Intelligent inspection system and method

Abstract

The invention discloses an intelligent inspection system and method, and relates to the technical field of industrial automatic inspection, a data acquisition module is integrated on mobile equipment to realize synchronous acquisition of instrument images and field environment parameters, and an environment adaptive calibration module is combined to perform targeted preprocessing optimization on the instrument images according to the environment parameters, so that the accuracy of the instrument images is improved. Structured reading data is generated through visual identification analysis of the data processing module, and finally bidirectional verification and hierarchical feedback of field data and remote transmission data are completed through the data comparison and management module, so that the inspection process gets rid of high dependence on manual reading. The influence of complex environments such as uneven illumination and light reflection on data acquisition accuracy is reduced, meanwhile, the isolation state of field inspection data and plant-level control system data is broken, an abnormal information transmission and response chain is shortened, and the effects of improving the inspection data accuracy, enhancing the system data collaboration and optimizing the inspection management efficiency are achieved.

Description

Technical Field

[0001] This application relates to the field of industrial automation inspection technology, and in particular to an intelligent inspection system and method. Background Technology

[0002] In chemical production processes, regular inspections of process parameters such as pressure, temperature, and liquid level in equipment are crucial for ensuring safe and stable operation. Currently, many chemical plants still rely heavily on pointer-type instruments with only local indication functions, such as pressure gauges, thermometers, and level gauges. The mainstream method for inspecting these instruments is for operators to periodically visit the site to manually read and record the data. Based on this, electronic inspection systems based on barcodes / Radio Frequency Identification (RFID) have been developed, but these systems can only record the inspection route.

[0003] However, existing technologies have significant drawbacks: First, manual interpretation is susceptible to factors such as personnel fatigue, differences in perspective, and ambient lighting, leading to reading errors, missed detections, and incorrect recordings, making it difficult to guarantee the accuracy of inspection data. Second, manually recorded data requires secondary entry later, easily creating data silos and making it impossible to automatically compare and analyze with real-time databases such as plant-level distributed control systems (DCS) and supervisory control and data acquisition and monitoring systems (SCADA), resulting in poor data collaboration. Third, the on-site abnormal information transmission chain is long, and coupled with the high noise and complex environment of the plant, the communication efficiency between the on-site and central control is low, easily causing delays in abnormal response. Fourth, the inspection results are presented in tabular numerical form, which cannot be intuitively linked with equipment status and location information, hindering trend analysis and early warning management.

[0004] In summary, existing inspection technologies can no longer meet the needs of efficient and precise control in chemical production, and a new inspection solution is urgently needed to solve the above problems. Summary of the Invention

[0005] To address the above problems, this application provides an intelligent inspection system and method, including the following:

[0006] Firstly, this application provides an intelligent inspection system, which includes:

[0007] Data acquisition module, data processing module, data comparison and management module, and environmental adaptive calibration module;

[0008] The data acquisition module is integrated into the mobile device and is used to acquire pointer instrument images and field environmental parameters;

[0009] The environmental adaptive calibration module is used to preprocess and optimize the acquired instrument images based on the field environmental parameters;

[0010] The data processing module is used to perform visual recognition and analysis on the optimized instrument images and output structured reading data;

[0011] The data comparison and management module is used to receive structured reading data, retrieve remote instrument data from corresponding monitoring points and perform bidirectional verification and comparison, and output hierarchical comparison results and trend analysis reports.

[0012] Optionally, the data acquisition module may include a built-in camera, multiple sensors, and a function configuration unit.

[0013] The built-in camera is used to capture images from the pointer-type instrument; multiple sensors are used to collect on-site environmental parameters.

[0014] The functional configuration unit is equipped with image assistance function and environmental parameter acquisition function. The image assistance function includes assistive focusing, image quality detection and light compensation adjustment. The image quality detection is used to determine the sharpness, whether there is shaking and the degree of reflection of the acquired image. The environmental parameter acquisition function obtains data on light intensity, ambient humidity and vibration frequency through multiple sensors.

[0015] Optionally, the environment adaptive calibration module specifically includes:

[0016] The image preprocessing unit and parameter matching unit are included. The parameter matching unit pre-stores image optimization algorithms corresponding to different combinations of environmental parameters.

[0017] The image preprocessing unit is used to call the corresponding optimization algorithm according to the on-site environmental parameters to perform anti-glare and noise reduction, brightness equalization and jitter correction on the instrument image.

[0018] Optionally, the data processing module uses an adaptive target detection algorithm to perform image analysis. The image analysis includes instrument model recognition, dial detection and positioning, accurate pointer recognition and angle calculation, and reading conversion based on range scale rules. The adaptive target detection algorithm is used to automatically match the corresponding recognition model parameters according to the instrument model.

[0019] Optionally, the data processing module may also include:

[0020] Data temporary storage unit, secondary confirmation unit, and anomaly marking unit;

[0021] The data storage unit is used to store the optimized image, recognition results, original environmental parameters and recognition confidence level. The secondary confirmation unit is used to display the relevant data of all inspection points in the form of image and text association for manual verification. The anomaly marking unit allows manual marking of questionable data and additional explanations.

[0022] Optionally, the data comparison and management module retrieves remote instrument data from the distributed control system or data acquisition and monitoring system through the OPC UA standard interface. The bidirectional verification and comparison includes the calculation of the difference between the field structured reading data and the remote instrument data, the comparison of historical data from the same period, and the trend fitting analysis.

[0023] Optionally, the data comparison and management module specifically includes:

[0024] The system includes a threshold dynamic configuration unit, a graded early warning unit, a trend analysis unit, and a linkage response unit. The threshold dynamic configuration unit is used to adaptively adjust the allowable deviation threshold of the process according to the importance level of the process parameters and the equipment operating cycle. The graded early warning unit is used to divide the system into multiple levels of early warning based on the difference results and the rate of trend change. The trend analysis unit is used to generate equipment operating trend curves and fault prediction reports based on historical inspection data and real-time comparison results. The linkage response unit is used to automatically push abnormal information, on-site images, and handling suggestions to the central control system when an early warning is triggered.

[0025] Optionally, the adaptive target detection algorithm includes a dynamic switching mechanism for the YOLO algorithm, SSD algorithm, and Faster R-CNN algorithm. The dynamic switching mechanism automatically selects the optimal algorithm based on the complexity of the instrument model, image clarity, and recognition confidence threshold.

[0026] Optionally, the system may also include:

[0027] The offline working module and data synchronization unit support image acquisition, environmental parameter recording, local preprocessing and recognition analysis in the absence of network environment, and temporarily store the data on mobile device. The data synchronization unit is used to automatically upload offline data to the server after the network connection is restored, and perform data comparison and update with the data comparison and management module.

[0028] Secondly, this application provides an intelligent inspection method, which includes:

[0029] The data acquisition module integrated into the mobile device acquires pointer-type instrument images and field environmental parameters.

[0030] The environmental adaptive calibration module is used to preprocess and optimize the acquired instrument images based on the on-site environmental parameters;

[0031] The data processing module performs visual recognition and analysis on the optimized instrument images and outputs structured reading data.

[0032] The system receives structured reading data using the data comparison and management module, retrieves remote instrument data from the corresponding monitoring points, performs bidirectional verification and comparison, and outputs hierarchical comparison results and trend analysis reports.

[0033] Optional, the acquisition of pointer-type instrument images and field environmental parameters includes:

[0034] The data acquisition module uses a built-in camera to capture images from pointer-type instruments and multiple sensors to collect environmental parameters. The image assistance function of the function configuration unit assists the acquisition process, including focusing assistance, image quality detection, and light compensation adjustment. The image quality detection determines the clarity, jitter, and reflectivity of the acquired image. The environmental parameter acquisition function of the function configuration unit uses the multiple sensors to acquire data on light intensity, ambient humidity, and vibration frequency.

[0035] Optionally, preprocessing and optimizing the acquired instrument images includes: calling the image optimization algorithm corresponding to the field environment parameters through the parameter matching unit of the environment adaptive calibration module, wherein the parameter matching unit pre-stores the image optimization algorithms corresponding to different combinations of environmental parameters; and using the image optimization algorithm through the image preprocessing unit to perform anti-glare and noise reduction, brightness equalization and jitter correction processing on the instrument images.

[0036] Optionally, visual recognition analysis of the optimized instrument image includes: using an adaptive target detection algorithm to perform image analysis on the optimized instrument image, including instrument model recognition, dial detection and positioning, accurate pointer recognition and angle calculation, and reading conversion based on range scale rules; automatically matching the corresponding recognition model parameters according to the identified instrument model, and outputting structured reading data.

[0037] Optionally, the method also includes data storage, secondary confirmation, and anomaly marking steps. Specifically, the data storage unit stores the optimized image, recognition results, original environmental parameters, and recognition confidence level; the secondary confirmation unit displays relevant data of all inspection points in a graphic-text association format for manual verification; and the anomaly marking unit allows manual marking of questionable data with additional explanations.

[0038] Optionally, the remote instrument data of the corresponding monitoring point and the two-way verification and comparison specifically include: retrieving remote instrument data from the distributed control system or data acquisition and monitoring system through the OPC UA standard interface; performing two-way verification and comparison between the field structured reading data and the remote instrument data, including difference calculation, comparison with historical data of the same period and trend fitting analysis.

[0039] Optionally, the output of hierarchical comparison results and trend analysis reports specifically includes: adaptively adjusting the process allowable deviation threshold according to the importance level of process parameters and equipment operating cycle; dividing the warning into multiple levels according to the difference results and trend change rate; generating equipment operating trend curves and fault prediction reports based on historical inspection data and real-time comparison results; and automatically pushing abnormal information, on-site images and handling suggestions to the central control system when a warning is triggered.

[0040] Optionally, the adaptive target detection algorithm includes a dynamic switching mechanism for YOLO, SSD and Faster R-CNN algorithms, which automatically selects the optimal algorithm for image analysis based on the complexity of the instrument model, image clarity and recognition confidence threshold.

[0041] Optionally, the method also includes offline work and data synchronization. Specifically, in a network-free environment, image acquisition, environmental parameter recording, local preprocessing and recognition analysis are completed, and the data is temporarily stored on the mobile device. After the network connection is restored, the offline data is automatically uploaded to the server, and the data comparison and management module is used to complete the comparison and update the data.

[0042] Thirdly, this application provides an apparatus including a memory and a processor, the memory for storing instructions or code, and the processor for executing the instructions or code to cause the apparatus to perform the method described in any of the implementations of the second aspect above.

[0043] Fourthly, this application provides a computer-readable storage medium storing code, wherein when the code is run, the device running the code implements the method described in any of the implementations of the second aspect above.

[0044] This application provides an intelligent inspection system, which includes a data acquisition module, a data processing module, a data comparison and management module, and an environmental adaptive calibration module. The data acquisition module is integrated into a mobile device and is used to acquire pointer-type instrument images and on-site environmental parameters. The environmental adaptive calibration module is used to preprocess and optimize the acquired instrument images according to the on-site environmental parameters. The data processing module is used to perform visual recognition analysis on the optimized instrument images and output structured reading data. The data comparison and management module is used to receive the structured reading data, retrieve remote instrument data from corresponding monitoring points and perform bidirectional verification and comparison, and output hierarchical comparison results and trend analysis reports. In this way, by integrating the data acquisition module into mobile devices, instrument images and on-site environmental parameters are simultaneously acquired. Combined with an environmental adaptive calibration module, the instrument images undergo targeted preprocessing and optimization based on environmental parameters. Then, the data processing module generates structured reading data through visual recognition and analysis. Finally, the data comparison and management module completes two-way verification and hierarchical feedback between on-site and remote data. This reduces the heavy reliance on manual readings during inspections, minimizes the impact of uneven lighting, glare, and other complex environmental factors on data acquisition accuracy, and breaks down the isolation between on-site inspection data and plant-level control system data. It also shortens the abnormal information transmission and response chain, achieving the effects of improving inspection data accuracy, enhancing system data synergy, and optimizing inspection management efficiency. This reduces oversights and errors caused by manual operation, provides more comprehensive data support for equipment operating status assessment, contributes to the refinement and intelligence of production management, and ensures the safe and stable operation of chemical production processes. Attached Figure Description

[0045] To more clearly illustrate the technical solutions in this embodiment or the prior art, the drawings used in the description of the embodiment or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0046] Figure 1 This is a schematic diagram of the structure of an intelligent inspection system provided in an embodiment of this application;

[0047] Figure 2 A flowchart of an intelligent inspection method provided in an embodiment of this application. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0049] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0050] Figure 1 This is a structural schematic diagram of an intelligent inspection device provided in an embodiment of this application. (Combined with...) Figure 1 As shown, the intelligent inspection device 100 provided in this application embodiment includes:

[0051] The system comprises a data acquisition module 110, a data processing module 120, a data comparison and management module 130, and an environmental adaptive calibration module 140. These modules work together to form a complete intelligent inspection chain. The data acquisition module 110, integrated into a mobile device, serves as the core data input for the system, collecting pointer-type instrument images and on-site environmental parameters to provide basic data support for subsequent processing. The environmental adaptive calibration module 140 performs targeted preprocessing and optimization of the instrument images based on the collected on-site environmental parameters, eliminating interference from complex environments on image quality. The data processing module 120 receives the optimized instrument images, extracts key information through visual recognition and analysis technology, and outputs standardized structured reading data. The data comparison and management module 130 receives this structured reading data, simultaneously retrieves remote instrument data from corresponding monitoring points for bidirectional verification and comparison, and finally outputs graded comparison results and trend analysis reports, providing decision-making basis for production management.

[0052] To more clearly demonstrate the specific implementation logic and technical details of each module, the following sections will provide a detailed introduction to the data acquisition module 110, data processing module 120, data comparison and management module 130, and environmental adaptive calibration module 140.

[0053] In one implementation of this application, the data acquisition module, as the core unit for system data input, specifically includes a built-in camera, multiple sensors, and a function configuration unit. The built-in camera uses a high-definition imaging component, capable of clearly capturing key details such as the dial scale and pointer shape of pointer-type instruments, and is specifically used for acquiring images of pointer-type instruments. The multiple sensors integrate a light sensor, a humidity sensor, and a vibration sensor, enabling simultaneous acquisition of on-site environmental parameters. The function configuration unit provides comprehensive support for the acquisition process, equipped with image assistance and environmental parameter acquisition functions. The image assistance function covers assisted focusing, image quality detection, and light compensation adjustment. Assisted focusing can automatically lock the instrument dial area through an algorithm. Light compensation adjustment can dynamically adjust shooting parameters according to the on-site light intensity. Image quality detection can accurately determine the clarity, jitter, and reflectivity of the acquired image. When an image is detected as not meeting preset standards, a prompt to retake the image will be issued to the inspection personnel to ensure that the image meets subsequent processing requirements. The environmental parameter acquisition function accurately acquires light intensity, ambient humidity, and vibration frequency data through multiple sensors, providing reliable environmental reference data for image preprocessing.

[0054] In one implementation of this application, the environment adaptive calibration module connected to the data acquisition module specifically includes an image preprocessing unit and a parameter matching unit. The parameter matching unit pre-stores a large number of image optimization algorithms corresponding to different combinations of environmental parameters that have been verified in actual scenarios, such as polarization filtering algorithms for strong light reflection scenarios, brightness enhancement algorithms for weak light environments, and image stabilization algorithms for vibration scenarios, which can achieve accurate matching between environmental parameters and optimization algorithms. After receiving the on-site environmental parameters transmitted by the data acquisition module, the image preprocessing unit quickly retrieves the appropriate optimization algorithm through the parameter matching unit, and then performs anti-reflection noise reduction, brightness equalization, and jitter correction processing on the acquired instrument image. Anti-reflection noise reduction can eliminate image noise caused by instrument glass reflection and on-site electromagnetic interference. Brightness equalization can smooth out the brightness differences in different areas of the dial. Jitter correction compensates for slight jitter during shooting through image alignment technology, effectively eliminating the interference of complex on-site environment on image quality, and laying a high-quality data foundation for subsequent visual recognition analysis.

[0055] In one implementation of this application, the data processing module that receives the optimized image output by the environmental adaptive calibration module uses an adaptive target detection algorithm to perform image analysis. This image analysis process sequentially includes instrument model identification, dial detection and positioning, precise pointer identification and angle calculation, and reading conversion based on range scale rules. Instrument model identification extracts features such as the dial's outline, scale distribution, and markings, and compares them with the system's built-in instrument feature library to determine the specific instrument model. Dial detection and positioning uses a target bounding box algorithm to accurately lock onto the effective instrument dial in the image. The system eliminates background clutter and ensures accurate pointer recognition by distinguishing the color and grayscale differences between the pointer and the dial scale. It then calculates the pointer's deflection angle relative to zero by combining the dial's center coordinates. The reading conversion based on the range and scale rules matches the corresponding range and scale interval to the identified instrument model, converting the calculated pointer angle into specific physical quantities such as MPa and °C. Furthermore, the adaptive target detection algorithm automatically matches the corresponding recognition model parameters based on the identified instrument model, ensuring accurate recognition of pointer instruments of different models and specifications.

[0056] In one implementation of this application, to further ensure data accuracy, the data processing module further includes a data temporary storage unit, a secondary confirmation unit, and an anomaly marking unit. The data temporary storage unit uses local encrypted storage to completely store the optimized image, recognition result, original environmental parameters, and recognition confidence level. The recognition confidence level is presented as a percentage, intuitively reflecting the reliability of the recognition result and achieving complete retention of key data. The secondary confirmation unit centrally displays the relevant data of all inspection points in a graphic and text-based format. That is, the original acquisition image of the instrument, the optimized image, the recognition reading, the confidence level, and the corresponding inspection point information are presented simultaneously on the same interface, which facilitates inspection personnel to conduct visual verification one by one and quickly identify recognition deviations. The anomaly marking unit provides a convenient marking entry, allowing inspection personnel to manually mark questionable data with labels such as "to be reviewed" and "abnormal reading," and add text descriptions, forming a dual guarantee mechanism of "automatic recognition + manual review," further reducing the risk of data error.

[0057] In one implementation of this application, after receiving the structured reading data output by the data processing module, the data comparison and management module establishes stable communication with the distributed control system or data acquisition and monitoring system through the Open Platform Communications Unified Architecture (OPC UA) standard interface. According to a preset point mapping relationship, it accurately retrieves the remote instrument data of the corresponding monitoring point and then performs bidirectional verification and comparison. This bidirectional verification and comparison specifically includes: real-time difference calculation between the on-site structured reading data and the remote instrument data, intuitively reflecting the magnitude of the deviation between the two sets of data; historical data comparison, comparing the current inspection data with historical data from the same period (7 days, 30 days, etc.) to identify abnormal data trends; and trend fitting analysis, using algorithms such as linear regression and moving average to fit multiple consecutive sets of inspection data to determine whether parameter changes conform to normal process rules, achieving multi-dimensional cross-validation of the data.

[0058] In one implementation of this application, to enhance the intelligence and practicality of data management, the data comparison and management module specifically includes a threshold dynamic configuration unit, a hierarchical early warning unit, a trend analysis unit, and a linkage response unit. The threshold dynamic configuration unit can adaptively adjust the allowable deviation threshold of the process according to the importance level of the process parameters (such as critical process parameters, general process parameters) and the equipment operating cycle (such as break-in period, stable operation period, aging period). For example, it can set a stricter deviation threshold for critical process parameters and appropriately relax the threshold for aging equipment, making the threshold setting more closely match actual production needs. The hierarchical early warning unit divides abnormal situations into four levels of early warning: normal, attention, warning, and emergency, based on the difference results and trend change rate of the two-way verification comparison. Different color codes and response priorities are used to achieve hierarchical control of abnormal situations; the trend analysis unit automatically generates equipment operation trend curves based on historical inspection data and real-time comparison results, intuitively showing the changes in parameters over time. At the same time, combined with process thresholds and equipment fault models, it generates fault prediction reports, clearly marking potential fault risk points and the possible time range, providing forward-looking guidance for equipment maintenance; when a warning or higher level alert is triggered, the linkage response unit automatically pushes abnormal information, on-site images, and handling suggestions to the central control system via industrial Ethernet. The handling suggestions are generated based on historical fault handling cases and process specifications. At the same time, an electronic work order is generated and distributed to the mobile terminals of the corresponding maintenance personnel, shortening the abnormal response cycle.

[0059] In one implementation of this application, to adapt to the recognition needs of different scenarios, the adaptive target detection algorithm includes a dynamic switching mechanism between a You Only Look Once (YOLO) algorithm, a Single Shot MultiBox Detector (SSD) algorithm, and a Faster Regions with Convolutional Neural Network Features (Faster R-CNN) algorithm. This dynamic switching mechanism incorporates an algorithm performance evaluation model that automatically selects the optimal algorithm based on the complexity of the instrument model (e.g., single-scale instruments, multi-scale instruments, irregularly shaped dial instruments), image clarity (e.g., high-definition with no noise, blurry with interference), and the recognition confidence threshold. For instruments with simple structures and clear images, the YOLO algorithm is preferred, balancing recognition speed and accuracy. For instruments with complex structures and rich details, the Faster R-CNN algorithm is selected to improve recognition accuracy. For scenarios with poor image quality and a lot of interference, the SSD algorithm is selected to balance anti-interference capability and recognition efficiency, ensuring both recognition accuracy and efficiency.

[0060] In one implementation of this application, considering the potential network instability or lack of network access in some areas of the chemical plant, the system further includes an offline working module and a data synchronization unit. The offline working module pre-stores inspection routes, location information, instrument feature libraries, and basic algorithm models locally on the mobile device, supporting normal image acquisition, environmental parameter recording, local preprocessing, and identification analysis even in network-free environments. It also temporarily stores relevant data in encrypted format on the mobile device's local storage space to ensure uninterrupted inspection work. The data synchronization unit has a real-time network status monitoring function. After detecting that the mobile device has restored its network connection, it automatically initiates the data synchronization process, uploading all data stored offline to the server in chronological order. This data is then compared and updated with the data comparison and management module, while simultaneously verifying data integrity. If data is missing, it automatically triggers retransmission to ensure data integrity and consistency.

[0061] The above embodiments describe the specific functions of each module in the intelligent inspection system of this application. The following detailed description uses a specific embodiment to illustrate the intelligent inspection system of this application, ensuring precise alignment between the functional implementation of each module and the actual inspection process:

[0062] The implementation scenario is set in the pressure vessel unit area of ​​a chemical plant. Due to the needs of production safety management, the inspection personnel need to conduct regular inspections of the local pressure gauges at 10 key points every 2 hours. The core objective is to accurately obtain the pressure parameters of each pressure gauge, verify the data with the remote transmission system, and promptly identify equipment malfunctions.

[0063] The implementation steps are as follows:

[0064] Step 1: Inspection personnel arrive at inspection point 1 carrying an explosion-proof smartphone with an integrated data acquisition module. After opening the system's dedicated app, they scan the QR code (system identification unit) on the device. The app immediately retrieves the standard pressure gauge legend, range information (0-1.6MPa), and historical inspection data corresponding to the inspection point, providing a reference benchmark for subsequent data collection and identification. At the same time, the multi-sensor of the data acquisition module automatically starts, collecting environmental parameters such as ambient light intensity, ambient humidity, and vibration frequency in real time, and transmitting them synchronously to the environmental adaptive calibration module for backup.

[0065] Step Two: Inspection personnel use the built-in camera of their explosion-proof smartphones to photograph the pressure gauge. Simultaneously, the data acquisition module's configuration unit activates the image assistance function, the auxiliary focusing algorithm automatically locks onto the dial area, and the light compensation adjustment dynamically adjusts the shooting parameters according to the ambient light intensity. The image quality detection function judges the clarity, jitter, and reflectivity of the acquired image in real time. After the image passes inspection, the App triggers a local lightweight adaptive target detection algorithm (defaulting to the YOLO algorithm). First, the parameter matching unit of the environmental adaptive calibration module uses the environmental parameters acquired in Step One to preprocess and optimize the image using the corresponding anti-reflection and noise reduction algorithm. Then, the data processing module analyzes the optimized image, sequentially completing instrument model identification (matching the built-in instrument feature library), dial detection and positioning (selecting the effective area), and precise pointer identification and angle calculation (locating the pointer tip and calculating the deflection angle). Finally, based on the 0-1.6MPa range information and scale rules, the pointer angle is converted into a specific reading of 1.02MPa. The structured readings, original acquired images, optimized images, and recognition confidence levels (e.g., 98%) are all encrypted and stored locally by the data processing module's temporary data storage unit.

[0066] Step 3: Following the steps above, the inspection personnel will sequentially complete the image acquisition, environmental parameter recording, image preprocessing, and recognition storage for the remaining 9 inspection points. After all points have been inspected, the secondary confirmation unit of the data processing module will generate an inspection list containing all points in the App interface in a text-image association format. This list will simultaneously display the inspection location, standard legend, original image, optimized image, recognition reading, and confidence level for each point. The inspection personnel will then visually verify each point. If any reading at a point is found to be questionable, it can be marked with a "Pending Verification" label and an explanation using the anomaly marking unit. After all verifications are completed, the personnel will click the "Submit" button to initiate a data upload request.

[0067] Step 4: If the on-site network is available, the data processing module uploads the structured readings, raw environmental parameters, image files, and manual verification results of all points to the server via wireless network. If there is no network on-site, the offline working module temporarily stores the above data on the explosion-proof smartphone. Once the network is restored, the data synchronization unit will automatically upload and complete the data. The server-side data comparison and management module retrieves the remote transmission data of the pressure transmitters corresponding to these 10 inspection points from the factory's distributed control system (DCS) via the OPC UA standard interface (e.g., the remote transmission value of point 1 is 1.05MPa). Then, it initiates a two-way verification and comparison process, which not only calculates the real-time difference between the on-site structured readings and the remote transmission data (difference of point 1 |1.02-1.05|=0.03MPa), but also compares the trend with historical data from the same period, and uses fitting analysis to determine whether the parameter changes conform to the process rules.

[0068] Step 5: The threshold dynamic configuration unit of the data comparison and management module presets the allowable deviation threshold for this process point to 0.05MPa (based on the importance level of this parameter and the stable operation cycle of the equipment). The graded early warning unit determines the status of this point as "normal" based on the difference of 0.03MPa at point 1 (less than the allowable deviation threshold) and displays it in green on the management interface. If the difference at a point exceeds 0.05MPa but does not reach 0.1MPa, a "Caution" level yellow warning is triggered. If the difference exceeds 0.1MPa, a "Warning" level red alarm is triggered. At the same time, the trend analysis unit generates the operating trend curve of the pressure parameters of each point based on the current and historical inspection data, and generates a prediction report in combination with the fault model. When the linkage response unit triggers a "Warning" level or higher alarm, it automatically pushes abnormal point information, on-site collected images, optimized images and standardized handling suggestions to the central control system, and generates an electronic work order to be distributed to the mobile terminal of the corresponding maintenance personnel to ensure timely handling of abnormal situations.

[0069] The above is a schematic diagram of the structure of an intelligent inspection device provided in an embodiment of this application. Based on this, this application also provides a corresponding method. The intelligent inspection method provided in the embodiment of this application will be described below with reference to specific flowcharts.

[0070] Figure 2 This is a flowchart illustrating an intelligent inspection method provided in an embodiment of this application. (In conjunction with...) Figure 1 As shown, the intelligent inspection method provided in this application embodiment may include:

[0071] S201. Data acquisition modules integrated into mobile devices are used to acquire pointer-type instrument images and field environmental parameters.

[0072] S202. The environmental adaptive calibration module is used to preprocess and optimize the acquired instrument images based on the field environmental parameters.

[0073] S203. The optimized instrument image is visually recognized and analyzed by the data processing module, and structured reading data is output.

[0074] S204. The data comparison and management module receives the structured reading data, retrieves the remote instrument data of the corresponding monitoring point and performs bidirectional verification and comparison, and outputs the hierarchical comparison results and trend analysis report.

[0075] In one implementation of this application, acquiring pointer-type instrument images and on-site environmental parameters includes:

[0076] The data acquisition module uses a built-in camera to capture images from pointer-type instruments and multiple sensors to collect environmental parameters. The image assistance function of the configuration unit assists the acquisition process, including focus assistance, image quality detection, and light compensation adjustment. The image quality detection determines the clarity, jitter, and reflectivity of the acquired image. The environmental parameter acquisition function of the configuration unit uses the multiple sensors to obtain data on light intensity, ambient humidity, and vibration frequency.

[0077] In one implementation of this application, preprocessing and optimizing the acquired instrument image includes: calling an image optimization algorithm corresponding to the field environment parameters through the parameter matching unit of the environment adaptive calibration module, wherein the parameter matching unit pre-stores image optimization algorithms corresponding to different combinations of environment parameters; and performing anti-glare and noise reduction, brightness equalization, and jitter correction processing on the instrument image by the image preprocessing unit using the image optimization algorithm.

[0078] In one implementation of this application, visual recognition analysis of the optimized instrument image includes: performing image analysis on the optimized instrument image using an adaptive target detection algorithm, wherein the image analysis includes instrument model recognition, dial detection and positioning, accurate pointer recognition and angle calculation, and reading conversion based on range scale rules; automatically matching the corresponding recognition model parameters according to the identified instrument model, and outputting structured reading data.

[0079] In one implementation of this application, the method further includes data temporary storage, secondary confirmation, and anomaly marking steps. Specifically, the data temporary storage unit stores the optimized image, recognition result, original environmental parameters, and recognition confidence level; the secondary confirmation unit displays relevant data of all inspection points in a graphic-text association format for manual verification; and the anomaly marking unit allows manual marking of questionable data with additional explanations.

[0080] In one implementation of this application, the bidirectional verification and comparison of remote instrument data corresponding to monitoring points specifically includes: retrieving remote instrument data from a distributed control system or data acquisition and monitoring system through the OPC UA standard interface; performing bidirectional verification and comparison between on-site structured reading data and remote instrument data, wherein the bidirectional verification and comparison includes difference calculation, comparison of historical data from the same period, and trend fitting analysis.

[0081] In one implementation of this application, the output of the hierarchical comparison results and trend analysis report specifically includes: adaptively adjusting the process allowable deviation threshold according to the importance level of process parameters and equipment operating cycle; dividing the warning into multiple layers according to the difference results and trend change rate; generating equipment operating trend curves and fault prediction reports based on historical inspection data and real-time comparison results; and automatically pushing abnormal information, on-site images and handling suggestions to the central control system when a warning is triggered.

[0082] In one implementation of this application, the adaptive target detection algorithm includes a dynamic switching mechanism between the YOLO algorithm, the SSD algorithm, and the Faster R-CNN algorithm, which automatically selects the optimal algorithm for image analysis based on the complexity of the instrument model, image clarity, and recognition confidence threshold.

[0083] In one implementation of this application, the method further includes offline work and data synchronization. Specifically, in a network-free environment, image acquisition, environmental parameter recording, local preprocessing and recognition analysis are completed, and the data is temporarily stored on the mobile device. After the network connection is restored, the offline data is automatically uploaded to the server, and the data is compared and updated by the data comparison and management module.

[0084] This application also provides corresponding devices and computer storage media for implementing the solutions provided in this application.

[0085] The device includes a memory and a processor. The memory stores instructions or code, and the processor executes the instructions or code to cause the device to perform the method described in any embodiment of this application.

[0086] The computer storage medium stores code, and when the code is run, the device running the code implements the method described in any embodiment of this application.

[0087] As can be seen from the above description of the embodiments, those skilled in the art can clearly understand that all or part of the steps in the methods of the above embodiments can be implemented by means of software plus a general-purpose hardware platform. Based on this understanding, the technical solution of this application can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as a read-only memory (ROM) / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network communication device such as a router) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0088] It is understood that in the specific embodiments of this application, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved need to obtain user permission or consent when the above embodiments of this application are applied to specific products or technologies, and the collection, use and processing of related data need to comply with the relevant laws, regulations and standards of relevant countries and regions.

[0089] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0090] It should also be noted that the various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, for the device and apparatus embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the description of the method embodiments. The device and apparatus embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components indicated as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the solution in this embodiment according to actual needs. Those skilled in the art can understand and implement this without creative effort.

[0091] The above description is merely one specific embodiment 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 technical scope 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. An intelligent inspection system, characterized in that, The system comprises: a data acquisition module, a data processing module, a data comparison and management module, and an environment adaptive calibration module; the data acquisition module is integrated in a mobile device and is used for acquiring images of a pointer instrument and field environment parameters; the environment adaptive calibration module is used for pre-processing and optimizing the acquired instrument images according to the field environment parameters; the data processing module is used for visually recognizing and analyzing the optimized instrument images and outputting structured reading data; the data comparison and management module is used for receiving the structured reading data, calling remote instrument data of a corresponding monitoring point and performing bidirectional verification comparison, and outputting a hierarchical comparison result and a trend analysis report.

2. The system of claim 1, wherein, The data acquisition module specifically comprises a built-in camera, multiple sensors, and a functional configuration unit; the built-in camera is used for acquiring the images of the pointer instrument; the multiple sensors are used for acquiring the field environment parameters; the functional configuration unit is configured with image auxiliary functions and environment parameter acquisition functions, the image auxiliary functions include auxiliary focusing, image quality detection, and light compensation adjustment, the image quality detection is used for judging the definition, whether shaking, and the degree of reflection of the acquired images, and the environment parameter acquisition functions acquire light intensity, environmental humidity, and vibration frequency data through the multiple sensors.

3. The system of claim 1, wherein, The environment adaptive calibration module specifically comprises: an image preprocessing unit and a parameter matching unit, the parameter matching unit pre-stores image optimization algorithms corresponding to different combinations of environment parameters; the image preprocessing unit is used for calling corresponding optimization algorithms according to the field environment parameters, and performing anti-reflection noise reduction, light and dark balance, and shaking correction processing on the instrument images.

4. The system of claim 1, wherein, The data processing module adopts an adaptive target detection algorithm to realize image analysis, the image analysis includes instrument model identification, dial detection and positioning, pointer accurate identification and angle calculation, and reading conversion based on range scale rules, and the adaptive target detection algorithm is used for automatically matching corresponding identification model parameters according to the instrument model.

5. The system of claim 1, wherein, The data processing module further comprises: a data temporary storage unit, a secondary confirmation unit, and an abnormality marking unit; the data temporary storage unit is used for storing the optimized images, identification results, original environment parameters, and identification confidence; the secondary confirmation unit is used for displaying related data of all inspection points in a graphic-textual association form for manual verification, and the abnormality marking unit allows manual annotation of suspicious data and adds an explanation.

6. The system of claim 1, wherein, The data comparison and management module calls remote instrument data in a distributed control system or a data acquisition and monitoring system through an OPC UA standard interface, and the bidirectional verification comparison includes difference calculation of field structured reading data and remote instrument data, historical same period data comparison, and trend fitting analysis.

7. The system of claim 1, wherein, The data comparison and management module specifically comprises: The threshold dynamic configuration unit is configured to self-adaptively adjust a process allowable deviation threshold according to a process parameter importance level and a device operation cycle; the hierarchical early warning unit is configured to divide into multiple layers of early warning according to a difference result and a trend change rate; the trend analysis unit is configured to generate a device operation trend curve and a fault prediction report based on historical inspection data and real-time comparison results; and the linkage response unit is configured to automatically push abnormal information, a field image and a disposal suggestion to a central control system when triggering early warning.

8. The system of claim 4, wherein, The adaptive target detection algorithm includes a dynamic switching mechanism of YOLO algorithm, SSD algorithm and Faster R-CNN algorithm, and the dynamic switching mechanism automatically selects an optimal algorithm according to an instrument model complexity, an image definition and an identification confidence threshold.

9. The system of claim 1, wherein, The system further includes: An offline working module and a data synchronization unit, the offline working module supports completing image acquisition, environment parameter recording, local preprocessing and identification analysis in a network-free environment, and temporarily storing data in a mobile device, and the data synchronization unit is configured to automatically upload offline data to a server after restoring network connection, and complete comparison and data update with the data comparison and management module.

10. An intelligent inspection method, characterized in that, The method is applied to the system of any one of claims 1-9, and the method includes: Collecting a pointer instrument image and a field environment parameter through a data acquisition module integrated in a mobile device; Using an environment adaptive calibration module to pre-process and optimize the collected instrument image according to the field environment parameter; Performing visual identification analysis on the optimized instrument image through a data processing module, and outputting structured reading data; Receiving the structured reading data by means of a data comparison and management module, calling remote instrument data of a corresponding monitoring point and performing bidirectional verification comparison, and outputting hierarchical comparison results and a trend analysis report.

Images