Dynamic sounding detection data management method, system, equipment and medium

Abstract

The invention relates to a dynamic sounding detection data management method, system and device and a medium, and relates to the technical field of civil engineering, and the method comprises the steps: receiving sounding data and a field picture uploaded by a construction area terminal through a CDN node, and embedding the detection data into picture metadata; identifying a key region of the picture by using a pre-trained CNN, implementing multi-resolution hierarchical compression, generating a thumbnail set, and mapping the thumbnail set with detection data; a cross-modal index database is established based on geographic coordinates and detection parameters, and retrieval according to multi-dimensional conditions such as coordinates, N values and soil layer types is supported; and when a terminal request is responded, a resolution demand is dynamically matched, the associated thumbnail is accurately transmitted, and efficient management and low-bandwidth transmission of the investigation data are realized. According to the invention, the data recording efficiency and accuracy of the dynamic penetrometer can be improved in a construction site with poor network conditions.

Description

Technical Field

[0001] The present application relates to the field of civil engineering technology, and in particular to a method, system, equipment and medium for managing dynamic penetration detection data. Background Art

[0002] In the civil engineering inspection industry, with the continuous advancement of infrastructure construction and increasing demands for project quality, accurate management of inspection data has become increasingly important. Inspection data not only contributes to project quality assessment but also provides a crucial basis for subsequent maintenance and improvement. While traditional data management can record and preserve information to a certain extent, its limitations are becoming increasingly apparent when faced with large-scale and complex engineering projects. Effective data management can provide reliable support for engineering decision-making and ensure project safety and stability. However, the current data management situation is unable to meet the needs of industry development.

[0003] Currently, in the field of civil engineering inspection, the industry primarily uses two methods to manage inspection data. On the one hand, there's the traditional manual record-keeping and paper reporting method. Workers manually record inspection data at the construction site and then compile it into paper reports. This method was widely used for a certain period of time and is relatively simple to operate, requiring no complex equipment or technical support. On the other hand, there's the desktop software-based data management system. It enables electronic storage and preliminary processing of data, representing a significant improvement over manual record-keeping and paper reporting. However, this system typically requires running on fixed computer equipment, lacking flexibility and mobility. Furthermore, the data transfer process to the back-end processing system is cumbersome and requires multiple steps.

[0004] Existing technologies have numerous drawbacks. Traditional manual record-keeping and paper-based reporting methods are inconvenient for data entry, prone to human error, and impacting data accuracy. Desktop-based data management systems lack mobility, preventing immediate data entry at the construction site. The data transmission process is cumbersome and leads to data transmission delays. Furthermore, there are issues such as inconsistent data formats and insufficient data security. Data upload speeds are limited, especially on remote sites or those with poor network conditions, severely impacting work efficiency and the overall inspection process. Summary of the Invention

[0005] The first purpose of this application is to provide a dynamic penetration detection data management method, which can improve the data recording efficiency and accuracy of the dynamic penetration instrument in construction sites with poor network conditions.

[0006] In a first aspect, the present application provides a method for managing dynamic penetration testing data, which adopts the following technical solutions: A method for managing dynamic penetration testing data, comprising: Receive test data and several associated site photos containing a dynamic penetration tester uploaded by mobile terminals in the construction area through CDN edge nodes, and embed test data into the metadata of each site photo. The test data includes the number of hits (N), penetration depth, and soil layer number of the dynamic penetration tester. Using a pre-trained convolutional neural network to identify key and non-key areas in several scene images, performing multi-resolution hierarchical compression processing on each scene image, wherein lossless compression is used for the key areas and adaptive block lossy compression is used for the non-key areas. The images are then combined to generate a set of thumbnails with different resolutions, and a mapping relationship between the thumbnails and the detection data is established; Building an index database based on the coordinate identifiers of the dynamic penetration instrument detection points and a thumbnail set corresponding to the detection data, wherein the index database supports cross-modal retrieval; In response to the query instruction of the receiving terminal and the resolution requirement for the thumbnail, a thumbnail of the required resolution is obtained from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of each dynamic probe, and the thumbnail is transmitted to the receiving terminal.

[0007] By adopting the above technical solution, CDN edge nodes are used to receive data and images uploaded by mobile terminals, which ensures stable data reception even in poor network environments. By embedding detection data in image metadata, a close association is achieved between detection data and images, facilitating subsequent queries to obtain thumbnails of different resolutions. Different parts of the on-site images are compressed in a hierarchical manner to generate thumbnail sets of different resolutions, which can not only ensure the integrity of key area data but also reduce storage costs. An index database supporting cross-modal retrieval is constructed based on the coordinate identification and detection data of the dynamic probe, which can accurately and quickly obtain thumbnails of the required resolution based on multi-dimensional conditions, thereby improving the efficiency and accuracy of data queries for the dynamic probe.

[0008] In a preferred example, the present application may be further configured as follows: before the step of receiving, via a CDN edge node, the detection data uploaded by a mobile terminal located in the construction area and a plurality of associated on-site pictures including a dynamic penetration instrument, and embedding the detection data into the metadata of each on-site picture, wherein the detection data includes the number of hits N, the penetration depth, and the soil layer number of the dynamic penetration instrument, the step further includes: Obtaining a video clip containing the detection point of the dynamic penetration instrument through a CDN edge node; Based on the inter-frame difference method, the number of hammer determinations of the dynamic penetration instrument in the video clip is counted, and the video clip is intercepted according to the time node of the statistical number of hammer determinations to obtain an on-site picture, and watermark data of the number of hammer determinations N, the penetration depth and the soil layer number are attached to the preset watermark area of ​​the on-site picture; Optical character recognition is performed on a preset watermark area of ​​the scene picture, and the recognized watermark data is compared with the detection data in the metadata of the scene picture. If there is a conflict between the watermark data and the detection data in the metadata, an abnormal marking signal is triggered.

[0009] By adopting the above technical solution, CDN edge nodes are used to obtain video clips, which improves the transmission efficiency of video clips at construction sites with poor network conditions. On-site pictures are captured based on the inter-frame difference method to record the number of hammer strikes. At the same time, watermark data is attached to the on-site pictures. This can enhance the correlation between the detection data and the on-site conditions, and improve the authenticity and traceability of the data. By comparing the watermark data and metadata and triggering an abnormal marking signal when a conflict occurs, data errors can be discovered in a timely manner to ensure the accuracy of the detection data.

[0010] In a preferred example, the present application can be further configured as follows: identifying key areas and non-key areas in a plurality of scene images using a pre-trained convolutional neural network, performing multi-resolution hierarchical compression processing on each scene image, wherein lossless compression is used for the key areas and adaptive block lossy compression is used for the non-key areas, merging and generating a plurality of thumbnail sets of different resolutions, and establishing a mapping relationship between the plurality of thumbnails and the detection data, including the following steps: The pre-trained convolutional neural network is used to identify the key areas and non-key areas of the dynamic probe in the on-site pictures, where the key areas include the core hammer and scale lines of the probe rod of the dynamic probe, and the non-key areas are the remaining areas except the key areas.

[0011] By adopting the above technical solution, the scale lines of the core hammer and probe rod of the dynamic probe are taken as key areas and losslessly compressed, which can accurately record the number of blows N and penetration depth that need to be detected by the dynamic probe, while other non-critical areas are losslessly compressed through adaptive block segmentation, and then different thumbnails and sets of thumbnails with different resolutions are generated for the key area part, and a mapping relationship between thumbnails and detection data is established, which improves the accuracy of key information while reducing storage costs and facilitates the search for corresponding thumbnails based on detection data.

[0012] In a preferred example, the present application may be further configured as follows: the step of constructing an index database based on the coordinate identifiers of the detection points of the dynamic penetration instrument and the thumbnail set corresponding to the detection data, wherein the index database supports cross-modal retrieval, includes: The GPS coordinate identification, detection time and detection data of the dynamic penetration instrument detection point are used as index dimensions to establish the index database, which supports retrieval based on the combination conditions of any data in the detection data; The video clips whose shooting range includes the GPS coordinates of the detection points of the dynamic probe are synchronously associated with the on-site pictures containing the dynamic probe to form an evidence chain.

[0013] By adopting the above technical solution, an index database is established with the GPS coordinate identification, detection time and detection data of the dynamic probe detection point as index dimensions, which can support retrieval based on the combination conditions of any data in the detection data, thereby improving the flexibility and convenience of data retrieval; the video clips containing the GPS coordinates of the dynamic probe detection point in the shooting range are synchronously associated with the on-site pictures containing the dynamic probe to form an evidence chain, thereby enhancing the authenticity of the detection data and facilitating traceability.

[0014] In a preferred example, the present application may be further configured as follows: after the step of synchronously associating the video clips including the GPS coordinates of the detection points of the dynamic penetration instrument with the on-site pictures including the dynamic penetration instrument to form an evidence chain, the application may further include: The camera device monitors the gradient change rate of the number of hits N of the dynamic penetration instrument in real time, and automatically retrieves the detection data and penetration process video associated with the dynamic penetration instrument when the difference between adjacent number of hits N exceeds the threshold value; Generate a spatiotemporal thermal diagram of the dynamic probe to mark abnormal areas, and support the layer-by-layer development of related evidence according to the chain of evidence.

[0015] By adopting the above technical solution, the gradient change rate of the number of hits N value is monitored in real time, and the abnormal situation where the difference between adjacent number of hits N values ​​exceeds the threshold can be discovered in time. The automatic retrieval of related detection data and penetration process video is helpful for rapid problem detection; the generation of spatiotemporal thermal diagrams to mark abnormal areas can intuitively display the abnormal situation of the number of hits N value of the dynamic probe; at the same time, it supports the layer-by-layer expansion of related evidence according to the chain of evidence, which can deeply trace the cause of the abnormality and enhance the traceability and reliability of the detection data.

[0016] In a preferred example, the present application may be further configured as follows: in response to the query instruction of the receiving terminal and the resolution requirement of the thumbnail, the step of obtaining a thumbnail of the required resolution from the index database according to the coordinate identification, the number of hits N value and the multi-dimensional conditions of the soil layer type of each dynamic penetration instrument, and transmitting the thumbnail to the receiving terminal also includes: In response to the query instruction of the receiving terminal, data is filtered from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of the query instruction, and a test report is quickly generated according to a preset report template.

[0017] By adopting the above technical solution, it is possible to filter data from the index database based on the multi-dimensional conditions of the receiving terminal query instructions, and then quickly generate inspection reports using preset report templates, unify the report format, save time for manual report compilation, and improve the data processing efficiency of the dynamic probe.

[0018] In a preferred example, the present application may be further configured as follows: the step of receiving, via a CDN edge node, detection data uploaded by a mobile terminal located in a construction area and a plurality of associated on-site pictures including a dynamic penetration instrument, and embedding detection data in metadata of each on-site picture, wherein the detection data includes the number of hits N, penetration depth, and soil layer number of the dynamic penetration instrument, further comprising: Real-time evaluation of the transmission performance of each CDN node and dynamic selection of the optimal node to establish a data channel; When the network is interrupted, the check value of the inspection data transmitted by the mobile terminal is recorded, and after the network is restored, the breakpoint resume is performed, and a timestamp mark is added to the data packet of the inspection data.

[0019] By adopting the above technical solution, the transmission performance of each CDN node is evaluated in real time and the optimal node is dynamically selected to establish a data channel. This can optimize the data transmission path, so that on-site pictures carrying inspection data can be stably and efficiently uploaded from mobile terminals to CDN edge nodes. When the network is interrupted, the checksum value of the transmitted inspection data is recorded and the breakpoint is resumed after the network is restored. At the same time, a timestamp is added to the data packet to minimize data loss and improve the reliability of data transmission. Even in construction locations with unstable networks, data transmission can be successfully completed.

[0020] In a second aspect, the present application provides a dynamic penetration test data management system, which adopts the following technical solutions: A dynamic penetration detection data management system, comprising: Image acquisition module: used to receive test data and several associated site photos containing a dynamic penetration tester uploaded by mobile terminals in the construction area through CDN edge nodes, and embed test data into the metadata of each site photo. The test data includes the number of hits (N), penetration depth, and soil layer number of the dynamic penetration tester; Image thumbnail module: used to identify key areas and non-key areas in a number of scene images through a pre-trained convolutional neural network, perform multi-resolution hierarchical compression processing on each scene image, wherein lossless compression is used for the key areas and adaptive block lossy compression is used for the non-key areas, and merge to generate a number of thumbnail sets with different resolutions, and establish a mapping relationship between the thumbnail sets and the detection data; Database construction module: used to construct an index database based on the coordinate identifiers of the dynamic penetration instrument detection points and the thumbnail collection corresponding to the detection data, wherein the index database supports cross-modal retrieval; Thumbnail transmission module: used to respond to the query instructions of the receiving terminal and the resolution requirements for the thumbnail, obtain the thumbnail of the required resolution from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of each dynamic probe, and transmit the thumbnail to the receiving terminal.

[0021] In a third aspect, the present application provides an electronic device, which adopts the following technical solution: An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the above-mentioned method for managing dynamic penetration detection data are implemented.

[0022] In a fourth aspect, the present application provides a computer storage medium, including the following technical solutions: A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for managing dynamic penetration detection data are implemented.

[0023] In summary, this application has the following beneficial technical effects: This application is based on the CDN edge node receiving the detection data uploaded by the mobile terminal, and the on-site pictures captured and uploaded at the time node determined by the number of hammer strikes from the on-site video clips, and dividing the non-critical areas in the on-site pictures for lossy compression, while using lossless compression to retain the details of the core hammer and probe rod scale lines of the dynamic probe. Even in a construction site with poor network conditions, the key detection data of the dynamic probe can be obtained, thereby improving the data recording efficiency and accuracy; the on-site pictures are hierarchically compressed to generate thumbnail sets of different resolutions and an index database supporting cross-modal retrieval is constructed. The required thumbnails can be quickly provided and the detection report can be generated according to the query instructions, which unifies the report format, saves report compilation time, and improves data processing efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a flow chart of a method for managing dynamic penetration detection data in one embodiment of the present application.

[0025] Figure 2 This is a flowchart of the steps added before step S1 in one embodiment of the present application.

[0026] Figure 3 This is a flowchart of the sub-steps of step S2 in one embodiment of the present application.

[0027] Figure 4 This is a flowchart of the sub-steps of step S3 in one embodiment of the present application.

[0028] Figure 5 This is a flowchart of the steps added after step S31 in one embodiment of the present application.

[0029] Figure 6 This is a flowchart of the sub-steps of step S4 in one embodiment of the present application.

[0030] Figure 7 This is a flowchart of the sub-steps of step S1 in one embodiment of the present application.

[0031] Figure 8 It is a structural diagram of a dynamic penetration detection data management system according to one embodiment of the present application.

[0032] Figure 9 It is a principle block diagram of an electronic device in one embodiment of the present application.

[0033] Figure numerals: 1. Picture acquisition module; 2. Picture thumbnail module; 3. Database construction module; 4. Thumbnail transmission module. DETAILED DESCRIPTION

[0034] The following is combined with Figure 1-9 This application is described in further detail.

[0035] It should be noted that all actions of obtaining data or information or data in this application are carried out in compliance with the relevant data protection laws and policies of the country where they are located and with the authorization of the corresponding users.

[0036] refer to Figure 1 , a dynamic penetration detection data management method, specifically comprising: S1. Receive test data and several associated on-site photos containing a dynamic penetration tester uploaded by mobile terminals in the construction area through CDN edge nodes, and embed test data into the metadata of each on-site photo. The test data includes the number of hits (N), penetration depth, and soil layer number of the dynamic penetration tester.

[0037] Specifically, the dynamic penetration instrument includes a core hammer and a probe rod. A hammer seat is provided at the top of the probe rod. The core hammer slides through the part of the probe rod located at the top of the hammer seat. A probe is provided at the bottom of the probe rod. The probe directly contacts the soil. The core hammer slides on the top of the probe rod and drops vertically, directly impacting the top of the hammer seat, so that the probe inserted into the soil penetrates into the soil. Each time the core hammer falls and impacts the top of the hammer seat, it is considered to have completed one hammer strike. When a dynamic penetration instrument is used to conduct a soil survey in the same soil layer number, when one hammer strike is completed, the number of strikes N is set to 1. When two hammer strikes are completed, the number of strikes N is set to 2, and so on. At the same time, the probe rod at the top of the probe is vertically provided with a scale line for recording the penetration depth of the probe rod.

[0038] By deploying CDN edge nodes at the construction site, we can receive the on-site pictures uploaded by mobile terminals, including the number of hits N, penetration depth, soil layer number and related pictures, and directly embed the structured detection data of the dynamic penetration instrument in the metadata of the on-site pictures. We can use the CDN network to shorten the data transmission path and reduce the delay caused by network fluctuations, thereby achieving local transmission and strong binding of data and images, solving the problems of low data upload efficiency and separation of data and images in weak network environments, improving the real-time data synchronization capability and integrity, making the key detection parameters correspond to the image information one by one, and avoiding manual recording errors as much as possible.

[0039] S2. Identify key areas and non-key areas in several scene pictures through pre-trained convolutional neural networks, perform multi-resolution hierarchical compression processing on each scene picture, using lossless compression for key areas and adaptive block lossy compression for non-key areas, and merge to generate several thumbnail sets with different resolutions, and establish a mapping relationship between several thumbnails and detection data.

[0040] Specifically, lossless compression is used to retain detail accuracy in key areas, and non-key areas are adaptively divided into blocks and lossy compressed according to texture complexity. A multi-resolution thumbnail set is generated, and a mapping relationship between the thumbnails and the detection data is established. A differentiated compression strategy is implemented to balance the losslessness of key information and the high compression rate of non-key areas, reduce the storage volume of a single image, and improve the efficiency of image transmission at construction sites with poor network conditions.

[0041] Furthermore, in order to generate thumbnails of the target resolution, the compressed full image, including the key areas of lossless compression and the non-key areas of lossy compression, is first normalized and scaled or cropped, and then adjusted to the target resolution using an anti-aliasing algorithm. This ensures that the clarity of the key areas is not affected by scaling, while the compression distortion of the non-key areas can be effectively weakened due to the smaller thumbnail size. Ultimately, a set of thumbnails with uniform resolution and taking into account the integrity of key information is obtained.

[0042] S3. Based on the coordinate identifiers of the dynamic penetration instrument detection points and the thumbnail collection corresponding to the detection data, an index database is constructed, and the index database supports cross-modal retrieval.

[0043] Furthermore, cross-modal retrieval technology enables hybrid queries involving text, numerical values, and images. Through structured indexing and multimodal association, it overcomes the limitations of traditional databases' single retrieval modes, enabling rapid cross-matching of geolocation, detection parameters, and image data. This reduces retrieval response times to seconds, significantly improving the accuracy and efficiency of data queries under complex conditions.

[0044] S4. In response to the query instruction of the receiving terminal and the resolution requirement for the thumbnail, a thumbnail of the required resolution is obtained from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of each dynamic probe, and the thumbnail is transmitted to the receiving terminal.

[0045] Specifically, the system dynamically selects the appropriate thumbnail version based on the receiving terminal's network status and resolution requirements. For example, low-resolution images are prioritized in weak network environments. Furthermore, the system selects thumbnails of the target resolution based on multiple criteria, such as the dynamic penetrometer's coordinate identifier, the number of hits (N), and soil layer type. This effectively alleviates transmission congestion in weak network environments while ensuring the readability of critical information. This allows users operating receiving terminals to quickly obtain thumbnails of the desired resolution, reducing resource consumption on both the cloud and the terminal.

[0046] refer to Figure 2 Furthermore, in one embodiment, before step S1, steps S10, S11, and S12 are added: S10. Obtain a video clip containing a detection point of the dynamic penetration instrument through a CDN edge node.

[0047] Specifically, video clips of the construction site's dynamic penetration instrument inspection process are obtained in real time through CDN edge nodes, and the CDN network is used to receive and cache video stream data nearby, reducing network delays and packet loss risks in long-distance transmission. This step quickly collects original video data through edge nodes, providing high-timeliness input for subsequent statistics of the number of hammer strikes and the capture of on-site pictures, minimizing video acquisition interruptions due to network fluctuations, and making the entire inspection process of each dynamic penetration instrument traceable.

[0048] S11. Based on the inter-frame difference method, the number of hammer judgments of the dynamic probe in the video clip is counted, and the video clip is intercepted according to the time node of the statistical number of hammer judgments to obtain the on-site picture, and the watermark data of the number of hammer judgments N, penetration depth and soil layer number are attached to the preset watermark area of ​​the on-site picture.

[0049] Specifically, a plaintext watermark containing the real-time blow count (N), penetration depth, and soil layer number recorded during the test period is embedded in the preset watermark area of ​​the image. Using the inter-frame difference method, the system analyzes the changes in consecutive frames of the core hammer's hammering action in the video clip. When the bottom of the core hammer lands on the top of the hammer seat, the system automatically counts the number of effective hammer blows and accurately captures the key frame at the time of the hammering completion to generate the on-site image.

[0050] At the same time, the keyframe recorded at the time of the hammering completion uses the built-in three-axis accelerometer and gyroscope data of the camera to obtain the device's pitch, roll, and yaw angles in real time, establishing a mapping relationship between the device's posture and the image coordinate system. Combined with the preset standardized markers on the probe surface and their positions in the image, and using perspective transformation models such as the Homography matrix, the probe image from the oblique perspective is corrected to a vertical projection to eliminate geometric distortion. The penetration depth is also accurately calculated based on the actual physical spacing of the markers and the corrected pixel ratio.

[0051] When the penetration depth reaches the preset standard detection depth, the number of key frames recorded at this penetration depth is counted and recorded as the number of hammer strikes in this section. For example, a light hammer (N10) records a hammer strike every 30 cm of penetration, and a heavy hammer (N63.5) records a hammer strike every 10 cm of penetration. By automatically determining the time nodes of hammering, replacing manual counting and screenshots, the objectivity and timeliness of hammering data records are improved. The direct superposition of watermark data enhances the intuitive correlation between images and detection data, reduces human operation errors, and improves the robustness of automated detection in complex construction scenarios.

[0052] S12. Perform optical character recognition on the preset watermark area of ​​the scene picture, compare the recognized watermark data with the detection data in the metadata of the scene picture, and trigger an abnormal marking signal if there is a conflict between the watermark data and the detection data in the metadata.

[0053] Specifically, optical character recognition (OCR) is performed on the text in the watermark area of ​​the image, extracting the watermark data and automatically comparing it with the detection value in the metadata. If any inconsistency is found in the number of hits, depth, or soil layer number, an anomaly flag is triggered and a manual review is notified. This dual data verification mechanism improves the authenticity and consistency of the test results, minimizing data distortion caused by transmission errors or human tampering. Automated anomaly detection also reduces manual verification costs and improves the credibility and integrity of the test data.

[0054] For example, during a dynamic penetration test, a tester captured a site image. The watermark area indicated the number of penetrations (N) as "15," the penetration depth as "3.2m," and the soil layer number as "③." After extracting the watermark data through OCR recognition, it was automatically compared with the test values ​​in the metadata (N=15, Depth=3.2m, Soil Layer=③). If the results were consistent, the data would be considered normal. If the image watermark indicated N=15, Depth=3.2m, Soil Layer=③, but a transmission error caused the soil layer number to be recorded as "②," the inconsistency would be detected through comparison, triggering an anomaly flag signal immediately. An alarm log would be generated and pushed to the management backend, prompting a manual review of the site image transmission process and test records.

[0055] In addition, reference Figure 3 Furthermore, in one embodiment, step S2 is further divided into the following sub-steps: S20. Identify the key areas and non-key areas of the dynamic penetration instrument in the on-site picture through a pre-trained convolutional neural network, where the key areas include the penetration hammer and the scale lines of the probe rod of the dynamic penetration instrument, and the non-key areas are the remaining areas except the key areas.

[0056] Specifically, semantic segmentation is performed on the acquired on-site images using a pre-trained convolutional neural network (CNN). Based on the structural features of the core hammer and the scale lines of the probe rod of the dynamic penetrator, such as the geometric shape of the core hammer and the texture patterns of the scale lines of the probe rod, the core hammer and scale lines as key areas are accurately identified, as well as the background environment and unrelated equipment as non-key areas. Transfer learning is then used to optimize the model to enhance the generalization capability of the complex lighting and occlusion scenarios of the soil corresponding to each soil layer number at the construction site. Automatic identification of key areas replaces manual labeling, enabling rapid positioning of the core hammer action and scale line data, minimizing regional division errors caused by manual misjudgment, and providing a precise basis for target region division for subsequent hierarchical compression. This reduces the time spent on redundant data processing, improves data preprocessing efficiency and the accuracy of key information extraction, and adapts to the needs of efficient data stream transmission at construction sites when network conditions are poor.

[0057] In addition, reference Figure 4 Furthermore, in one embodiment, step S3 is further divided into the following sub-steps: S30. Use the GPS coordinate identifier, detection time and detection data of the dynamic probe detection point as index dimensions to establish an index database, supporting combined condition retrieval based on any data in the detection data.

[0058] Specifically, a structured database is constructed with the GPS coordinates, detection time, number of hits N, penetration depth and soil layer number of the dynamic probe detection point as the core index dimensions. Through multi-dimensional labeling storage of spatiotemporal labels and detection parameters, it supports rapid retrieval by any combination of conditions such as coordinate range, time interval, soil layer type or number of hits threshold, thereby replacing the traditional single keyword retrieval mode with a structured index, solving the problem of low efficiency of complex queries in weak network environments at construction sites, shortening data retrieval response time, and improving the accuracy of cross-dimensional data correlation analysis, meeting the needs of construction sites for rapid query of any detection data of the dynamic probe.

[0059] S31. Synchronously associate the video clips whose shooting range includes the GPS coordinates of the detection points of the dynamic probe with the on-site pictures containing the dynamic probe to form an evidence chain.

[0060] Specifically, the video clips recording the hammering process under the GPS coordinates of the same dynamic probe detection point are temporally and spatially associated with the on-site pictures recording the key frames and watermark data, and a complete chain of evidence is established by matching the coordinates and timestamps. This supports tracing back the detection process from the video and verifying the detection results from the pictures, realizing cross-verification of the video and the pictures, solving the problem of insufficient credibility of the traditional single data source, enhancing the traceability and anti-dispute ability of the detection data, and providing multimodal evidence support for the review of abnormal data. For example, the authenticity of the picture watermark data can be verified through video, reducing the risk of misjudgment due to data isolation, and improving the efficiency of data integrity management in complex scenarios.

[0061] In addition, reference Figure 5 Furthermore, in one embodiment, after step S31, steps S310 and S311 are added: S310. Monitor the gradient change rate of the number of hits N of the dynamic probe in real time through a shooting device. When the difference between adjacent number of hits N exceeds the threshold value of the number of hits, automatically retrieve the detection data and penetration process video of the associated dynamic probe.

[0062] Specifically, the system collects the N-value sequence of the dynamic penetration instrument's hit counts in real time and calculates the gradient change rate between adjacent test points. Through real-time gradient monitoring and dynamic threshold determination, it quickly identifies sudden changes in soil hardness or abnormal equipment operating conditions, reducing blind spots in manual inspections. This allows abnormal data to be captured immediately and correlated with multimodal evidence, improving the timeliness and accuracy of data anomaly analysis and avoiding safety hazards caused by missed inspections.

[0063] In this embodiment, the gradient change rate δ = ΔN / Δd, where Δd is the preset standard testing depth. If the δ value remains stable within a continuous testing section, such as a ΔN fluctuation within ±5 blows per 30 cm in a light penetration test, it indicates that the soil density or strength distribution is uniform. If the δ value suddenly changes, such as a sudden increase or decrease in ΔN in a certain section, it may reflect the presence of interlayers in the soil layer, such as a hard gravel layer or a soft silt layer. For example, in a heavy penetration test, if ΔN suddenly increases from 5 blows to 20 blows in a 10 cm section (δ = 15 blows / 10 cm), it may indicate the presence of a dense sand layer; if ΔN decreases from 20 blows to 5 blows (δ = -15 blows / 10 cm), it may indicate a void or loose backfill.

[0064] S311. Generate a spatiotemporal thermal diagram of the dynamic penetration instrument to mark abnormal areas, and support the layer-by-layer development of related evidence based on the chain of evidence.

[0065] Specifically, a spatiotemporal heat map is generated based on the GPS coordinates, timestamps, and hit count N gradient data of the detection points. In this embodiment, the spatiotemporal heat map is constructed with time / depth on the horizontal axis and hit count gradient on the vertical axis. The color maps indicate the degree of anomaly, anomaly mutation areas are marked, and the map supports layer-by-layer expansion of the chain of evidence from the heat map nodes. This improves the efficiency of problem diagnosis and decision-making reliability under complex geological conditions, while also providing data support for subsequent report generation and engineering rectification.

[0066] In addition, reference Figure 6 Furthermore, in one embodiment, step S4 is further divided into the following sub-steps: S40. In response to the query instruction of the receiving terminal, data is filtered from the index database according to the coordinate identification, the number of hits N value and the multi-dimensional conditions of the soil layer type of the query instruction, and a test report is quickly generated according to a preset report template.

[0067] Specifically, by pre-defining multi-dimensional screening rules, which are set as coordinate range, hit count N value interval, and soil layer type in this embodiment, matching test data can be quickly extracted from the index database, such as the test points with N value > 20 and soil layer ② in the coordinates of a certain area. The key parameters are automatically filled in with the standardized report template to generate a structured test report, thereby replacing the process of manual data collation and report writing, solving the problem of time-consuming report generation and chaotic format in weak network environments, and meeting the needs of instant reporting on the construction site.

[0068] In addition, reference Figure 7 Furthermore, in one embodiment, step S1 is further divided into the following sub-steps: S13: Evaluate the transmission performance of each CDN node in real time and dynamically select the optimal node to establish a data channel.

[0069] Specifically, by real-time monitoring of the network performance indicators of each CDN node, such as latency, bandwidth, and packet loss rate, the node availability is dynamically evaluated and the node with the lowest latency and best stability is selected as the data transmission channel; when the network fluctuates or the node is congested, it automatically switches to the backup node to ensure continuous data upload, improve the success rate of data transmission on the construction site, and ensure that the detection data is synchronized to the cloud in real time, so as to minimize data retention or loss caused by improper node selection.

[0070] S14. When the network is interrupted, the check value of the inspection data transmitted by the mobile terminal is recorded, and after the network is restored, the breakpoint transmission is resumed, and a timestamp is added to the data packet of the inspection data.

[0071] Specifically, when the network is interrupted, the checksum value and breakpoint position of the transmitted data packet are recorded, and a timestamp mark is added to the data packet to be transmitted. After the network is restored, the integrity of the transmitted data is verified based on the checksum value, and only the unfinished part is retransmitted, and repeated transmission is avoided as much as possible. This solves the problem of damaged data integrity or repeated transmission in scenarios with frequent network interruptions, ensures that the data packet is uploaded completely in time sequence, and improves the recording reliability and transmission efficiency of construction inspection data.

[0072] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0073] The embodiment of the present application also provides a dynamic penetration detection data management system, which corresponds one-to-one to the dynamic penetration detection data management method in the embodiment.

[0074] refer to Figure 8 A dynamic penetration test data management system includes: an image acquisition module 1, an image thumbnail module 2, a database construction module 3, and a thumbnail transmission module 4. The functional modules are described in detail as follows: Image acquisition module 1: used to receive test data and several associated site photos containing a dynamic penetration tester uploaded by mobile terminals in the construction area through CDN edge nodes, and embed the test data into the metadata of each site photo. The test data includes the number of hits (N), penetration depth, and soil layer number of the dynamic penetration tester; Image thumbnail module 2: Used to identify key and non-key areas in several scene images through a pre-trained convolutional neural network, perform multi-resolution hierarchical compression processing on each scene image, using lossless compression for key areas and adaptive block lossy compression for non-key areas, and merge to generate several sets of thumbnails with different resolutions, and establish a mapping relationship between the thumbnails and the detection data; Database construction module 3: used to construct an index database based on the coordinate identifiers of the dynamic penetration instrument detection points and the thumbnail collection corresponding to the detection data. The index database supports cross-modal retrieval; Thumbnail transmission module 4: used to respond to the query instructions of the receiving terminal and the resolution requirements for the thumbnail, obtain the thumbnail of the required resolution from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of each dynamic probe, and transmit the thumbnail to the receiving terminal.

[0075] Among them, the image acquisition module 1 uses the CDN edge node to receive the detection data and on-site images uploaded by the mobile terminal and embeds them into the detection data, which can improve the data recording efficiency and accuracy of the dynamic probe. Even in an environment with poor network conditions, the CDN can ensure the timeliness of data reception. The image thumbnail module 2 can improve the efficiency of on-site image processing and reduce storage costs by identifying the key and non-key areas of the on-site image and performing hierarchical compression processing, generating thumbnail sets and establishing mapping relationships. The database construction module 3 builds an index database that supports cross-modal retrieval based on the coordinate identification and thumbnail set of the dynamic probe, facilitating data management and retrieval. The thumbnail transmission module 4 obtains and transmits thumbnails according to the query instructions and resolution requirements of the receiving terminal, meeting the user's query needs for data and improving the user experience. Through the combination of various modules, the data recording efficiency and accuracy of the dynamic probe can be improved in construction sites with poor network conditions.

[0076] For the specific definition of the dynamic penetration detection data management system, please refer to the definition of the dynamic penetration detection data management method in the context, which will not be repeated here. Each module in the above-mentioned dynamic penetration detection data management system can be implemented in whole or in part by software, hardware and their combination. The above-mentioned modules can be embedded in or independent of the processor in the electronic device in the form of hardware, or can be stored in the memory of the electronic device in the form of software, so that the processor can call and execute the operations corresponding to the above modules. In one embodiment, an electronic device is provided, which is a user terminal. Reference Figure 9 The electronic device includes a processor, a memory, a network interface, and a database connected via a system bus. The processor of the electronic device is used to provide computing and control capabilities. The memory of the electronic device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the electronic device is used to store a detection data table. The network interface of the electronic device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for managing dynamic probing detection data is implemented.

[0077] In one embodiment, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed: S1. Receive test data and several associated on-site photos containing a dynamic penetration tester uploaded by mobile terminals in the construction area through CDN edge nodes, and embed test data into the metadata of each on-site photo. The test data includes the number of hits (N), penetration depth, and soil layer number of the dynamic penetration tester.

[0078] S2. Identify key areas and non-key areas in several scene pictures through pre-trained convolutional neural networks, perform multi-resolution hierarchical compression processing on each scene picture, using lossless compression for key areas and adaptive block lossy compression for non-key areas, and merge to generate several thumbnail sets with different resolutions, and establish a mapping relationship between several thumbnails and detection data.

[0079] S3. Based on the coordinate identifiers of the dynamic penetration instrument detection points and the thumbnail collection corresponding to the detection data, an index database is constructed, and the index database supports cross-modal retrieval.

[0080] S4. In response to the query instruction of the receiving terminal and the resolution requirement for the thumbnail, a thumbnail of the required resolution is obtained from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of each dynamic probe, and the thumbnail is transmitted to the receiving terminal.

[0081] In one embodiment, the steps added before step S1 include: S10. Obtain a video clip containing a detection point of the dynamic penetration instrument through a CDN edge node.

[0082] S11. Based on the inter-frame difference method, the number of hammer judgments of the dynamic probe in the video clip is counted, and the video clip is intercepted according to the time node of the statistical number of hammer judgments to obtain the on-site picture, and the watermark data of the number of hammer judgments N, penetration depth and soil layer number are attached to the preset watermark area of ​​the on-site picture.

[0083] S12. Perform optical character recognition on the preset watermark area of ​​the scene picture, compare the recognized watermark data with the detection data in the metadata of the scene picture, and trigger an abnormal marking signal if there is a conflict between the watermark data and the detection data in the metadata.

[0084] In one embodiment, the sub-steps of step S2 include: S20. Identify the key areas and non-key areas of the dynamic penetration instrument in the on-site picture through a pre-trained convolutional neural network, where the key areas include the penetration hammer and the scale lines of the probe rod of the dynamic penetration instrument, and the non-key areas are the remaining areas except the key areas.

[0085] In one embodiment, the detailed sub-steps of step S3 include: S30. Use the GPS coordinate identifier, detection time and detection data of the dynamic probe detection point as index dimensions to establish an index database, supporting combined condition retrieval based on any data in the detection data.

[0086] S31. Synchronously associate the video clips whose shooting range includes the GPS coordinates of the detection points of the dynamic probe with the on-site pictures containing the dynamic probe to form an evidence chain.

[0087] In one embodiment, the steps added after step S31 include: S310. Monitor the gradient change rate of the number of hits N of the dynamic probe in real time through a shooting device. When the difference between adjacent number of hits N exceeds the threshold value of the number of hits, automatically retrieve the detection data and penetration process video of the associated dynamic probe.

[0088] S311. Generate a spatiotemporal thermal diagram of the dynamic penetration instrument to mark abnormal areas, and support the layer-by-layer development of related evidence based on the chain of evidence.

[0089] In one embodiment, the sub-steps of step S4 include: S40. In response to the query instruction of the receiving terminal, data is filtered from the index database according to the coordinate identification, the number of hits N value and the multi-dimensional conditions of the soil layer type of the query instruction, and a test report is quickly generated according to a preset report template.

[0090] In one embodiment, the sub-steps of step S1 include: S13: Evaluate the transmission performance of each CDN node in real time and dynamically select the optimal node to establish a data channel.

[0091] S14. When the network is interrupted, the check value of the inspection data transmitted by the mobile terminal is recorded, and after the network is restored, the breakpoint transmission is resumed, and a timestamp is added to the data packet of the inspection data.

[0092] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM).

[0093] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

Claims

1. A method for managing dynamic penetration testing data, characterized in that: include: Receive test data and several associated site photos containing a dynamic penetration tester uploaded by mobile terminals in the construction area through CDN edge nodes, and embed test data into the metadata of each site photo. The test data includes the number of hits (N), penetration depth, and soil layer number of the dynamic penetration tester. Using a pre-trained convolutional neural network to identify key and non-key areas in several scene images, performing multi-resolution hierarchical compression processing on each scene image, wherein lossless compression is used for the key areas and adaptive block lossy compression is used for the non-key areas. The images are then combined to generate a set of thumbnails with different resolutions, and a mapping relationship between the thumbnails and the detection data is established; Building an index database based on the coordinate identifiers of the dynamic penetration instrument detection points and a thumbnail set corresponding to the detection data, wherein the index database supports cross-modal retrieval; In response to the query instruction of the receiving terminal and the resolution requirement for the thumbnail, a thumbnail of the required resolution is obtained from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of each dynamic probe, and the thumbnail is transmitted to the receiving terminal.

2. The method according to claim 1, characterized in that Before the step of receiving, via a CDN edge node, the test data uploaded by a mobile terminal located in the construction area and a plurality of associated on-site pictures including a dynamic penetration instrument, and embedding the test data into the metadata of each on-site picture, wherein the test data includes the number of hits N, the penetration depth, and the soil layer number of the dynamic penetration instrument, the method further includes: Obtaining a video clip containing the detection point of the dynamic penetration instrument through a CDN edge node; Based on the inter-frame difference method, the number of hammer determinations of the dynamic penetration instrument in the video clip is counted, and the video clip is intercepted according to the time node of the statistical number of hammer determinations to obtain an on-site picture, and watermark data of the number of hammer determinations N, the penetration depth and the soil layer number are attached to the preset watermark area of ​​the on-site picture; Optical character recognition is performed on a preset watermark area of ​​the scene picture, and the recognized watermark data is compared with the detection data in the metadata of the scene picture. If there is a conflict between the watermark data and the detection data in the metadata, an abnormal marking signal is triggered.

3. The method according to claim 2, characterized in that The steps of identifying key areas and non-key areas in a plurality of scene images by using a pre-trained convolutional neural network, performing multi-resolution hierarchical compression processing on each scene image, wherein lossless compression is used for the key areas and adaptive block lossy compression is used for the non-key areas, merging and generating a plurality of thumbnail sets of different resolutions, and establishing a mapping relationship between the plurality of thumbnails and the detection data include: The pre-trained convolutional neural network is used to identify the key areas and non-key areas of the dynamic probe in the on-site pictures, where the key areas include the core hammer and scale lines of the probe rod of the dynamic probe, and the non-key areas are the remaining areas except the key areas.

4. The method according to claim 3, characterized in that The step of constructing an index database based on the coordinate identifiers of the detection points of the dynamic penetration instrument and the thumbnail set corresponding to the detection data, wherein the index database supports cross-modal retrieval, includes: The GPS coordinate identification, detection time and detection data of the dynamic penetration instrument detection point are used as index dimensions to establish the index database, which supports retrieval based on the combination conditions of any data in the detection data; The video clips whose shooting range includes the GPS coordinates of the detection points of the dynamic probe are synchronously associated with the on-site pictures containing the dynamic probe to form an evidence chain.

5. The method according to claim 4, characterized in that After the step of synchronously associating the video clips including the GPS coordinates of the detection points of the dynamic penetration instrument with the on-site pictures including the dynamic penetration instrument to form an evidence chain, the method further includes: The camera device monitors the gradient change rate of the number of hits N of the dynamic penetration instrument in real time, and automatically retrieves the detection data and penetration process video associated with the dynamic penetration instrument when the difference between adjacent number of hits N exceeds the threshold value; Generate a spatiotemporal thermal diagram of the dynamic probe to mark abnormal areas, and support the layer-by-layer development of related evidence according to the chain of evidence.

6. The method according to claim 1, characterized in that The step of responding to the query instruction of the receiving terminal and the resolution requirement of the thumbnail, obtaining a thumbnail of the required resolution from the index database according to the coordinate identification, the number of hits N value and the multi-dimensional conditions of the soil layer type of each dynamic penetration instrument, and transmitting the thumbnail to the receiving terminal also includes: In response to the query instruction of the receiving terminal, data is filtered from the index database according to the coordinate identification, number of hits N value and multi-dimensional conditions of the soil layer type of the query instruction, and a test report is quickly generated according to a preset report template.

7. The method according to claim 2, characterized in that The step of receiving, via a CDN edge node, detection data uploaded by a mobile terminal located in a construction area and a plurality of associated on-site pictures including a dynamic penetration instrument, and embedding detection data in metadata of each on-site picture, wherein the detection data includes the number of hits N, penetration depth, and soil layer number of the dynamic penetration instrument, further includes: Real-time evaluation of the transmission performance of each CDN node and dynamic selection of the optimal node to establish a data channel; When the network is interrupted, the check value of the inspection data transmitted by the mobile terminal is recorded, and after the network is restored, the breakpoint resume is performed, and a timestamp mark is added to the data packet of the inspection data.

8. A dynamic penetration detection data management system, characterized in that: include: Image acquisition module (1): used for receiving detection data uploaded by mobile terminals in the construction area and a number of associated on-site pictures including a dynamic penetration instrument through CDN edge nodes, and embedding detection data into metadata of each on-site picture, wherein the detection data includes the number of hits N, penetration depth and soil layer number of the dynamic penetration instrument; Image thumbnail module (2): used to identify key areas and non-key areas in a number of scene images through a pre-trained convolutional neural network, perform hierarchical compression processing based on multiple resolutions on each scene image, wherein lossless compression is used for the key areas and adaptive block lossy compression is used for the non-key areas, and merge to generate a number of thumbnail sets with different resolutions, and establish a mapping relationship between the thumbnail sets and the detection data; Database construction module (3): used for constructing an index database based on the coordinate identifiers of the detection points of the dynamic penetration instrument and the thumbnail set corresponding to the detection data, wherein the index database supports cross-modal retrieval; Thumbnail transmission module (4): used for responding to the query instruction of the receiving terminal and the resolution requirement of the thumbnail, obtaining the thumbnail of the required resolution from the index database according to the coordinate identification, the number of hits N value and the multi-dimensional conditions of the soil layer type of each dynamic probe, and transmitting the thumbnail to the receiving terminal.

9. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and execute the method for managing dynamic penetration detection data according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The device stores a computer program that can be loaded by a processor and execute the method for managing dynamic penetration detection data according to any one of claims 1 to 7.

Images