A highway tunnel concealed engineering construction quality visual online detection and reporting method and system and a storage medium

Abstract

The application discloses a highway tunnel concealed engineering construction quality visual online detection and reporting method and system and a storage medium. It relates to the technical field of intelligent detection of tunnel construction. The method comprises the following steps: collecting construction quality image data and binding with space-time information to generate image data packets with space-time correlation identifiers; performing authenticity verification on the image data packets; using a multi-exposure image fusion technology, performing quality enhancement processing on the verified image data to generate a high dynamic range enhanced image; based on the high dynamic range enhanced image, determining the actual size parameters of the target component; based on the actual size parameters of the target component, generating a standardized online reporting report, and according to a preset rule, pushing the reporting report to a terminal with corresponding authority for approval, realizing credible, efficient and intelligent control of the construction quality of the highway tunnel concealed engineering.

Description

Technical Field

[0001] This invention relates to the field of intelligent inspection technology for tunnel construction, and in particular to a method, system, and storage medium for online visualization inspection and reporting of the construction quality of concealed works in highway tunnels. Background Technology

[0002] Concealed works in highway tunnels (including initial support shotcrete, steel arch installation, anchor bolt installation, drainage system, invert construction, etc.) are covered by subsequent processes after construction and cannot be directly re-inspected. According to the "Regulations on Quality Supervision and Management of Highway and Waterway Engineering" and local transportation department regulations, concealed works must undergo self-inspection and supervision acceptance after completion, and compliant image data must be retained before proceeding to the next process. Existing measurement methods cannot simultaneously ensure both testing accuracy and the standardization and traceability of image data. Currently, self-inspection of concealed works during construction relies entirely on manual on-site inspection, and the reporting and approval process is mainly based on paper-based list management. In practice, it can be roughly divided into the following steps:

[0003] The construction unit conducts self-inspection; the supervision unit conducts re-inspection; data is retained and archived; the construction unit submits inspection applications through the information platform and simultaneously submits paper documents; after the supervision unit conducts on-site acceptance and verifies that the documents are correct, it signs the acceptance opinion.

[0004] However, current mainstream practices in the industry suffer from two major flaws. First, data distortion: to avoid rework, some projects reuse the same image data for different parts of the inspection report or falsify data by modifying image content with software, altering data such as shotcrete thickness and anchor bolt anchoring force. Supervisory re-inspections are superficial, failing to verify data authenticity, creating quality "blind spots." Second, chaotic archiving management: image data is not accurately linked to station numbers and mileage, folder naming is inconsistent, making it impossible to locate specific parts during later verification; electronic data is only stored locally without cloud backup, making it susceptible to loss due to equipment failure, and data sharing among all participating parties is not achieved, resulting in low traceability efficiency. It is difficult to achieve tamper-proof traceability of image data throughout the entire process. Summary of the Invention

[0005] This invention provides a method, system, and storage medium for visualized online detection and reporting of construction quality of concealed works in highway tunnels, in order to solve a series of pain points in the quality control of concealed works in the highway tunnel construction industry, such as unreliable data, low detection efficiency, and difficulty in traceability.

[0006] Firstly, a method for visualized online detection and reporting of construction quality of concealed works in highway tunnels is provided, including:

[0007] Collect construction quality image data and bind it with spatiotemporal information to generate image data packets with spatiotemporal correlation identifiers;

[0008] The authenticity of the image data packet is verified.

[0009] Multi-exposure image fusion technology is used to enhance the quality of verified image data and generate high dynamic range enhanced images.

[0010] Based on the high dynamic range enhanced image, the actual size parameters of the target component are determined;

[0011] Based on the actual size parameters of the target component, a standardized online inspection report is generated, and the inspection report is pushed to the terminal with the corresponding permissions for approval according to preset rules.

[0012] Optionally, the step of collecting construction quality image data and binding it with spatiotemporal information to generate an image data packet with spatiotemporal correlation identifiers includes:

[0013] A benchmark mapping table between tunnel mileage station numbers and actual three-dimensional coordinates is pre-established;

[0014] When collecting the construction quality image data, the current three-dimensional coordinates are obtained through a mobile terminal, the current tunnel mileage station is determined based on the benchmark mapping table, and a timestamp is generated simultaneously.

[0015] The tunnel mileage station number, the current three-dimensional coordinates, and the timestamp are combined to generate a unique associated identifier;

[0016] Using binary encapsulation technology, the unique association identifier is written into the custom metadata field of the construction quality image data to generate an image data packet with spatiotemporal association identifier.

[0017] Optionally, the authenticity verification of the image data packet includes:

[0018] Based on the unique association identifier in the image data packet, the time and location of the image data packet are verified for compliance.

[0019] The image feature matching algorithm is used to detect the repetition and tampering of the construction quality image data in the image data packet.

[0020] Optionally, the step of employing multi-exposure image fusion technology to perform quality enhancement processing on the verified image data in the image data packet to generate a high dynamic range enhanced image includes:

[0021] The verified single image in the image data packet is adaptively divided into brightness ranges to obtain multiple brightness ranges corresponding to the single image.

[0022] Calculate the gamma correction value for each brightness range, and perform gamma transformation processing on each single image based on the gamma correction value to generate pseudo-exposure images that present details in different brightness ranges;

[0023] A multi-resolution image pyramid technique is used to weight and fuse the pseudo-exposure images to generate a high dynamic range enhanced image.

[0024] Optionally, determining the actual size parameters of the target component based on the high dynamic range enhanced image using a bifocal measurement model includes:

[0025] For the high dynamic range enhanced image, a scale-invariant feature transform algorithm is used to filter candidate images;

[0026] For each selected candidate image, key imaging parameters are extracted;

[0027] An optical calculation model for pinhole imaging is established based on the key imaging parameters corresponding to each candidate image.

[0028] In three-dimensional space, the optical axes of optical calculation models from different perspectives are unified to establish a standard bifocal measurement model;

[0029] Based on the standard bifocal measurement model, the actual dimensional parameters of the target component are calculated.

[0030] Optionally, based on the actual size parameters of the target component, a standardized online inspection report is generated, and the inspection report is pushed to the terminal with corresponding permissions for approval according to preset rules, including:

[0031] Based on the type of concealed works to which the target component belongs, a preset inspection template is invoked, the actual size parameters of the target component are written into the preset inspection template, and a standardized online inspection report is generated;

[0032] The terminal with the corresponding permissions is determined according to preset rules, and the inspection report is pushed to the terminal with the corresponding permissions for approval.

[0033] Optionally, after generating a standardized online inspection report based on the actual size parameters of the target component, and pushing the inspection report to the terminal with corresponding permissions for approval according to preset rules, the method further includes:

[0034] The entire process data is chronologically sequenced to generate encrypted hash values ​​and uploaded to the blockchain node, and a data access control mechanism is designed.

[0035] Secondly, a visualized online detection and reporting system for the construction quality of concealed works in highway tunnels is provided, characterized in that the system is used to execute the visualized online detection and reporting method for the construction quality of concealed works in highway tunnels as described in any one of the embodiments of the present invention, including:

[0036] The acquisition module is used to acquire construction quality image data and bind it with spatiotemporal information to generate image data packets with spatiotemporal correlation identifiers;

[0037] The verification module is used to verify the authenticity of the image data packet;

[0038] The enhancement module is used to perform quality enhancement processing on the verified image data in the image data packet using multi-exposure image fusion technology to generate a high dynamic range enhanced image;

[0039] The determination module is used to determine the actual size parameters of the target component based on the high dynamic range enhanced image and through a dual-focus measurement model;

[0040] The generation module is used to generate a standardized online inspection report based on the actual size parameters of the target component, and push the inspection report to the terminal with corresponding permissions for approval according to preset rules.

[0041] Optionally, the enhancement module is specifically used for:

[0042] The input single image is adaptively divided into multiple brightness ranges based on its brightness distribution;

[0043] Calculate the gamma correction value for each brightness range, and perform gamma transformation processing on the image based on each gamma correction value to generate a pseudo-exposure image that presents details in different brightness ranges;

[0044] A multi-resolution image pyramid technique is used to weight and fuse the pseudo-exposure images to generate a high dynamic range enhanced image.

[0045] Thirdly, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being used to cause a processor to execute and implement the online visualization detection and reporting method for the construction quality of concealed works in highway tunnels as described in any embodiment of the present invention.

[0046] The technical solution of this invention uses a mobile app as the data acquisition terminal. First, it precisely binds captured images with spatial location and time through a three-dimensional association protocol of "tunnel mileage markers + timestamps + mobile phone positioning," and ensures data authenticity from the source by using binary encapsulation and database dual recording. Subsequently, multi-exposure fusion image enhancement technology is employed to improve the detail and quality of images in low-light tunnel environments, laying the foundation for subsequent intelligent recognition. In the parameter extraction stage, key components are identified based on the YOLO model, and an imaging model is constructed using a dual-focus image matching method combined with mobile phone sensor data to achieve non-contact, automated measurement of key parameters such as exposed anchor bolt length and steel arch spacing. All inspection and reporting processes are completed online. The system automatically verifies the integrity of reporting materials and pushes them to the supervisor's and owner's terminals according to permissions, improving approval efficiency. All data generated throughout the process, from original images and analysis results to approval opinions, is encrypted and sequentially uploaded to the blockchain for evidence storage, ultimately achieving traceability of the source of quality problems, process verification, and accountability.

[0047] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0048] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 This invention provides a flowchart of a method for visual online detection and reporting of the construction quality of concealed works in highway tunnels, as described in Embodiment 1 of the present invention.

[0050] Figure 2 This is a flowchart of an image data quality enhancement process provided in Embodiment 2 of the present invention;

[0051] Figure 3 This is an image brightness grayscale histogram provided in Embodiment 2 of the present invention;

[0052] Figure 4 This is a schematic diagram of multiple pseudo-exposure images provided in Embodiment 2 of the present invention;

[0053] Figure 5 This is a schematic diagram of a high dynamic range enhanced image provided in Embodiment 2 of the present invention.

[0054] Figure 6This is a schematic diagram of the imaging model of image 1 provided in Embodiment 2 of the present invention;

[0055] Figure 7 A schematic diagram of the image 2 imaging model provided in Embodiment 2 of the present invention;

[0056] Figure 8 This is a schematic diagram of the bifocal measurement calculation model after unifying the optical axes of two images provided in Embodiment 2 of the present invention;

[0057] Figure 9 This is a framework diagram of a visual online detection and reporting system for the construction quality of concealed works in highway tunnels, provided in Embodiment 3 of the present invention. Detailed Implementation

[0058] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0059] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0060] Application Overview:

[0061] In the field of hidden works construction quality management for large-scale infrastructure projects such as highway tunnels, the core contradiction has long been the difficulty in simultaneously achieving "reliable data acquisition," "efficient non-destructive testing," and "auditable traceability." The traditional model, relying on manual paper records and on-site sampling, while obtaining first-hand data, is inefficient, poses security risks, and is prone to data tampering and loss, resulting in a fragile traceability chain. Current attempts at digital and intelligent technologies, such as inspection solutions based on fixed 3D scanning or specific robots, while improving the automation level of inspection, suffer from expensive equipment, inflexible deployment, and difficulty in covering all construction blind spots. Furthermore, these solutions typically focus only on the "inspection" process itself, failing to deeply integrate with management processes such as "reporting," "approval," and "evidence storage," creating new "data silos." More critically, these solutions generally lack effective technical constraints on the authenticity of data sources, failing to fundamentally address the industry's persistent problems of falsified image data and distorted inspection data, leading to a weak foundation of trust in digital management. Therefore, the market urgently needs an integrated solution that can run through the entire construction quality control chain, ensure the authenticity and reliability of data from the source, and achieve efficient automated testing and seamless process collaboration.

[0062] To address the aforementioned contradictions, this invention proposes a visualized online inspection and reporting method for the construction quality of concealed works in highway tunnels, integrating mobile data acquisition, artificial intelligence, and blockchain technologies. The technological breakthrough of this solution lies in constructing a "spatiotemporal three-dimensional association" protocol. This protocol uses ordinary mobile terminals (such as smartphones) to strongly encrypt and bind images with precise tunnel station numbers, millisecond-level timestamps, and equipment location information during image acquisition, establishing anti-counterfeiting and unique identifiers from the source of data generation, thus solving the fundamental problem of data reliability. Building upon this, the invention innovatively applies multi-exposure image fusion technology from computational photography to image enhancement in low-light tunnel environments. Furthermore, a non-contact measurement model based on dual-focus vision and mobile terminal inertial measurement unit (IMU) data is designed, enabling high-precision, automated, and non-destructive measurement of key parameters such as steel arch spacing and exposed anchor bolt length in complex field environments, significantly improving inspection safety and efficiency.

[0063] Furthermore, it establishes a fully digital closed loop encompassing on-site data collection, intelligent analysis, online reporting, and multi-party approval. Blockchain technology is used to hash and store all data (including original images, analysis results, and approval records), forming an immutable, access-controlled, and traceable quality evidence chain. Compared to existing technologies, this invention is the first to systematically integrate data anti-counterfeiting, intelligent detection, and process traceability in an engineering quality management scenario. It achieves a paradigm shift from "human-based" to "technology-based" and from "result inspection" to "process transparency," providing the industry with a reliable, efficient, and complete digital quality management system.

[0064] Example 1:

[0065] Figure 1 This invention provides a flowchart of a method for online visualization detection and reporting of the construction quality of concealed works in highway tunnels, as described in Embodiment 1 of the present invention. This embodiment is applicable to situations involving online visualization detection and reporting of the construction quality of concealed works in highway tunnels. This method can be executed by an online visualization detection and reporting system for the construction quality of concealed works in highway tunnels. Figure 1 As shown, the method includes: S110, acquiring construction quality image data and binding it with spatiotemporal information to generate an image data packet with spatiotemporal correlation identifier; S120, verifying the authenticity of the image data packet; S130, using multi-exposure image fusion technology to enhance the quality of the verified image data in the image data packet to generate a high dynamic range enhanced image; S140, based on the high dynamic range enhanced image, determining the actual size parameters of the target component through a bifocal measurement model; S150, based on the actual size parameters of the target component, generating a standardized online inspection report, and pushing the inspection report to the terminal with corresponding permissions for approval according to preset rules.

[0066] In this embodiment, construction quality image data is collected and bound to spatiotemporal information to generate image data packets with spatiotemporal correlation identifiers. The authenticity of these image data packets is verified. Multi-exposure image fusion technology is used to enhance the quality of the verified image data, generating high dynamic range (HDL) enhanced images. Based on these HDL enhanced images, the actual size parameters of the target component are determined. Based on the actual size parameters of the target component, a standardized online inspection report is generated, and the report is pushed to the terminal with the corresponding permissions for approval according to preset rules. This ensures the authenticity and tamper-proof nature of the inspection data from the source, improves inspection security and accuracy through non-contact visual measurement, and upgrades the traditional manual, paper-based quality management model to a digital, automated, and reliable evidence chain system through full-process online collaboration and traceability. This systematically solves industry pain points such as data distortion, low efficiency, and difficulty in traceability. It achieves reliable, efficient, and intelligent control over the construction quality of hidden works in highway tunnels.

[0067] Example 2:

[0068] The technical solution in this embodiment is a further refinement based on the above embodiments.

[0069] In step S110, construction quality image data is acquired and bound to spatiotemporal information to generate an image data packet with spatiotemporal correlation identifiers, including:

[0070] A benchmark mapping table between tunnel mileage station numbers and actual three-dimensional coordinates is pre-established;

[0071] When collecting the construction quality image data, the current three-dimensional coordinates are obtained through a mobile terminal, the current tunnel mileage station is determined based on the benchmark mapping table, and a timestamp is generated simultaneously.

[0072] The tunnel mileage station number, the current three-dimensional coordinates, and the timestamp are combined to generate a unique associated identifier;

[0073] Using binary encapsulation technology, the unique association identifier is written into the custom metadata field of the construction quality image data to generate an image data packet with spatiotemporal association identifier.

[0074] The tunnel mileage marker refers to the unique number used to identify a specific longitudinal location within a tunnel in highway tunnel design drawings and actual construction. It is a sequential location code, such as "K5+280," indicating a location 5 kilometers and 280 meters from the tunnel's starting point. Actual three-dimensional coordinates refer to three values ​​representing the location of a point in space within a specific coordinate system (such as the National Geodetic Coordinate System or the Engineering Independent Coordinate System), typically including east (X), north (Y), and elevation (Z). Inside the tunnel, it is obtained through positioning technology (such as an indoor positioning system) to accurately describe the physical location when the mobile phone is taking the picture. A reference mapping table can refer to a lookup table pre-created and stored in the system's backend database. This table establishes the correspondence between tunnel mileage markers and their corresponding design three-dimensional coordinates. A mobile terminal specifically refers to a smartphone with a dedicated application installed. As a front-end tool for image acquisition, primary data processing, and spatiotemporal information binding, it integrates a camera, positioning module, time module, and computing unit. The current three-dimensional coordinates refer to the three-dimensional coordinates of the terminal's location obtained in real time at the moment the image is captured by the positioning module built into the mobile terminal (such as GPS, Beidou, or indoor positioning technologies such as UWB and Bluetooth beacons used in tunnels).

[0075] A timestamp can refer to a numerical sequence accurate to the millisecond level, used to uniquely identify the absolute point in time when an image was captured; it is synchronously generated by the system server or mobile phone system when the shutter is triggered, ensuring the authority and tamper-proof nature of the time. A unique identifier can refer to a string composed of elements such as "tunnel mileage marker," "timestamp," and "device identifier" combined according to specific rules. Its core characteristic is global uniqueness, enabling a strong association between a single shooting action and its generated data packet, much like a digital fingerprint. Binary encapsulation technology can refer to a data processing method that writes the "unique identifier" information directly into a system-defined data area within the image file in the form of raw binary data stream, without any visual character encoding. This is distinct from the general EXIF ​​information area, making it more concealed and less susceptible to modification by general software. Custom metadata fields can refer to reserved or custom-extended data storage areas in standard JPEG, PNG, and other image file formats, in addition to the areas storing pixel data and the standard EXIF ​​information area. These areas are used to store the "unique identifier," achieving a physical binding between system-private data and the image file. An image data packet refers to the final generated complete data unit containing the original image data and embedded spatiotemporal correlation information (unique identifier). It is not just a photograph, but a data entity carrying an indivisible spatiotemporal identity.

[0076] Specifically, the process involves pre-extracting the "mileage markers" of all milestones along the entire route and their corresponding precise "3D coordinates" in the design coordinate system from the construction design drawings of the highway tunnel. Then, the paired "mileage marker-coordinate" data is batch-imported into the system server's database using a backend management tool, creating a data table named "Baseline Mapping Table." This table's structure includes at least the fields "mileage marker," "design coordinate X," "design coordinate Y," and "design coordinate Z." For example, the marker "K5+250" corresponds to coordinates (1000.000, 2000.000, 50.000). The wireless positioning network (such as UWB) within the tunnel will be calibrated against this table to ensure that the positioning results of mobile terminals match the design markers, with a positioning error ≤5cm in the tunnel's indoor environment.

[0077] At highway tunnel construction sites, workers can use mobile devices (e.g., smartphones) with a dedicated application (APP) to access their cameras and photograph key, concealed areas within the tunnel, such as anchor bolts, steel arch frames, waterproofing membrane paving surfaces, and invert arch pouring surfaces. During photography, the application automatically activates the mobile device's location function. Inside the tunnel, the mobile device connects to a pre-deployed indoor high-precision positioning system (such as UWB) and obtains the phone's current 3D coordinates (e.g., X: 1000.523, Y: 2000.312, Z: 50.148) in real time. The dedicated application or backend server then matches these 3D coordinates with a reference mapping table in a database. The distance between the current 3D coordinates and all preset 3D coordinate points in the mapping table is calculated, and the closest point is determined. The corresponding "mileage marker" of this point is then "determined" as the current location's marker (e.g., if the nearest point is K5+280, then the marker is determined to be K5+280). The moment the user presses the shutter button, a dedicated application requests an authoritative, millisecond-accurate current time from the system time server, or calls the phone's high-precision clock to synchronously generate a "timestamp" in the format "YYYY-MM-DD-HH-MM-SS-SSS", for example, "2026-01-22-14-30-25-123". This action is triggered almost simultaneously with obtaining coordinates and determining the station number.

[0078] After obtaining the mileage marker, current 3D coordinates, and timestamp, the system concatenates them according to a preset, immutable string format rule. For example, it uses the format "mileage marker - last 6 digits of timestamp - last 4 digits of mobile device ID". Specifically, it takes the "mileage marker" (e.g., K5+280), the last 6 digits of the timestamp (e.g., 025123), and the last four digits of the mobile device's unique identifier (e.g., 3579). Then, it connects them with a hyphen "-" to generate the final unique identifier "K5+280-25123-3579". This ensures that the probability of repeatedly taking photos at the same location, within the same millisecond, using the same mobile phone is extremely low, thus achieving global uniqueness of the identifier.

[0079] Using binary encapsulation technology, the unique association identifier is transformed into a binary data stream through encoding conversion (such as UTF-8 encoding). The file structure of the newly captured JPEG or PNG image file is determined via a file operation interface. Outside the standard EXIF ​​information block, a reserved or custom data segment (i.e., a "custom metadata field") is located. This binary data stream is written to the specified location in this field, generating an image data packet with a spatiotemporal association identifier. This is distinct from the native EXIF ​​information, preventing tampering. This image data packet contains the original pixel data and also embeds a spatiotemporal identity that cannot be easily separated—the unique association identifier. Simultaneously, a relational database table is established in the APP backend, consisting of "association identifier - image file - mileage marker information - time information - location information." This achieves dual binding between image data and three-dimensional relational information (i.e., tunnel mileage marker + timestamp + mobile phone location coordinates), ensuring that even if the image file is copied separately, its source can be traced through the relational database table.

[0080] In addition, the dedicated application guides users to take photos of the same target object from at least two different angles or distances to ensure the feasibility of subsequent measurements. Furthermore, supplementary lighting can be turned on according to the tunnel site environment to improve the quality of the acquired images.

[0081] In this embodiment, by using a mobile APP as the data acquisition terminal, the captured images are first precisely bound to spatial location and time through a three-dimensional association protocol of "tunnel mileage marker + timestamp + mobile phone positioning". The authenticity of the data is ensured from the source by using binary encapsulation and database dual recording.

[0082] Optionally, after shooting, the system (either locally on the mobile app or on a backend server) reverse-parses the "custom metadata field" from the newly generated "image data packet," reads the written binary data, and decodes it to restore the "unique association identifier" stored in the file, denoted as identifier A. Simultaneously, the system retrieves the same "unique association identifier" generated at the moment of shooting and prepared for writing from the temporary cache or backend transaction records of this shooting event, denoted as identifier B. Identifier A and identifier B are compared to see if they are completely identical. If they are identical, it proves that the encapsulation process was successful and the identifier is completely bound to the image file; if they are inconsistent or the reading fails, it means that the file is corrupted or an error occurred in the encapsulation process, and the verification fails.

[0083] Furthermore, the system obtains two key coordinates from this shooting record: one is the "current three-dimensional coordinates" (measured value) obtained through the mobile phone positioning module; the other is the "design three-dimensional coordinates" (baseline value) corresponding to the "tunnel mileage station" determined by matching the measured coordinates with the "baseline mapping table".

[0084] Calculate the three-dimensional Euclidean distance between the measured coordinates and the corresponding design coordinates. This distance value is the "deviation between the positioning coordinates and the mileage marker reference coordinates". Compare this deviation value with a preset allowable threshold (e.g., 5 cm mentioned in the document).

[0085] If the "match integrity" verification is successful and the calculated "coordinate deviation" is less than or equal to the preset threshold (e.g., ≤5cm), the system determines that the collected data is valid and the process proceeds to the next stage (e.g., uploading or analysis).

[0086] If the "match integrity" check fails, or the "coordinate deviation" exceeds a preset threshold (e.g., >5cm), the system will automatically trigger an alert. The alert can be immediately displayed to the operator via the mobile app interface, accompanied by a pop-up dialog box, an audible alert, or a screen flashing. The alert will clearly indicate the problem, such as "Data binding error, please retake the photo" or "Positioning deviation is too large, please calibrate or reposition before taking the photo," guiding the operator to retake the photo or check the positioning environment.

[0087] In this embodiment, through internal data consistency verification and external spatial accuracy check, it is ensured that each image data packet is not only complete and reliable, but also that the spatial location information it records meets the accuracy requirements of engineering management, thus guaranteeing the reliability and availability of the data from the source.

[0088] In step S120, the authenticity of the image data packet is verified, including:

[0089] Based on the unique association identifier in the image data packet, the time and location of the image data packet are verified for compliance.

[0090] The image feature matching algorithm is used to detect the repetition and tampering of the construction quality image data in the image data packet.

[0091] Authenticity verification refers to a verification process aimed at confirming that the claimed acquisition time, location, and image content of the "image data package" are genuine and tamper-proof, and not reused, forged later, or misappropriated from other locations. Location compliance refers to whether the actual physical location of the mobile terminal at the time of image acquisition is spatially consistent with the tunnel construction site (such as a specific mileage marker interval) specified in the verification application. Compliance check determines whether the shooting occurred at the designated work site. Image feature matching algorithms refer to a class of computer vision algorithms used to detect and compare the similarity of local features in different images. The most typical algorithm is scale-invariant feature transform. This algorithm can extract key points (feature points) and their descriptors from an image that remain invariant to rotation, scaling, and brightness changes, and is used to accurately match the same scene or object in two images.

[0092] Specifically, standard EXIF ​​information is read from the image data packet to extract the "capture time T1" recorded by the phone itself. Simultaneously, the system timestamp T2 generated during spatiotemporal binding is retrieved from the associated database table linked to this data packet. The difference between T1 and T2 is calculated. Due to the extremely short delays in capturing, processing, and binding, a very small, reasonable time difference is allowed, for example, within a few seconds, such as 6 seconds. If the time difference is within a reasonable range, the time verification passes; if the time difference is abnormal, for example, several days, such as 6 days; or if T1 is significantly earlier than T2, indicating a logical error, the time is deemed questionable.

[0093] Retrieve the "mileage marker L" and "location coordinates P1" bound to this data collection from the associated database table. Simultaneously, retrieve the "tunnel section R" declared by the construction party for acceptance from the submitted "inspection application," which is typically a mileage marker range, such as K5+250 to K5+280. Check whether mileage marker L falls within the declared section R. Additionally, spatial relationship calculations can be performed between coordinates P1 and the design coordinate range of section R. If the locations match, the location compliance verification passes; if they do not match, it indicates "a photo was taken at location A but declared as location B," and the verification fails.

[0094] Image feature matching algorithms such as SIFT are used to extract full-image feature points from the current "image data package". Then, a full-database scan and matching of feature points is performed in the historical image database (especially historical photos of the same tunnel or similar locations). The algorithm calculates the number of matching feature points between the current image and each historical image. If the number of matching points with a historical image exceeds a high preset threshold, for example, if the matching point ratio exceeds 90%, it indicates that the two images are almost identical, and the current image is judged to be "reused" and suspected of being forged.

[0095] Furthermore, also based on feature matching algorithms, the consistency of features within a single image is analyzed. For images suspected of being stitched together, there may be subtle differences in the perspective relationships, lighting conditions, and noise patterns of different regions. More advanced detection methods analyze the statistical consistency of local image features or look for unnatural boundary seams. If the algorithm detects obvious "feature breaks" or "geometric perspective contradictions" in the image that do not conform to the imaging rules of a single shot, it determines that the image may have been stitched together and tampered with. For example, in a photograph of an anchor bolt, the features of some anchor bolt areas and the features of the background concrete show inexplicable abrupt changes in scale and direction.

[0096] In this embodiment, a triple verification mechanism is implemented to compare the original EXIF ​​information of the image with the system timestamp; verify whether the mobile phone location coordinates match the reported area; and use the SIFT feature matching algorithm to detect duplicate or stitched images. Through the integration of multiple technologies, the image data achieves spatiotemporal uniqueness, immutability, and verifiable authenticity from the source of acquisition, laying a reliable data foundation for subsequent intelligent recognition and blockchain evidence storage.

[0097] In step S130, multi-exposure image fusion technology is used to enhance the quality of the verified image data in the image data packet, generating a high dynamic range enhanced image, including:

[0098] The verified single image in the image data packet is adaptively divided into brightness ranges to obtain multiple brightness ranges corresponding to the single image.

[0099] Calculate the gamma correction value for each brightness range, and perform gamma transformation processing on each single image based on the gamma correction value to generate pseudo-exposure images that present details in different brightness ranges;

[0100] A multi-resolution image pyramid technique is used to weight and fuse the pseudo-exposure images to generate a high dynamic range enhanced image.

[0101] In this context, "brightness range" refers to dividing the brightness values ​​(from the darkest 0 to the brightest 1 or 0-255) of all pixels in a digital image into several consecutive brightness value ranges. For example, the entire brightness range can be divided into 8 ranges, each containing pixels within a certain brightness range. This division is for the purpose of differentiating image regions with different levels of brightness. "Gamma correction value" refers to the specific value of γ, a parameter that controls the shape of the gamma transform curve. γ < 1 will brighten the overall image (enhance dark areas), while γ > 1 will darken the overall image (suppress bright areas). Gamma transform processing can be considered a non-linear operation used to adjust the brightness response curve of an image. Its basic formula is: Output brightness = (Input brightness)^γ. By applying different γ values, the contrast of different brightness regions of the image can be compressed or expanded in a targeted manner, thereby revealing hidden details.

[0102] Pseudo-exposure images refer to a series of images generated by applying different "gamma correction values" to the same original input image. These images appear as a collection of photographs of the same scene taken under different exposure parameters; some emphasize details in shadows, while others emphasize details in highlights, hence the name "pseudo-exposure" images. Multi-resolution image pyramid technology refers to a multi-scale image representation method that constructs an image as a series of sets arranged from high to low resolution (from large to small size), resembling a pyramid. This technique allows algorithms to analyze and process image features at different scales. High dynamic range enhanced images refer to computationally generated images with high dynamic range. Compared to ordinary photographs, they can simultaneously and clearly present details in both extremely dark and extremely bright areas of a scene, possessing richer brightness levels and more realistic detail, providing a higher-quality data foundation for subsequent visual analysis tasks.

[0103] Specifically, due to the complex environment inside tunnels, simple image preprocessing methods cannot effectively enhance images. This invention, based on the multi-exposure image fusion technique of computational photography, uses adaptive gamma transform to generate multiple pseudo-exposure images from a single low dynamic range image, fully exposing details in each region of the tunnel face image. Finally, the Laplacian pyramid technique is used to fully fuse the images, ultimately generating a detailed high dynamic range image of the tunnel face. High-quality images are a prerequisite for accurate identification of construction quality parameters. Figure 2 This is a flowchart of an image data quality enhancement process provided in Embodiment 2 of the present invention, as follows: Figure 2 As shown, the specific steps of image data quality enhancement processing include:

[0104] (1) Adaptive segmentation of the image brightness region is performed on the input single low-quality image. The brightness value range of the image is normalized to the [0,1] interval. In order to balance detail preservation and computational efficiency, the brightness interval is divided into 8 sub-intervals in this embodiment. The segmentation method adopts the maximum inter-class variance method. Through iterative calculation, the segmentation threshold that maximizes the pixel brightness variance within the interval is found, thereby ensuring that each segmented interval has obvious visual differences. Figure 3 This is an image brightness grayscale histogram provided in Embodiment 2 of the present invention, such as... Figure 3 As shown, after segmentation, the first and last two intervals, the darkest (almost black) and the brightest (overexposed white), contain very little effective detail. To ensure the quality of the fused image, the first and last brightness intervals are removed, and only the brightness of the six middle intervals is calculated using regional gamma values.

[0105] (2) In order to present the image details in different brightness ranges, it is necessary to calculate the most suitable gamma transformation value. For the six brightness regions of the image, six adaptive gamma values ​​need to be calculated respectively. Then, the image is processed by gamma transformation to generate a series of pseudo-exposure images. Each pseudo-exposure image can present different local geological information well. Finally, the different local information is fused on a complete image by image fusion, so as to expose the information details of the entire face to the maximum extent.

[0106] However, excessive correction can introduce too much noise, affecting image sharpness and color accuracy. Therefore, choosing an appropriate gamma value is crucial for gamma correction. This invention determines the gamma value based on the overall brightness of the image to be corrected. The overall brightness of the image can be quantified using the brightness tendency index. The gamma value is linearly correlated with the overall brightness of the image, and its calculation method is as follows:

[0107] ;

[0108] ;

[0109] ;

[0110] Where L(x,y) represents the pixel brightness value at each point in the image. This represents the maximum image brightness. This represents the minimum image brightness. Let N be the average of the logarithms of the image brightness, and N be the total number of pixels in the image. It is a decimal, and to avoid the logarithm of L(x,y) being 0 causing the value to tend to infinitesimal. The gamma value for adaptive brightness adjustment. Figure 4 This is a schematic diagram of multiple pseudo-exposure images provided in Embodiment 2 of the present invention, as shown below. Figure 4 As shown, after obtaining γ, different coefficients k are combined with γ to generate six different transformed gamma values ​​for the six brightness ranges. (i=1 to 6). Next, using each Perform a gamma transform on the original image to obtain 6 pseudo-exposure images. Each This will allow the corresponding brightness range details to be presented in the best possible way.

[0111] (3) Based on multiple pseudo-exposure images, a multi-resolution Laplacian image pyramid fusion operation is further adopted to obtain a high dynamic range image with rich details. During the image fusion process, the proportion of each image in the fusion process is determined by measuring the three major image quality indicators of contrast, saturation and exposure. Flat, colorless underexposed and overexposed areas are removed during the fusion process by weight allocation.

[0112] The local contrast of an image is represented by its high-frequency amplitude values. To preserve as much detail as possible, images with high contrast are given higher weights. The contrast index value is calculated as follows:

[0113] ;

[0114] in, Let be the pixel value of the k-th image in the image sequence.

[0115] For saturation, the variance of the R, G, and B channels is used to represent contrast. Areas with low saturation are assigned lower weights to eliminate overexposed and underexposed areas, while areas with high saturation are assigned higher weights to retain richer colors. The saturation index is calculated as follows:

[0116] ;

[0117] Exposure quality is measured by the distance between the image's brightness value and the brightness value of 0.5. A closer distance indicates a greater distance from underexposed brightness (0) and overexposed brightness (1), resulting in better exposure quality. The exposure index is calculated as follows:

[0118]

[0119] in, This represents the variance of the image pixel values.

[0120] Based on the above index calculation results, the image fusion method is as follows:

[0121] ;

[0122]

[0123] in, The pixel values ​​of the merged image. For image fusion weights, This is the weighted normalized value. Figure 5 This is a schematic diagram of a high dynamic range enhanced image provided in Embodiment 2 of the present invention. The final fusion effect is as follows: Figure 5 As shown.

[0124] In this embodiment, for the complex lighting environment of the tunnel, a pseudo-exposure image generation based on adaptive gamma transform and Laplacian pyramid fusion technology are used to generate a high dynamic range image to improve recognition accuracy.

[0125] In step S140, determining the actual size parameters of the target component based on the high dynamic range enhanced image and using a bifocal measurement model includes:

[0126] For the high dynamic range enhanced image, a scale-invariant feature transform algorithm is used to filter candidate images;

[0127] For each selected candidate image, key imaging parameters are extracted;

[0128] An optical calculation model for pinhole imaging is established based on the key imaging parameters corresponding to each candidate image.

[0129] In three-dimensional space, the optical axes of optical calculation models from different perspectives are unified to establish a standard bifocal measurement model;

[0130] Based on the standard bifocal measurement model, the actual dimensional parameters of the target component are calculated.

[0131] The actual size parameters of the target component refer to the physical dimensions of specific objects in the tunnel concealed engineering that need to be measured, such as steel arches, anchor bolts, and reinforcing bars, in real three-dimensional space. Common parameters include spacing (such as the spacing between arches), length (such as the exposed length of anchor bolts), and diameter. Scale-invariant feature transform algorithms can refer to a classic computer vision local feature detection and description algorithm. Its core capability is to extract key points from images that are invariant to scale, rotation, and brightness changes, and generate descriptors describing the areas surrounding these key points. It is mainly used to filter out multiple images of the same scene (i.e., the same target component) by comparing the degree of matching of feature points between different images.

[0132] Candidate images refer to images obtained after being filtered by the "scale-invariant feature transform algorithm". These images have enough matching feature points between each other, thus being determined by the system to be images of the same target component taken from different perspectives. These images form the data foundation for subsequent 3D geometric calculations.

[0133] Key imaging parameters refer to the mathematical and physical parameters used to accurately describe the imaging process of candidate images. They mainly include two categories: intrinsic parameters, which are usually obtained from the image's EXIF ​​information and primarily refer to information such as focal length (the distance from the lens center to the imaging sensor, which determines the angle of view) and sensor size, used to construct the camera's imaging geometry model; and extrinsic parameters, which mainly refer to the camera's attitude (such as pitch, roll, and yaw angles) at the moment of capture, recorded by inertial measurement units such as the phone's gyroscope and gravity sensor. These parameters define the camera's orientation in three-dimensional space.

[0134] The optical computational model of pinhole imaging refers to a simplified yet effective mathematical model of camera imaging. It abstracts the camera as a "pinhole" without lens distortion, through which light from objects in the scene is projected onto the imaging plane. Establishing this model for each "candidate image" essentially uses mathematical formulas to describe the precise geometric projection relationship between its three-dimensional spatial points, projection centers, and two-dimensional image pixels. Optical axis unification refers to a coordinate transformation process. Since each image is captured by a camera in different positions and orientations, its respective "pinhole imaging model" exists in different coordinate systems. Optical axis unification means transforming the camera coordinate systems corresponding to different images to the same global three-dimensional world coordinate system through mathematical transformations (rotation and translation). The optical axes (i.e., the rays passing from the projection center through the center of the imaging plane) and imaging relationships of all imaging models are expressed within a unified reference frame. A standard bifocal measurement model can refer to the joint geometric model of imaging systems from two different perspectives of the same scene obtained after completing optical axis unification. This model clearly expresses the triangular relationship between a three-dimensional spatial point, the projection centers of the two cameras, and the projection points of that point on the two images. It is the core mathematical model for calculating three-dimensional spatial coordinates from two-dimensional image coordinates. "Dual focal length" emphasizes that the model originates from observations at two different focal lengths or viewpoints.

[0135] Specifically, for multiple images uploaded in the same batch targeting the same concealed engineering location, a scale-invariant feature transform algorithm is used to detect and match feature points pairwise. Image pairs with more than a preset threshold of matching feature points are selected as candidate image pairs that can be used for 3D measurement. Each image pair provides two different perspectives of the same scene.

[0136] The EXIF ​​information of each image, along with gyroscope and gravity sensor information recorded during shooting, is extracted. The EXIF ​​information is primarily used to identify the focal length and pixel dimensions of the captured image. Focal length determines the camera's field of view, and pixel dimensions are used to convert pixel distances in the image into physical distances on the image plane. Gyroscope and gravity sensor information are mainly used to identify the phone's tilt attitude during shooting, such as pitch, yaw, and roll angles, primarily reflecting the rotational relationship between the image coordinate system (o-uv) and the world coordinate system (O-XYZ).

[0137] By using EXIF ​​data and sensor data, the intrinsic parameter matrix K and rotation matrix R can be determined, thus establishing a unique imaging geometry for each captured candidate image, known as the "pinhole imaging model." The pinhole imaging model is a geometric model in computer vision that projects three-dimensional spatial points (X, Y, Z) onto two-dimensional image points (u, v). Its mathematical model is typically expressed as: s*[u,v,1] T =K*[R|t]*[X,Y,Z,1] T ;

[0138] Where, [u, v, 1] T These are the homogeneous coordinates of the image pixels. s is a scaling factor. K is the camera intrinsic parameter matrix, derived from the focal length (…). , ) and principal point ( , The parameters consist of [R|t], which can be calculated or estimated from the EXIF ​​information. [R|t] is the camera extrinsic parameter matrix, comprising a 3x3 rotation matrix R and a 3x1 translation vector t. The rotation matrix R is calculated from the attitude angles provided by the phone's gyroscope and gravity sensor, describing the orientation of the phone (camera) coordinate system relative to a world coordinate system (e.g., a coordinate system with its origin at a point in the tunnel). The translation vector t represents the position of the camera's optical center in the world coordinate system; in the initial single-image model, its value may be unknown or set as a relative value.

[0139] In three-dimensional space, the optical axes of models from different perspectives are unified to establish a standard bifocal measurement model. Optical axis unification is essentially a coordinate system transformation. The specific steps are as follows:

[0140] 1. Establish a unified world coordinate system: Usually, the coordinate system of one of the cameras (e.g., camera O1) is chosen as the unified world coordinate system. In this case, the extrinsic parameter [R1|t1] of camera O1 can be set to [I|0] (i.e., rotation is the identity matrix and translation is the zero vector).

[0141] 2. Calculate relative pose: For another camera O2, we need to find its pose relative to O1. This can be achieved in one of the following two ways:

[0142] Method A (sensor-based): If the timestamps of the two cameras (phones) are very close, and their absolute poses relative to the global plane (such as the tunnel axis) are known (from the sensors), then the rotation matrix of O2 relative to O1 can be directly calculated. Translation vector .

[0143] Method B (Image Matching Based): A more general approach is to utilize epipolar geometry. By matching the pixel coordinates of the same object feature points in two images (i.e., SIFT matching point pairs), the essential matrix or fundamental matrix between the two cameras can be solved, and then the rotation of camera O2 relative to O1 can be obtained. Translation direction (The scale of the translation needs to be determined later.)

[0144] 3. Establish a unified model: Transform the model of camera O2 to the coordinate system of O1. At this time, for the same point P in space, the projection points p1 and p2 on the imaging planes of the two cameras satisfy certain geometric constraints (epochal constraints), and they form a triangle (O1-O2-P) with the spatial point P.

[0145] The standard dual-focus measurement model refers to the complete stereoscopic vision geometric model established after completing the coordinate system described above. This model includes the optical centers of two cameras (O1, O2), their imaging planes, and the measured object point (P). This model is the foundation for 3D reconstruction and distance measurement.

[0146] Figure 6 This is a schematic diagram of the imaging model of image 1 provided in Embodiment 2 of the present invention; Figure 7 This is a schematic diagram of the image 2 imaging model provided in Embodiment 2 of the present invention; Figure 8 This is a schematic diagram of the bifocal measurement calculation model after unifying the optical axes of two images provided in Embodiment 2 of the present invention; as shown. Figure 6-8 As shown. Based on the dual-focal distance measurement model with unified optical axes, the following three equations are established using similar triangle relationships. These equations can be solved simultaneously to obtain the actual parameters of the object being photographed. For example, assuming that the principal optical axes of the two cameras are parallel to each other and both are perpendicular to the plane where the object (H) is located, under this setting, the object distance L is the perpendicular distance from the optical center to the object plane, which greatly simplifies the calculation.

[0147] ;

[0148] ;

[0149] The above formulas are directly derived from similar triangles in the pinhole imaging model. Formulas with parameter subscript 1 represent the camera imaging relationship for the first image, and formulas with parameter subscript 2 represent the camera imaging relationship for the second image. Taking the camera imaging relationship of the first image as an example (parameter subscript 1):

[0150] There are two similar triangles: the object triangle, with the camera optical center as the vertex, the actual height H of the object as the opposite side, and the object distance L1 as the adjacent side; and the image triangle, with the camera optical center as the vertex, the image height r1 of the object as the opposite side, and the focal length f1 as the adjacent side. Based on the principle that "corresponding sides are proportional" in similar triangles, we can directly derive: H:r1 = L1:f1, which is the formula with the parameter subscript 1. The second formula is established in the same way, only using the parameters from the second image. Here, H is the actual physical size of the object being photographed, either its height or width. r1 and r2 are the image heights, referring to the pixel dimensions of the target object H on the imaging plane of the first and second candidate images. It directly reflects the size of the object in the image and can be accurately obtained through image recognition and feature point matching. f1 and f2 are the effective focal lengths, referring to the distance from the optical center of the camera lens to the imaging sensor plane when the first and second candidate images are captured. This is a key internal parameter, which can usually be directly read from the EXIF ​​information of the image. L1 and L2 are object distances, referring to the vertical distance from the optical center of the camera when the first and second candidate images are captured to the plane where the target object is located.

[0151] ;

[0152] This formula is a direct mathematical expression of the optical axis unification assumption. It implies that the two cameras are placed on a common measurement reference line. The physical meaning of L1 + f1 on the left side of the equation refers to the total distance from the imaging sensor plane of the first camera to the plane of the object being measured. This is because L1 is the distance from the optical center to the object, and f1 is the distance from the optical center to the sensor plane. Similarly, L2 + f2 on the right side is the total distance from the imaging sensor plane of the second camera to the same plane of the object being measured. This equation holds true only if the sensor planes of both cameras are parallel to the plane of the object being measured. In practice, this can be achieved by using data from the phone's attitude sensor (gyroscope, gravity sensor) and performing virtual rotation correction on the image during calculation, thus "unifying" the optical axis direction and allowing the ideal model to be applied.

[0153] Solving the three equations simultaneously for the actual parameter H, we find three unknowns: H, L1, and L2. Here, f1, f2, r1, and r2 are known measured values. From the above formulas, we can obtain:

[0154] ;

[0155] ;

[0156] Substitute the expressions for L1 and L2 into get:

[0157] = ;

[0158] By rearranging the equations, we obtain a linear equation in one variable H. Solving this equation yields the actual size H of the object. Once H is obtained, the object distance between the two cameras can be calculated using the formula above. This avoids complex absolute coordinate calculations; only the assumptions of focal length, image height, and parallel optical axes are needed to calculate the size through simple proportional relationships.

[0159] In this embodiment, the YOLO model is used to locate key components such as steel arch frames and anchor rods. By matching images of the same component from different perspectives, the EXIF ​​data of the images and mobile phone sensor data (gyroscope, gravity sensor) are extracted to construct a unified optical axis imaging model. Based on the principle of similar triangles, non-contact, high-precision automatic measurement of key parameters such as component spacing and diameter is realized, which solves the problems of low efficiency and high risk of traditional manual measurement in concealed engineering.

[0160] In step S150, based on the actual size parameters of the target component, a standardized online inspection report is generated, and the inspection report is pushed to the terminal with corresponding permissions for approval according to preset rules, including:

[0161] Based on the type of concealed works to which the target component belongs, a preset inspection template is invoked, the actual size parameters of the target component are written into the preset inspection template, and a standardized online inspection report is generated;

[0162] The terminal with the corresponding permissions is determined according to preset rules, and the inspection report is pushed to the terminal with the corresponding permissions for approval.

[0163] The standardized online inspection report refers to the concealed works quality acceptance application form generated in a unified, predefined electronic format. Its structure, required fields, and attachment requirements are all predefined by the system template, ensuring that all inspection data is formatted correctly, complete, and directly processable by a computer, forming the foundation for digital approval. Preset rules refer to the set of logical judgment conditions pre-configured in the system to drive business processes and control permissions. These mainly include two categories: approval process rules, which define the approval paths for different types of project inspections (e.g., initial support inspection requires review by a professional supervising engineer before approval by the chief supervising engineer); and permission and push rules, which define the approval permissions of different user roles and the priority of task notifications (e.g., priority reminders for key processes). The corresponding permission terminal refers to the client device and software logged in and used by users with specific approval permissions. According to the preset rules, the system will automatically push the inspection report to the user terminal with corresponding approval responsibilities, such as the mobile app of the supervising engineer responsible for the construction area or the web backend logged in by the owner's representative.

[0164] The concealed works type of the target component refers to the category of works that will be covered after construction, as defined in the tunnel engineering design. Examples include "initial support," "secondary lining," "drainage system," and "invert arch filling." Different types correspond to different quality inspection standards and report formats. The preset inspection template refers to an electronic form framework pre-created in the system corresponding to the "concealed works type." The template pre-sets the list of items that must be inspected for this type of work, the data fields to be filled in for each item, the corresponding design specification qualification standards, and the required supporting documentation (such as video footage).

[0165] Specifically, based on different types of concealed works (such as initial support and waterproofing), a pre-set inspection template is automatically invoked. This template includes a list of necessary inspection items, design specification values, and required accompanying image data for that type of project. The actual dimensional parameters of the target component are entered into the corresponding item. Simultaneously, the measured value of this parameter is automatically compared with the associated acceptance standard, such as the allowable deviation of ±50mm for arch spacing, and a preliminary judgment result is automatically generated, such as qualified / unqualified. The original image data that has passed authenticity verification and its corresponding image data package are automatically associated and linked to the inspection report as non-deletable attachments. After the above automatic filling and association are completed, a standardized online inspection report containing basic project information, automatically measured quantitative results, automatic comparison of results with standards, and tamper-proof image evidence is generated and awaits submission.

[0166] After the construction unit confirms the report content and clicks "Submit," it immediately triggers automatic verification using semantic analysis technology to check the report's descriptive text, required fields, and required attachments. If the verification fails, feedback is sent to the submitter in real time. If it passes, the report enters the approval pool. Based on the preset organizational structure and role-based access control model, the system automatically determines the approval process for the inspection report. For example, the rule could be that initial support inspection reports must first be sent to the professional supervising engineer, who will then review and approve them before sending them to the chief supervising engineer. The system matches the account of the professional supervising engineer responsible for the area with the inspection report.

[0167] The new inspection task will be automatically added to the target approver's to-do list based on role permissions. The system supports sorting by preset priority, for example, prioritizing critical processes or tasks that are about to expire. After sorting, the system will push notifications, such as sending a notification containing core information like the inspection number, project location, process type, submission time, and priority identifier to the approver's mobile and web-based systems.

[0168] Furthermore, to ensure timely response to approvals, multiple channels of notification are implemented, including: sending push notifications to the approver's mobile app, even if the app is running in the background; and highlighting the task in the "Pending Approvals" list on the web-based workbench logged in by the approver, possibly accompanied by in-app notifications or digital badges.

[0169] After receiving the notification on their terminal, the approver can directly access the details page of the inspection report by clicking the notification. They can view the automatically generated measurement data, system comparison results, and associated image data, and can also fill in their review comments and sign their electronic signature on the terminal to complete the online approval. The approval action will then trigger the workflow engine again, pushing the report to the next approver or archiving it.

[0170] In this embodiment, the completeness of the inspection and verification documents is automatically verified, and the documents are pushed to the supervisor's and owner's terminals according to their permissions, thereby improving approval efficiency. All data generated throughout the process, from the original images and analysis results to the approval opinions, is encrypted and uploaded to the blockchain in sequence for evidence storage, ultimately enabling traceability of the source of quality problems, verification of the process, and accountability.

[0171] Optionally, after generating a standardized online inspection report based on the actual size parameters of the target component, and pushing the inspection report to the terminal with corresponding permissions for approval according to preset rules, the method further includes:

[0172] The entire process data is processed in chronological order to generate encrypted hash values ​​and uploaded to the blockchain node, and a data access control mechanism is designed.

[0173] The "full-process data" refers to the collection of all key electronic data generated throughout the entire digital quality management process, from on-site image acquisition to final approval and archiving. This includes spatiotemporally bound images / videos, enhanced images, visualization reports, inspection applications, supervisor's approval opinions, rectification records, etc. It is a complete information chain recording the entire lifecycle of project quality. A cryptographic hash value is a fixed-length, unique alphanumeric sequence calculated using a cryptographic hash function. This calculation has a core characteristic: any slight change in the original data will produce a completely different hash value, and the original data cannot be derived from the hash value. A blockchain node refers to a single server or computer participating in the formation of the blockchain's distributed network. Each node stores a copy of the data block and maintains the consistency and immutability of network data through a consensus mechanism. A data access control mechanism refers to a set of rules pre-designed and implemented within the system that precisely manages data visibility and operational permissions based on user roles. This mechanism ensures that different roles (construction, supervision, owner, regulatory departments) can only view data corresponding to their permissions. For example, a supervisor can only view data for their assigned section, while a regulatory department may have broader query permissions.

[0174] Specifically, the entire process of data processing, including raw inspection data, preprocessing results, visualization reports, inspection applications, supervisor approval opinions, and rectification records, uses a cryptographic hash function (such as SHA-256) to calculate a unique encrypted hash value for each piece of data or each set of operation records. These hash values ​​are then uploaded sequentially to a pre-defined blockchain node network, strictly following the chronological order of the business data's generation. The blockchain network packages the uploaded hash values ​​into blocks linked by timestamps. Once uploaded to the chain, due to the distributed and encrypted nature of the blockchain, the hash value record cannot be tampered with or deleted, forming an unbreakable chain of evidence across time.

[0175] The system incorporates a data access control mechanism, ensuring that different roles (construction, supervision, owner, and regulatory authorities) can only view data within their authorized scope. Furthermore, all data operations (such as modification and approval) are logged in an unalterable manner. A one-click traceability function is also implemented, allowing for rapid retrieval of the entire process data by inputting the mileage marker, construction procedure, or inspection report number, thus enabling the tracing of quality issues back to their source.

[0176] In this embodiment, all process data, including testing data, inspection applications, and supervisory opinions, are generated into encrypted hash values ​​in chronological order and uploaded to the blockchain node. A data access control mechanism ensures that different roles can only view data corresponding to their permissions. All data operations are recorded as immutable blockchain operation logs. When traceability is required, by entering the mileage marker, construction procedure, or inspection number, the system can quickly retrieve the corresponding full-process hash record on the blockchain and, combined with the access verification mechanism, access the original data, enabling source tracing of quality issues.

[0177] Example 3:

[0178] Figure 9 This is a framework diagram of a visualized online detection and reporting system for the construction quality of concealed works in highway tunnels, provided in Embodiment 3 of the present invention. Figure 9 As shown, the system includes:

[0179] The acquisition module 210 is used to acquire construction quality image data and bind it with spatiotemporal information to generate image data packets with spatiotemporal correlation identifiers;

[0180] Verification module 220 is used to verify the authenticity of the image data packet;

[0181] Enhancement module 230 is used to perform quality enhancement processing on the verified image data in the image data packet using multi-exposure image fusion technology to generate a high dynamic range enhanced image;

[0182] The determination module 240 is used to determine the actual size parameters of the target component based on the high dynamic range enhanced image and through a dual-focus measurement model;

[0183] The generation module 250 is used to generate a standardized online inspection report based on the actual size parameters of the target component, and push the inspection report to the terminal with corresponding permissions for approval according to preset rules.

[0184] Optional, the acquisition module 210 is specifically used for:

[0185] A benchmark mapping table between tunnel mileage station numbers and actual three-dimensional coordinates is pre-established;

[0186] When collecting the construction quality image data, the current three-dimensional coordinates are obtained through a mobile terminal, the current tunnel mileage station is determined based on the benchmark mapping table, and a timestamp is generated simultaneously.

[0187] The tunnel mileage station number, the current three-dimensional coordinates, and the timestamp are combined to generate a unique associated identifier;

[0188] Using binary encapsulation technology, the unique association identifier is written into the custom metadata field of the construction quality image data to generate an image data packet with spatiotemporal association identifier.

[0189] Optional, the verification module 220 is specifically used for:

[0190] Based on the unique association identifier in the image data packet, the time and location of the image data packet are verified for compliance.

[0191] The image feature matching algorithm is used to detect the repetition and tampering of the construction quality image data in the image data packet.

[0192] Optional, enhancement module 230, specifically used for:

[0193] The verified single image in the image data packet is adaptively divided into brightness ranges to obtain multiple brightness ranges corresponding to the single image.

[0194] Calculate the gamma correction value for each brightness range, and perform gamma transformation processing on each single image based on the gamma correction value to generate pseudo-exposure images that present details in different brightness ranges;

[0195] A multi-resolution image pyramid technique is used to weight and fuse the pseudo-exposure images to generate a high dynamic range enhanced image.

[0196] Optionally, module 240 is defined, specifically for:

[0197] For the high dynamic range enhanced image, a scale-invariant feature transform algorithm is used to filter candidate images;

[0198] For each selected candidate image, key imaging parameters are extracted;

[0199] An optical calculation model for pinhole imaging is established based on the key imaging parameters corresponding to each candidate image.

[0200] In three-dimensional space, the optical axes of optical calculation models from different perspectives are unified to establish a standard bifocal measurement model;

[0201] Based on the standard bifocal measurement model, the actual dimensional parameters of the target component are calculated.

[0202] Optionally, module 250 is generated, specifically for:

[0203] Based on the type of concealed works to which the target component belongs, a preset inspection template is invoked, the actual size parameters of the target component are written into the preset inspection template, and a standardized online inspection report is generated;

[0204] The terminal with the corresponding permissions is determined according to preset rules, and the inspection report is pushed to the terminal with the corresponding permissions for approval.

[0205] Optionally, it also includes: an upload module, used for:

[0206] The entire process data is chronologically sequenced to generate encrypted hash values ​​and uploaded to the blockchain node, and a data access control mechanism is designed.

[0207] The online visualization detection and reporting system for the construction quality of concealed works in highway tunnels provided in this embodiment of the invention can execute the online visualization detection and reporting method for the construction quality of concealed works in highway tunnels provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0208] Example 4:

[0209] In some embodiments, the method for visual online detection and reporting of construction quality of concealed works in highway tunnels can be implemented as a computer program, which is tangibly contained in a computer-readable storage medium.

[0210] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0211] In the context of this invention, a computer-readable storage medium stores computer instructions that, when executed by a processor, implement the online visualization detection and reporting method for the construction quality of concealed works in highway tunnels provided by this invention. The computer-readable storage medium may be a tangible medium that may contain or store computer programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0212] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0213] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for online visualization and reporting of construction quality of concealed works in highway tunnels, characterized in that, include: Collect construction quality image data and bind it with spatiotemporal information to generate image data packets with spatiotemporal correlation identifiers; The authenticity of the image data packet is verified. Multi-exposure image fusion technology is used to enhance the quality of the verified image data in the image data packet, generating a high dynamic range enhanced image. Based on the high dynamic range enhanced image, the actual size parameters of the target component are determined using a dual-focus measurement model; Based on the actual size parameters of the target component, a standardized online inspection report is generated, and the inspection report is pushed to the terminal with the corresponding permissions for approval according to preset rules.

2. The method according to claim 1, characterized in that, The process of collecting construction quality image data and binding it with spatiotemporal information to generate image data packets with spatiotemporal correlation identifiers includes: A benchmark mapping table between tunnel mileage station numbers and actual three-dimensional coordinates is pre-established; When collecting the construction quality image data, the current three-dimensional coordinates are obtained through a mobile terminal, the current tunnel mileage station is determined based on the benchmark mapping table, and a timestamp is generated simultaneously. The tunnel mileage station number, the current three-dimensional coordinates, and the timestamp are combined to generate a unique associated identifier; Using binary encapsulation technology, the unique association identifier is written into the custom metadata field of the construction quality image data to generate an image data packet with spatiotemporal association identifier.

3. The method according to claim 1, characterized in that, The verification of the authenticity of the image data packet includes: Based on the unique association identifier in the image data packet, the time and location of the image data packet are verified for compliance. The image feature matching algorithm is used to detect the repetition and tampering of the construction quality image data in the image data packet.

4. The method according to claim 1, characterized in that, The method employs multi-exposure image fusion technology to enhance the quality of verified image data in the image data packet, generating a high dynamic range enhanced image, including: The verified single image in the image data packet is adaptively divided into brightness ranges to obtain multiple brightness ranges corresponding to the single image. Calculate the gamma correction value for each brightness range, and perform gamma transformation processing on each single image based on the gamma correction value to generate pseudo-exposure images that present details in different brightness ranges; A multi-resolution image pyramid technique is used to weight and fuse the pseudo-exposure images to generate a high dynamic range enhanced image.

5. The method according to claim 1, characterized in that, The process of determining the actual size parameters of the target component based on the high dynamic range enhanced image and using a dual-focus measurement model includes: For the high dynamic range enhanced image, a scale-invariant feature transform algorithm is used to filter candidate images; For each selected candidate image, key imaging parameters are extracted; An optical calculation model for pinhole imaging is established based on the key imaging parameters corresponding to each candidate image. In three-dimensional space, the optical axes of optical calculation models from different perspectives are unified to establish a standard bifocal measurement model; Based on the standard bifocal measurement model, the actual dimensional parameters of the target component are calculated.

6. The method according to claim 1, characterized in that, Based on the actual dimensional parameters of the target component, a standardized online inspection report is generated, and the inspection report is pushed to the terminal with corresponding permissions for approval according to preset rules, including: Based on the type of concealed works to which the target component belongs, a preset inspection template is invoked, the actual size parameters of the target component are written into the preset inspection template, and a standardized online inspection report is generated; The terminal with the corresponding permissions is determined according to preset rules, and the inspection report is pushed to the terminal with the corresponding permissions for approval.

7. The method according to claim 1, characterized in that, After generating a standardized online inspection report based on the actual size parameters of the target component, and pushing the inspection report to the terminal with corresponding permissions for approval according to preset rules, the process further includes: The entire process data is chronologically sequenced to generate encrypted hash values ​​and uploaded to the blockchain node, and a data access control mechanism is designed.

8. A visualized online detection and reporting system for the construction quality of concealed works in highway tunnels, characterized in that, The system is used to execute the online visualization detection and reporting method for the construction quality of concealed works in highway tunnels as described in any one of claims 1-7, including: The acquisition module is used to acquire construction quality image data and bind it with spatiotemporal information to generate image data packets with spatiotemporal correlation identifiers; The verification module is used to verify the authenticity of the image data packet; The enhancement module is used to perform quality enhancement processing on the verified image data in the image data packet using multi-exposure image fusion technology to generate a high dynamic range enhanced image; The determination module is used to determine the actual size parameters of the target component based on the high dynamic range enhanced image and through a dual-focus measurement model; The generation module is used to generate a standardized online inspection report based on the actual size parameters of the target component, and push the inspection report to the terminal with corresponding permissions for approval according to preset rules.

9. The system according to claim 8, characterized in that, The enhancement module is specifically used for: The input single image is adaptively divided into multiple brightness ranges based on its brightness distribution; Calculate the gamma correction value for each brightness range, and perform gamma transformation processing on the image based on each gamma correction value to generate a pseudo-exposure image that presents details in different brightness ranges; A multi-resolution image pyramid technique is used to weight and fuse the pseudo-exposure images to generate a high dynamic range enhanced image.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the method for visual online detection and reporting of the construction quality of concealed works in highway tunnels as described in any one of claims 1-7.

Images