchronization mechanism. The detection guidance client software adopts a state machine mechanism, which forces the operator to complete the measurement in the order of the detection plan. If the current measurement point is not completed, the operator cannot jump to the subsequent non-sequence measurement points.

[0021] Preferably, the mobile guidance and data acquisition module further includes an offline synchronization unit, and the mobile terminal has a built-in local database for storing detection operations and detection records when the network is interrupted;

[0022] When the network is restored, the offline synchronization unit automatically performs differential synchronization between the locally stored data and the server.

[0023] Preferably, the real-time monitoring and decision analysis module includes:

[0024] The real-time 3D visualization engine uses WebGL technology to render lightweight 3D models on the browser side and dynamically updates the visual features of the corresponding measurement control points in the model based on the deviation data received in real time from the detection records.

[0025] A process monitor is used to calculate and display one or more process indicators in real time, such as inspection progress, real-time pass rate, measurement cycle statistics, and out-of-tolerance point distribution.

[0026] The report generator is configured to automatically extract data from the database, perform statistical analysis, and generate a structured report document containing deviation chromatograms and process capability indices when the detection task triggers completion conditions.

[0027] Preferably, the system also has preset warning rules, which include:

[0028] Single-point early warning rule: When the absolute value of the measured deviation of a single measurement control point exceeds the preset proportion of its tolerance, the point is marked as an early warning in the interface of the real-time three-dimensional visualization engine.

[0029] Process early warning rule: Based on the real-time calculated process capability index estimate, when the estimate is continuously lower than the set threshold, a system-level alarm is triggered.

[0030] This invention also discloses an intelligent detection method for fixtures, comprising the following steps:

[0031] S1. Design Analysis and Task Generation: Analyze the 3D design model file of the fixture, extract the set of measurement control points defined therein and the theoretical data of each measurement point, and map the set of measurement control points to a lightweight 3D model to automatically generate a structured inspection task containing the sequence of measurement points.

[0032] S2. On-site guidance and data acquisition: The structured detection task is sent to the mobile terminal. In the three-dimensional visualization interface of the mobile terminal, the operator is forcibly guided according to the measurement point sequence. After the operator triggers the measurement command, the measured data from the measuring equipment is automatically received, and the measured data is bound in real time with the theoretical data of the currently activated measurement control point to form a detection record.

[0033] S3. Real-time monitoring and visualization: The detection records are uploaded to the server in real time, and the status indicators and visual features of the corresponding measurement control points are dynamically updated in a web-based 3D visualization view based on the deviation data in the detection records.

[0034] S4. Report Generation and Output: When the inspection task is completed or the triggering conditions are met, the system automatically performs statistical analysis based on the complete inspection records and generates an inspection report that includes deviation visualization charts and process capability indicators.

[0035] Preferably, the forced guidance in step S2 is implemented through a state machine, specifically including:

[0036] Each measurement control point is assigned a state controlled by a state machine in the 3D visualization interface. The states include inactive, pending measurement, measurement in progress, and completed.

[0037] The mobile terminal only allows the operator to interact with the measurement control point that is currently in the "to be tested" state;

[0038] Only after the measured data of the current point is successfully bound and the status changes to "completed" will the state machine automatically activate the next measurement control point in the sequence to the "to be measured" state.

[0039] Preferably, step S2 further includes an offline acquisition and synchronization sub-step:

[0040] When the mobile terminal and server network are interrupted, all operation and binding detection records are saved locally;

[0041] Once the network is restored, the locally stored data will be automatically synchronized with the server to ensure data integrity.

[0042] Preferably, the detection report generated in step S4 includes at least:

[0043] Two-dimensional deviation chromatogram generated based on 3D model rendering;

[0044] Calculation results and trend analysis of process capability index CP / CPK;

[0045] The report is automatically pushed to the preset recipients via at least one of the following methods: email, enterprise instant messaging tools, or OA system interface.

[0046] This invention provides a fixture intelligent inspection system and method with real-time three-dimensional visualization monitoring, which has the following beneficial effects:

[0047] By employing state machine-driven 3D visualization forcibly guided operation, error-proofing logic is embedded into the software workflow. The system strictly locks the measurement sequence, preventing operators from performing any non-sequence operations until the current point is completed. This completely eliminates the possibility of missed or incorrect measurements due to fatigue, negligence, or arbitrary skipping of measurements. It significantly improves the completeness and accuracy of a single inspection, providing a solid and reliable data foundation for subsequent quality decisions.

[0048] By constructing a real-time, closed-loop digital flow from design to decision-making, information silos in traditional processes are broken down. Design intent is seamlessly transformed into an executable inspection plan, and data collected on-site is transmitted back in milliseconds, dynamically driving the visualization update of the 3D model, forming a "digital twin" synchronized with the physical world. This end-to-end data connectivity not only ensures the consistency and traceability of information but also makes the quality control process completely transparent and controllable in real time.

[0049] By using a real-time 3D deviation chromatogram based on WebGL, abstract deviation values ​​are mapped to intuitive colors and spatial distributions, enabling managers to instantly perceive the overall quality distribution and areas where problems are concentrated. Combined with real-time monitoring and trend analysis of indicators such as the process capability index (CPK), the system can provide early risk warnings, driving the transformation of quality control models from lagging "post-event inspection" to proactive in-process intervention and even predictive maintenance.

[0050] The automated generation of inspection tasks eliminates the need for manual interpretation of drawings and entry of measurement points; automatic binding of on-site data eliminates errors and time-consuming manual recording and transcription; and one-click automatic report generation replaces the time-consuming manual report compilation of hours or even days. This significantly shortens the overall time to complete a comprehensive inspection, allowing professional engineers to focus their energy on creative work such as root cause analysis, process optimization, and continuous improvement, thereby enhancing the value output of human resources. Meanwhile, the robust design and intelligent offline synchronization mechanism of the dedicated industrial terminal ensure high robustness and continuous availability of the system under complex workshop environments and unstable network conditions. Attached Figure Description

[0051] To more clearly illustrate the technical solutions in this invention or the prior art, the accompanying drawings used in the description of this invention or the prior art will be briefly introduced below.

[0052] Figure 1 System architecture diagram of the present invention;

[0053] Figure 2 The system architecture data closed-loop diagram of this invention;

[0054] Figure 3 Overall flowchart of the intelligent detection method for fixtures of this invention;

[0055] Figure 4 The state machine transition diagram of the measurement point in this invention;

[0056] Figure 5 The flowchart of intelligent generation and distribution of test reports in this invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0058] Example 1, as Figures 1 to 5 As shown, a fixture intelligent inspection system with real-time 3D visualization monitoring includes:

[0059] The design data parsing and task generation module 100, deployed on an enterprise intranet server or private cloud platform, serves as the central control starting point for inspection tasks. It parses the 3D design model file of the fixture, extracts the defined set of measurement control points and the theoretical data of each measurement point, and maps the set of measurement control points to a lightweight 3D model for visualization, automatically generating a structured inspection task containing a sequence of measurement points. Specifically, it includes:

[0060] (1) Data Input and Parsing: The module provides standard interfaces (such as Web upload interfaces or integration interfaces with PDM / PLM systems) to receive 3D design model files (such as CATIA V5 .CATPart or .CATProduct files). Through CAD secondary development interfaces or parsing libraries that support intermediate formats (such as STEP AP242, which contains Product Manufacturing Information (PMI)), it reads the "Measurement Control Points" (MCPs) or "Functional Tolerances and Annotations" (FT&A) information embedded in the model.

[0061] (2) Information extraction and structuring: The parsing engine extracts the point identifier (ID / name), theoretical three-dimensional coordinates (X,Y,Z), normal vector (I,J,K) and dimensional tolerance of each measurement point from the above information to form a structured list of measurement points (such as stored in JSON or XML format).

[0062] (3) Lightweight Model Generation and Mapping: To meet the requirements of smooth rendering on Web and mobile devices, the system calls a lightweight conversion engine (such as using JT, 3D PDF, or a dedicated lightweight format conversion tool) to convert the original precise 3D design model into a lightweight model (such as .glTF or .obj format) with simplified facets but retaining macroscopic geometric features. Subsequently, the system automatically maps and associates the extracted measurement point list with the corresponding geometric positions in the lightweight model to ensure that each theoretical measurement point has an accurate virtual position on the model.

[0063] (4) Automatic generation of detection plan: The system has a built-in path optimization algorithm (such as a greedy algorithm based on the nearest neighbor principle) to calculate an approximately optimal measurement walking path based on the spatial coordinates of all measurement points, thereby generating an ordered sequence of measurement points. Finally, the system packages and generates an executable "structured detection task", which contains a lightweight model, an ordered sequence of measurement points, and all theoretical data for each point.

[0064] The mobile guidance and data acquisition module 200, using a dedicated mobile terminal 210 as its carrier, is the core of the on-site operation. Its implementation includes:

[0065] The mobile terminal 210 uses an industrial-grade ruggedized tablet PC as its hardware platform. It features a high-brightness display to adapt to workshop lighting environments, meets IP65 or higher protection standards for dust and water resistance, and integrates a 1D / 2D barcode scanner or UHF RFID reader module for rapid identification of the fixture to be inspected and automatic association with inspection tasks. The terminal establishes stable, low-latency wireless communication with the portable coordinate measuring machine 220 (such as an articulated arm measuring machine) via Bluetooth 5.0+ or ​​Wi-Fi Direct protocol.

[0066] The detection and boot client software functions include:

[0067] (1) Task loading and 3D display: After logging in, the operator downloads or directly receives the detection task pushed by the server. The client software uses the built-in 3D rendering engine (such as OpenGL ES) to load the lightweight model and displays the fixture in an interactive 3D view. All measurement points are displayed as icons floating in the corresponding positions of the model.

[0068] (2) Forced guidance driven by state machine: This is the core of error prevention. The software maintains a state machine for each measurement point, with states including: "Inactive" (gray), "Pending Measurement" (blinking, such as blue), "Measuring" (yellow), "Measured and Passed" (green), and "Measured and Out of Tolerance" (red). The interface interaction logic is strictly locked: only the current "Pending Measurement" point can respond to the operator's click confirmation; only after the current point is measured, data is bound, and the state changes will the system automatically activate the next point in the sequence as "Pending Measurement". The operator cannot manually select measurement points that are not in the current sequence, fundamentally eliminating skipped measurements and missed measurements.

[0069] (3) Intelligent Data Binding: When the operator clicks the "Measure" button, the client sends a trigger signal to the measuring device 220 via a wireless link and receives the measured coordinate values ​​returned by the device in real time. Based on the currently active measuring point ID, the software automatically binds the set of measured values ​​with the corresponding theoretical values ​​and tolerances in the task, calculates the deviation, and immediately updates the status color of the point in the 3D view. All records (including timestamps) are encrypted and stored in the local SQLite database on the terminal.

[0070] (4) Offline and Synchronization Mechanism: The client continuously monitors the network connection. When entering a Wi-Fi dead zone, it automatically switches to offline mode, and all operations and data are fully recorded locally. After the network is restored, the client automatically starts the background synchronization service, uploads the incremental detection records added locally to the server, and updates the task progress status.

[0071] The real-time monitoring and decision analysis module 300, deployed on a server or in the cloud, communicates with the mobile guidance and data acquisition module 200. It receives the detection records in real time and dynamically updates the status of each measurement control point in a web-based 3D visualization view. Simultaneously, it performs process monitoring and analysis and automatically generates a detection report based on the complete detection records. Specifically, it includes:

[0072] Real-time 3D Visualization Engine 310: This browser-based 3D rendering component utilizes a WebGL framework (such as Three.js). The engine subscribes to data streams from mobile devices. Whenever the server receives a new detection record, the engine drives the lightweight 3D model on the webpage in real time, updating the visual features (color, labels) of the corresponding measurement points. Administrators can view the overall progress in real time through the browser (gray areas decrease, green / red areas increase) and can view detailed deviation data by clicking on measurement points.

[0073] Process Monitor 320: The backend service analyzes incoming data in real time, calculates key performance indicators (KPIs), and displays them through a visual dashboard, such as: total completion percentage, current real-time pass rate, average measurement time for a single point, and spatial clustering analysis of out-of-tolerance points. These indicators provide quantitative basis for process monitoring.

[0074] Report Generator 330: Pre-configured report templates. When a testing task is completed (or manually triggered), the report engine automatically extracts all data for that task from the database, calls analysis routines to calculate process capability indices (such as CPK and PPK), and renders the final deviation chromatogram generated by the 3D visualization engine as a high-resolution 2D image. Finally, it automatically synthesizes a structured report (PDF / Word format) containing a cover, fixture information, testing conclusions, deviation chromatogram, detailed data tables, SPC charts, and trend analysis.

[0075] Early warning rule execution: The system background runs an early warning service, which makes real-time judgments based on preset rules (as described in claim 6). Once an early warning is triggered (such as a single-point deviation exceeding the tolerance by 80%), a visual alert (flashing yellow) is immediately displayed on the monitoring interface, and multi-level alarm push notifications can be configured (such as in-system messages, emails, SMS, or integration into WeChat / DingTalk) to enable real-time intervention.

[0076] Working principle:

[0077] Taking the periodic precision inspection of the "side panel welding fixture" of a certain automobile OEM as an example, the implementation process of this system is explained as follows:

[0078] S1: Task Preparation: The quality engineer uploads the CATIA digital model of the fixture to the system's web interface. The system automatically parses out 452 measurement control points, generates an inspection plan, and assigns it to the inspectors responsible for that area.

[0079] S2: On-site operations:

[0080] The inspector uses their equipped industrial tablet to log in to the inspection guidance client.

[0081] The operator selects the "Side Grip Fixture Inspection" task from the task list. The client automatically downloads the associated "Structured Inspection Task" file and "Lightweight 3D Model" file from the server.

[0082] The left side of the main interface of the testing guidance client displays the 3D model of the fixture, with all test points displayed as overlaid colored bubble icons at their respective positions on the model. Initially, all points are grayed out, indicating "to be measured". The right side of the main interface displays detailed information (ID, coordinates, tolerance, etc.) and operation buttons for the currently active test point.

[0083] Based on the task sequence, the system automatically highlights the first measurement point (e.g., MCP_001) in blue ("Measuring") and makes it the only interactive point. Operators cannot select other gray measurement points on the model.

[0084] The operator holds the articulated arm and moves the probe to the approximate position of the point on the fixture. Clicking the "Measure" button on the industrial tablet sends the command to the articulated arm via Bluetooth. The articulated arm performs a precise contact measurement and transmits the measured coordinates (X', Y', Z') back to the inspection guidance client in real time via Bluetooth. The inspection guidance client software automatically binds the received measured data to the currently active measurement point ID MCP_001, immediately calculates the deviation value, and compares it with the tolerance. Based on the deviation judgment result, the bubble icon of MCP_001 on the model instantly turns green (pass) or red (out of tolerance). Simultaneously, the status of MCP_001 changes to "completed," and the system automatically activates the next measurement point, MCP_002, in bright blue according to the sequence. This process forces the operator to complete the measurements in the order guided by the system, fundamentally eliminating skipped and missed measurements.

[0085] If the system enters a network dead zone during testing, all operation records and bound measurement data are encrypted and stored in the tablet's built-in SQLite local database. When the operator returns to a network coverage area, the client automatically starts a synchronization process in the background, uploading the local data to the server completely and systematically.

[0086] S3: Remote Monitoring: The quality supervisor / manager opens a browser in their office and logs into the system monitoring interface. They can see the color representing the inspector's entry point continuously spreading on the 3D model, with a current pass rate of 99.3%. When the 120th point is detected, the system triggers a yellow warning (deviation is approaching the upper limit), and the quality manager can immediately view the details of that point.

[0087] S4: Reporting and Closed-Loop: Within 5 minutes of the inspection completion, the system automatically generates a PDF report and sends it to the quality manager and process engineer via email. The report shows 3 points out of tolerance, and the deviation chromatogram clearly indicates that the problem is concentrated on a specific positioning mechanism of the fixture. Based on this, the process department arranges targeted maintenance to complete the quality closed loop.

[0088] This invention employs state machine-driven, 3D visual forced guidance to embed error-proofing logic into the software process. The system strictly locks the measurement sequence; operators cannot perform any non-sequence operations until the current point is completed. This technically eliminates the possibility of missed or incorrect measurements due to fatigue, negligence, or arbitrary skipping of measurements. This "forced guidance" transforms quality control from unreliable "human-based" to stable and reliable "technical-based," significantly improving the completeness and accuracy of a single inspection and providing a solid and reliable data foundation for subsequent quality decisions.

[0089] By constructing a complete, real-time closed loop from CAD design data (C) → inspection plan (P) → field data acquisition (M) → analysis and decision-making (A), design intent is seamlessly transmitted to the field, measurement data is transmitted back to the management end in milliseconds, and 3D visualization updates are driven in real time. This closed-loop data flow ensures data consistency and traceability, breaks down departmental silos and information islands, and makes the quality control process completely transparent, real-time, and controllable, achieving true digital, end-to-end quality management.

[0090] By using a real-time 3D deviation chromatogram based on WebGL, abstract deviation values ​​are mapped to intuitive colors and spatial distributions, allowing managers to "see at a glance" the overall quality status of the fixture and areas where problems are concentrated. Combined with real-time process monitoring dashboards (such as pass rate and CPK trend), the system can not only provide immediate warnings for single-point deviations but also issue risk alerts before batch problems occur by predicting process capabilities. This enables a leapfrog transformation in quality control from reactive "post-inspection" to synchronous "in-process control" and even proactive "predictive maintenance."

[0091] The automated generation of inspection tasks eliminates the need for manual interpretation of drawings and entry of measurement points; the automatic binding of on-site data eliminates errors and time-consuming manual recording and transcription; and the one-click automatic generation of reports (including SPC analysis and charts) replaces the manual report compilation that takes hours or even days. The "multiplier effect" generated by these automated processes significantly reduces the overall time to complete a comprehensive inspection, allowing professional engineers to focus their energy on creative work such as root cause analysis, process optimization, and continuous improvement, thereby enhancing the value output of human resources.

[0092] Example 2: This invention also discloses an intelligent detection method for fixtures, comprising the following steps:

[0093] S1. Design Analysis and Task Generation: This section aims to automatically translate design intent into executable detection instructions. Specifically, it includes:

[0094] S101: Input and Parsing: The quality engineer uploads the original 3D design model file (such as a CATIA .CATProduct file) of the fixture to be inspected (e.g., a welding fixture for a side panel assembly of a certain vehicle model) via the system's web interface. The system backend calls the parsing engine to read the set of "Measurement Control Points (MCPs)" embedded in the file through the CAD system's secondary development interface or intermediate format parser. This information is authoritative data defined during the design phase.

[0095] S102: Data Extraction and Structuring: From the MCP set, precisely extract the unique identifier (ID), theoretical spatial coordinates, normal vector (used to define the measurement direction), and dimensional tolerance value for each point. All information is structured into a machine-readable data list (such as JSON format).

[0096] S103: Model Lightweighting and Measurement Point Mapping: The system automatically transforms the original precision model into a lightweight 3D model (e.g., converted to .glb format) that retains the main geometric features but significantly reduces the data volume, using a lightweight conversion algorithm. Subsequently, each theoretical measurement point extracted in the previous step is precisely mapped and "anchored" to the corresponding surface of the lightweight model using a spatial coordinate matching algorithm.

[0097] S104: Detection Plan Generation: Based on the spatial distribution of all measurement points, the system runs a path optimization algorithm (e.g., a greedy algorithm based on nearest neighbors) to calculate a recommended measurement sequence that reduces the operator's travel distance. Finally, a "Structured Detection Task" file is generated, which includes: a lightweight model file, a list of measurement points with sequence numbers (containing all theoretical data), and the relationship between the two.

[0098] S2. On-site guidance and data collection:

[0099] This step, performed on-site in the workshop, is a crucial step in preventing errors and ensuring data accuracy. Specific implementation includes:

[0100] S201: Task Issuance and Loading: The generated inspection task is automatically issued to the industrial-grade mobile terminal (rugged tablet) held by the designated operator via the enterprise wireless network (Wi-Fi). After the operator logs in to the client, the task is loaded, and the interface displays the 3D model of the fixture, with all measurement points presented in icon form.

[0101] S202: Forced booting driven by state machine:

[0102] The client software maintains a state machine for each measurement point. In the initial state, the first measurement point in the sequence is set to the "to be measured" state (highlighted and flashing on the interface), and the remaining points are "inactive" (grayed out).

[0103] Interactive locking: The software interface logic is strictly programmed, and the operator can only click or select the measurement point that is currently in the "to be measured" state. After clicking, the status of the point changes to "under measurement".

[0104] The operator uses a terminal connected to the articulated arm measuring machine via Bluetooth to move the probe to the vicinity of the physical location of the point and confirms the triggering of the measurement on the terminal. The measuring machine performs precise data acquisition and transmits the measured coordinates back.

[0105] Automatic data binding and status transition: The terminal client automatically matches the received measured data with the corresponding theoretical data in the task and calculates the deviation based on the currently active measurement point ID. Once the binding is complete, the system immediately updates the status of the point to "Completed (Qualified - Green)" or "Completed (Out of Tolerance - Red)" depending on whether the deviation is within the tolerance.

[0106] Automatic sequence progression: The state machine automatically activates the next measurement point in the sequence as the "to be measured" state only after the above state transition occurs. This process repeats continuously, forming a mandatory closed-loop guidance, preventing the operator from skipping or performing measurements out of order.

[0107] S203: Offline data acquisition and synchronization:

[0108] When the terminal detects an interruption in the workshop Wi-Fi signal, it automatically and seamlessly switches to offline mode. All interactive operations, measurement data, and binding relationships are fully encrypted and recorded in the terminal's built-in SQLite local database, and the interface guidance logic remains unaffected.

[0109] When the terminal reconnects to the network, the client automatically starts a background synchronization process. By comparing the data versions on the local machine and the server, only newly added detection records are incrementally uploaded, and the task progress on the server is updated to ensure eventual data consistency.

[0110] S3, Real-time Monitoring and Visualization:

[0111] This step is performed almost simultaneously with S2, achieving transparent monitoring from the management end. Specific implementation includes:

[0112] S301: Real-time streaming data upload: Once the mobile terminal completes the data binding of a measurement point (whether online or offline and then synchronized), the detection record containing the point ID, theoretical value, measured value, deviation, and timestamp is sent to the central server.

[0113] S302: Dynamic Updates for 3D Visualization: The server-side WebGL visualization engine (e.g., based on Three.js) continuously monitors the data stream. Whenever a new record is received, the engine drives the 3D model in all connected monitoring browsers in real time, updating the visual features of the corresponding measurement points (color changes according to status, such as green / red). The entire fixture's inspection progress and quality distribution are clearly displayed on the monitoring dashboard.

[0114] S303: Real-time Calculation of Process Indicators: The server backend calculates process indicators in parallel and dynamically refreshes them on the monitoring dashboard, such as: number of completed points / total points (progress), current pass rate, and statistical charts of out-of-defect points. Managers can remotely and in real-time monitor the overall situation.

[0115] S4. Report Generation and Output:

[0116] This step is the final extraction and delivery of data value. Specific implementation includes:

[0117] S401: Triggering and Data Analysis: When the system detects that all measurement points for a task are marked as "completed" (or triggered manually by the administrator), the report generation process automatically starts. The system extracts the complete dataset for the task from the database, calls the built-in analysis module, and automatically calculates the process capability index (CPK / PPK) and generates statistical charts such as deviation distribution histograms.

[0118] S402: Report Content Composition: The report generator calls a preset template and automatically composes the following content:

[0119] (1) Two-dimensional deviation chromatogram: The final three-dimensional model, along with its color labels, is rendered to generate a high-definition two-dimensional overall deviation chromatogram, which is then embedded in the report.

[0120] (2) Structured data and charts: including detailed data tables of measurement points, historical trend comparison charts of key dimensions, process capability analysis tables, etc.

[0121] S403: Automatic Report Push: The generated complete test report (usually in PDF format) is automatically saved to the server for archiving. Simultaneously, the system automatically pushes the report through one or more of the following channels according to preset rules:

[0122] (1) Email: Send the report as an attachment to the relevant quality and process engineers and managers.

[0123] (2) Enterprise instant messaging tools: Send card messages containing a summary of conclusions and a link to an online preview of the report via WeChat / DingTalk robot.

[0124] (3) OA system interface: The report link or archived information is written into the OA workflow of the relevant personnel as a to-do item.

[0125] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A fixture intelligent inspection system with real-time three-dimensional visualization monitoring, characterized in that, include: The design data parsing and task generation module (100) is used to parse the three-dimensional design model file of the fixture, extract the set of measurement control points defined therein and the theoretical data of each measurement point, and map the set of measurement control points to a lightweight three-dimensional model for visualization display, so as to automatically generate a structured inspection task containing the sequence of measurement points. The mobile guidance and data acquisition module (200) includes a mobile terminal (210), which is used to load and execute structured detection tasks on site, force the operator to perform measurements according to the measurement point sequence through a three-dimensional visualization interface, and receive measured data from the measuring equipment (220), and bind the measured data with the corresponding theoretical data to form a detection record; The real-time monitoring and decision analysis module (300) is deployed on a server or in the cloud and is connected to the mobile guidance and data acquisition module (200) to receive the detection records in real time and dynamically update the status of each measurement control point in a web-based three-dimensional visualization view. At the same time, it performs process monitoring and analysis and automatically generates a detection report based on the complete detection records.

2. The intelligent fixture detection system with real-time three-dimensional visualization monitoring according to claim 1, characterized in that, The specific configuration of the design data parsing and task generation module (100) is as follows: Read and parse the 3D model file from the computer-aided design system, and extract the measurement control point information therein. The information includes at least the point identifier, theoretical 3D coordinates, normal vector and tolerance. The extracted measurement control point information is mapped to a lightweight 3D model for visualization rendering; Based on the spatial distribution of the measurement control points, a recommended measurement sequence is generated through a path optimization algorithm, forming a machine-readable inspection plan that includes the measurement point sequence, theoretical data, and the relationship with the lightweight 3D model.

3. The intelligent fixture detection system with real-time three-dimensional visualization monitoring according to claim 1, characterized in that, The mobile terminal (210) includes: Industrial-grade mobile terminal with protective features suitable for industrial environments and workpiece recognition function; The detection guidance client software supports 3D model interaction, real-time updates of measurement point status, and has an offline data storage and synchronization mechanism. The detection guidance client software adopts a state machine mechanism, which forces the operator to complete the measurement in the order of the detection plan. If the current measurement point is not completed, the operator cannot jump to the subsequent non-sequence measurement points.

4. The intelligent fixture detection system with real-time three-dimensional visualization monitoring according to claim 1, characterized in that, The mobile guidance and data acquisition module (200) also includes an offline synchronization unit. The mobile terminal (210) has a built-in local database for storing detection operations and detection records when the network is interrupted. When the network is restored, the offline synchronization unit automatically performs differential synchronization between the locally stored data and the server.

5. The intelligent fixture detection system with real-time three-dimensional visualization monitoring according to claim 1, characterized in that, The real-time monitoring and decision analysis module (300) includes: The real-time 3D visualization engine (310) uses WebGL technology to render a lightweight 3D model on the browser side, and dynamically updates the visual features of the corresponding measurement control points in the model based on the deviation data received in real time from the detection record. A process monitor (320) is used to calculate and display one or more process indicators in real time, such as inspection progress, real-time pass rate, measurement cycle statistics, and out-of-tolerance distribution. The report generator (330) is configured to automatically extract data from the database, perform statistical analysis, and generate a structured report document containing a deviation chromatogram and a process capability index when the detection task triggers the completion condition.

6. The intelligent fixture detection system with real-time three-dimensional visualization monitoring according to claim 1, characterized in that, The system also has preset warning rules, which include: Single-point early warning rule: When the absolute value of the measured deviation of a single measurement control point exceeds the preset proportion of its tolerance, the point is marked as an early warning in the interface of the real-time three-dimensional visualization engine (310). Process early warning rule: Based on the real-time calculated process capability index estimate, when the estimate is continuously lower than the set threshold, a system-level alarm is triggered.

7. A method for intelligent detection of fixtures, characterized in that, Includes the following steps: S1. Design Analysis and Task Generation: Analyze the 3D design model file of the fixture, extract the set of measurement control points defined therein and the theoretical data of each measurement point, and map the set of measurement control points to a lightweight 3D model to automatically generate a structured inspection task containing the sequence of measurement points. S2. On-site guidance and data acquisition: The structured detection task is sent to the mobile terminal. In the three-dimensional visualization interface of the mobile terminal, the operator is forcibly guided according to the measurement point sequence. After the operator triggers the measurement command, the measured data from the measuring equipment is automatically received, and the measured data is bound in real time with the theoretical data of the currently activated measurement control point to form a detection record. S3. Real-time monitoring and visualization: The detection records are uploaded to the server in real time, and the status indicators and visual features of the corresponding measurement control points are dynamically updated in a web-based 3D visualization view based on the deviation data in the detection records. S4. Report Generation and Output: When the inspection task is completed or the triggering conditions are met, the system automatically performs statistical analysis based on the complete inspection records and generates an inspection report that includes deviation visualization charts and process capability indicators.

8. The intelligent fixture detection method according to claim 7, characterized in that, The forced guidance in step S2 is implemented through a state machine, specifically including: Each measurement control point is assigned a state controlled by a state machine in the 3D visualization interface. The states include inactive, pending measurement, measurement in progress, and completed. The mobile terminal only allows the operator to interact with the measurement control point that is currently in the "to be tested" state; Only after the measured data of the current point is successfully bound and the status changes to "completed" will the state machine automatically activate the next measurement control point in the sequence to the "to be measured" state.

9. The intelligent fixture detection method according to claim 7, characterized in that, Step S2 also includes an offline acquisition and synchronization sub-step: When the mobile terminal and server network are interrupted, all operation and binding detection records are saved locally; Once the network is restored, the locally stored data will be automatically synchronized with the server to ensure data integrity.

10. The intelligent fixture detection method according to claim 7, characterized in that, The detection report generated in step S4 includes at least the following: Two-dimensional deviation chromatogram generated based on 3D model rendering; Calculation results and trend analysis of process capability index CP / CPK; The report is automatically pushed to the preset recipients via at least one of the following methods: email, enterprise instant messaging tools, or OA system interface.

Images