Bridge deck pavement layer hidden water seepage channel positioning method, system, equipment and medium

Abstract

The invention belongs to the technical field of geological radar detection, and discloses a bridge deck pavement layer hidden water seepage channel positioning method, system and device and a medium, and the method comprises the steps: obtaining first geological radar detection data of a target bridge deck pavement layer under a preset dry condition and second geological radar detection data of the target bridge deck pavement layer under a preset wet condition; determining a moisture abnormal area in the target bridge deck pavement layer; determining water seepage path information of the target bridge deck pavement layer based on the hydrological simulation data; and determining a hidden water seepage channel of the target bridge deck pavement layer based on a matching relationship between the moisture abnormal region and the water seepage path information so as to realize positioning of the hidden water seepage channel of the bridge deck pavement layer. According to the method, the geological radar data under the dry and wet conditions are compared, and the hydrological simulation seepage path is combined for matching analysis, so that the hidden water seepage channel in the pavement layer can be accurately and reliably positioned on the premise of not damaging the structure of the bridge deck pavement layer.

Description

Technical Field

[0001] This invention relates to the field of ground-penetrating radar detection technology, and in particular to a method, system, equipment, and medium for locating hidden seepage channels in bridge deck pavement. Background Technology

[0002] With the continuous expansion of highway and bridge infrastructure and the increasing service life, the bridge deck pavement, as a key component directly bearing vehicle loads and environmental effects, has a significant impact on the overall safety and durability of bridges. In actual operation, rainwater, snowmelt, and other moisture often seep into the internal structure through cracks, joints, or local defects on the pavement surface, forming hidden seepage channels. Over time, this can easily lead to pavement loosening, peeling, potholes, and other defects, and may even further induce steel corrosion and concrete durability problems in the main bridge structure. Therefore, accurately locating the hidden seepage channels within the bridge deck pavement is of significant engineering importance.

[0003] In related technologies, detection methods for water seepage in bridge deck pavement layers largely rely on visual inspection, experience-based judgment, or single non-destructive testing methods. However, due to the complex internal structure and diverse material media of bridge deck pavement layers, the results of single ground-penetrating radar detection are easily affected by factors such as differences in structural layers and material heterogeneity. Its abnormal responses often have multiple solutions, making it difficult to accurately distinguish between anomalies caused by water infiltration and radar reflections generated by the structure's own characteristics. It also fails to reflect the actual migration path and connectivity of water within the pavement layer, leading to the determination of seepage channels relying primarily on experience-based inference, resulting in low positioning accuracy and reliability. In other words, related technologies suffer from insufficient accuracy in locating hidden seepage channels in bridge deck pavement layers.

[0004] Therefore, how to provide methods, systems, equipment, and media for locating hidden seepage channels in bridge deck pavement is an urgent problem to be solved. Summary of the Invention

[0005] The present invention provides a method, system, equipment and medium for locating hidden seepage channels in bridge deck pavement, in order to solve the problems of the above-mentioned technologies in the prior art.

[0006] According to a first aspect of the present invention, a method for locating hidden seepage channels in bridge deck pavement is provided.

[0007] In one embodiment, a method for locating hidden seepage channels in bridge deck pavement includes:

[0008] Acquire first ground-penetrating radar detection data of the target bridge deck pavement layer under preset dry conditions and second ground-penetrating radar detection data under preset wet conditions;

[0009] Based on the detection data from the first and second ground-penetrating radars, the area of ​​abnormal moisture inside the pavement layer of the target bridge deck was identified.

[0010] The target bridge deck pavement layer is simulated and analyzed using a preset hydrological simulation to obtain the corresponding hydrological simulation data, and the seepage path information of the target bridge deck pavement layer is determined based on the hydrological simulation data.

[0011] Based on the matching relationship between areas of abnormal moisture and seepage path information, the hidden seepage channels of the target bridge deck pavement layer are identified, so as to realize the location of the hidden seepage channels of the bridge deck pavement layer.

[0012] In one embodiment, determining the abnormal moisture area inside the target bridge deck pavement layer based on first and second ground-penetrating radar detection data includes:

[0013] Imaging processing was performed on the first and second ground-penetrating radar detection data respectively to obtain the first radar image sequence and the second radar image sequence.

[0014] The second radar image sequence is compared with the first radar image sequence to obtain a radar response difference map.

[0015] Extract connected regions from radar response difference maps where the signal strength exceeds a preset strength threshold and the spatial continuity satisfies a preset continuity threshold.

[0016] Based on the spatial topological relationship of the connected regions, a seepage channel network is formed, and the channels that are distributed in a linear or sheet-like manner in the seepage channel network are identified as abnormal moisture areas inside the target bridge deck pavement layer.

[0017] In one embodiment, connecting the interconnected regions into a seepage channel network based on their spatial topological relationships includes:

[0018] Obtain the parameters of the connected regions, verify the parameters of the connected regions, remove connected regions with abnormal parameters, unify the range of values ​​for the direction angle, and divide the depth interval according to the thickness of the target bridge deck pavement layer, so that all connected region parameters meet the spatial characteristic requirements of bridge deck seepage detection.

[0019] Calculate the planar distance between the center points of any two connected regions, quantify the orientation similarity between each connected region, and determine the depth fit between regions by the overlap of depth intervals, so as to complete the multi-dimensional quantitative analysis of the spatial topological relationship of connected regions.

[0020] Based on the actual distribution pattern of seepage channels in the target bridge deck pavement, scenario-based topological association judgment conditions are set, and the topological association between connected regions is judged according to the plane spacing, direction similarity, and depth adaptability. At the same time, the structural features in the design drawings of the target bridge deck pavement are used for auxiliary verification, and the topological association of connected regions that meet the scenario-based association characteristics is further judged.

[0021] For connected regions that are determined to be topologically related, draw connecting line segments with the nearest point on the region boundary as the endpoint to form a preliminary network structure of seepage channels;

[0022] Redundant connections are removed and boundaries are smoothed in the initial network structure, and finally structured seepage channel network data is output.

[0023] In one embodiment, a pre-set hydrological simulation is used to analyze the target bridge deck pavement layer to obtain corresponding hydrological simulation data. Based on the hydrological simulation data, the seepage path information of the target bridge deck pavement layer is determined, including:

[0024] Acquire dynamic response data of the target bridge deck pavement layer within a preset historical period;

[0025] Extract structural response anomaly patterns from dynamic response data that are temporally correlated with rainfall events or bridge water accumulation.

[0026] Based on the structural response anomaly pattern, the distribution of water-sensitive areas in the target bridge deck pavement layer was determined;

[0027] The surface of the target bridge deck pavement was sprayed with water in different areas, and the spatiotemporal variation data of humidity and the dynamic image sequence of runoff on the surface of the target bridge deck pavement were obtained respectively.

[0028] Spatiotemporal correlation analysis was performed on humidity spatiotemporal variation data and runoff dynamic image sequences to obtain the seepage path information of the target bridge deck pavement layer.

[0029] In one embodiment, determining the distribution of water-sensitive areas in the target bridge deck pavement layer based on structural response anomaly patterns includes:

[0030] Based on the design drawings of the target bridge deck pavement layer, a global plane coordinate system for the bridge deck is established, the installation coordinates of various sensors are retrieved, and the deviation is eliminated through coordinate transformation. A precise mapping relationship between the spatial coordinates of the sensors and the plane coordinates of the bridge deck is established to ensure that the accuracy of the sensor position projection meets the requirements.

[0031] We organized the triggering sensors corresponding to the abnormal response modes of each structure, extracted the bridge deck projection coordinates of various sensors, and combined them with the triggering frequency characteristics of the abnormal modes to screen out the set of key sensors that are strongly correlated with structural anomalies caused by water seepage.

[0032] Based on the sensor type and monitoring accuracy parameters, the bridge deck monitoring coverage range of each key sensor is set, the monitoring area is delineated with the sensor bridge deck projection coordinates as the center, and the overlapping monitoring areas of adjacent key sensors are merged to form a joint monitoring area.

[0033] Based on the pre-set bridge structural mechanics model, the load transfer range of the bridge deck pavement layer corresponding to the key sensor monitoring area is analyzed, and the intersection of the sensor monitoring area and the load transfer range is obtained to obtain preliminary candidate areas for water seepage sensitivity.

[0034] Isolated candidate regions with areas smaller than a preset value are removed. Based on the structural features of the bridge deck, the boundaries of the candidate regions are adjusted and aligned with the structural features. Adjacent candidate regions with consistent features are merged, and finally, a distribution map of water-sensitive areas and visual annotations of the bridge deck are generated.

[0035] In one embodiment, spatiotemporal correlation analysis is performed on humidity spatiotemporal variation data and runoff dynamic image sequences to obtain seepage path information of the target bridge deck pavement layer, including:

[0036] Standardized preprocessing was performed on the spatiotemporal variation data of humidity and the dynamic image sequence of runoff, respectively, to establish the coordinate system mapping relationship between the spatiotemporal variation data of humidity and the dynamic image sequence of runoff, and to achieve accurate spatiotemporal matching based on timestamps;

[0037] Calculate the humidity change rate and peak humidity occurrence time of each humidity sensor, and select humidity sensors that meet the preset screening conditions; extract runoff regions from the runoff dynamic image sequence, analyze the runoff direction vector, and identify the confluence concentration area;

[0038] The physical locations of humidity sensors that meet the preset screening criteria are mapped onto the runoff dynamic image, and potential infiltration points that simultaneously meet the following criteria are marked: humidity change index meets the standard, peak occurrence time is earlier, they are located in the confluence area, and the time difference between humidity increase and runoff occurrence meets the preset requirements.

[0039] A sequence is constructed based on the humidity peak time series of humidity sensors that meet the preset screening conditions. The seepage source path is traced by combining the runoff direction vector. The migration order is determined by associating the humidity time series differences of surrounding humidity sensors. The seepage channels inside the target bridge deck pavement layer are supplemented based on the humidity gradient between adjacent humidity sensors.

[0040] By integrating information on potential infiltration points, seepage migration direction angles, key migration nodes, and migration speeds, and verifying and correcting the data on structural joints and porosity in the design drawings of the target bridge deck pavement layer, the migration direction of paths exceeding the preset deviation range is adjusted, and finally, standardized seepage path information containing text descriptions and visual path diagrams is output.

[0041] In one embodiment, determining the hidden seepage channels in the target bridge deck pavement layer based on the matching relationship between areas of abnormal moisture and seepage path information, in order to locate the hidden seepage channels in the bridge deck pavement layer, includes:

[0042] The spatial coordinate set of the water anomaly area is spatially superimposed with the path spatial coordinate set of the seepage path information to determine the spatial overlap, direction consistency index, and depth correlation index between the water anomaly area and the seepage path information.

[0043] Based on spatial overlap, directional consistency index, and depth correlation index, the comprehensive score of each path segment in the seepage path information is determined.

[0044] All paths with comprehensive scores exceeding the preset matching threshold are segmented and combined to obtain hidden seepage channels, thereby enabling the location of hidden seepage channels in the bridge deck pavement layer.

[0045] According to a second aspect of the present invention, a system for locating hidden seepage channels in bridge deck pavement is provided.

[0046] In one embodiment, the bridge deck pavement hidden seepage channel locating system includes:

[0047] The data acquisition unit is used to acquire first ground-penetrating radar detection data of the target bridge deck pavement layer under preset dry conditions and second ground-penetrating radar detection data under preset wet conditions.

[0048] The area determination unit is used to determine the area of ​​abnormal moisture inside the pavement layer of the target bridge deck based on the detection data of the first and second ground-penetrating radars.

[0049] The path determination unit is used to simulate and analyze the target bridge deck pavement layer using a preset hydrological simulation, obtain the corresponding hydrological simulation data, and determine the seepage path information of the target bridge deck pavement layer based on the hydrological simulation data.

[0050] The channel determination unit is used to determine the hidden seepage channels of the target bridge deck pavement layer based on the matching relationship between the abnormal moisture area and the seepage path information, so as to realize the positioning of the hidden seepage channels of the bridge deck pavement layer.

[0051] According to a third aspect of the present invention, a computer device is provided.

[0052] In some embodiments, the computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0053] According to a fourth aspect of the present invention, a computer-readable storage medium is provided.

[0054] In one embodiment, a computer program is stored on the computer-readable storage medium, which, when executed by a processor, implements the steps of the above method.

[0055] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0056] (1) This invention compares geological radar data under dry and wet conditions and combines hydrological simulation of seepage paths for matching analysis. Without damaging the structure of the bridge deck pavement layer, it can achieve accurate and reliable positioning of hidden seepage channels inside the pavement layer.

[0057] (2) By acquiring two sets of ground-penetrating radar detection data of the bridge deck pavement layer under dry and wet conditions respectively, and comparing and analyzing the two, the present invention can identify the abnormal water area caused by the entry of water, and can distinguish the abnormal response caused by the material's own structural characteristics and water changes, thereby reducing the ambiguity of single ground-penetrating radar detection results and improving the accuracy of identifying abnormal water areas inside the pavement layer.

[0058] (3) By performing a pre-set hydrological simulation on the target bridge deck pavement layer and obtaining the corresponding hydrological simulation data, this invention can reflect the seepage process and path characteristics of water in the bridge deck pavement layer, transforming the seepage path from an invisible state into analyzable and traceable information, providing a reliable hydrological basis for the subsequent judgment of hidden seepage channels; by matching the water anomaly area obtained based on ground-penetrating radar data with the seepage path information determined based on hydrological simulation, and verifying the consistency of spatial and behavioral characteristics, areas that simultaneously possess water anomaly characteristics and actual seepage characteristics can be screened out, thereby more reliably determining the hidden seepage channels inside the bridge deck pavement layer and reducing misjudgment and omission; without damaging the structure of the bridge deck pavement layer, it can realize the location of hidden seepage channels inside the pavement layer, which helps to discover potential seepage problems when surface defects are not obvious, and provides a basis for subsequent targeted treatment.

[0059] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit the invention. Attached Figure Description

[0060] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0061] Figure 1 This is a flowchart illustrating a method for locating hidden seepage channels in bridge deck pavement according to an exemplary embodiment;

[0062] Figure 2This is a schematic diagram of a hidden seepage channel positioning system for bridge deck pavement layer according to an exemplary embodiment.

[0063] Figure 3 This is a schematic diagram of the structure of a computer device according to an exemplary embodiment. Detailed Implementation

[0064] The following description and accompanying drawings fully illustrate specific embodiments described herein to enable those skilled in the art to practice them. Some portions and features of certain embodiments may be included in or replace portions and features of other embodiments. The scope of the embodiments herein includes the entire scope of the claims and all available equivalents thereof. The various embodiments described herein are presented in a progressive manner, with each embodiment focusing on its differences from other embodiments; similar or identical parts between embodiments can be referred to interchangeably.

[0065] The modules in the apparatus or system of this application can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0066] Where there is no conflict, the embodiments and features in the embodiments of the present invention can be combined with each other.

[0067] Figure 1 An embodiment of the method for locating hidden seepage channels in bridge deck pavement layer according to the present invention is shown.

[0068] In this optional embodiment, the method for locating hidden seepage channels in the bridge deck pavement layer includes:

[0069] Step S101: Obtain the first ground-penetrating radar detection data of the target bridge deck pavement layer under preset dry conditions and the second ground-penetrating radar detection data under preset wet conditions;

[0070] Specifically, the target bridge deck pavement layer is the entire bridge pavement structure to be tested for hidden water seepage channels. It encompasses the surface structure that directly bears vehicle loads and environmental effects, the intermediate transition layer, and the bottom layer that connects to the main bridge structure. For example, the target bridge deck pavement layer could be the right-hand pavement layer of a continuous beam bridge on an urban expressway. This pavement layer consists of a 4 cm thick asphalt concrete surface layer, a 6 cm thick cement concrete leveling layer, and the base layer connection area. Preset drying conditions are environmental conditions designed to ensure the accuracy of ground-penetrating radar detection data, guaranteeing that the target bridge deck pavement layer is free from moisture intrusion and in a stable, dry state. The criteria for determining preset drying conditions are: the target bridge deck has been free from rainfall for at least 48 hours, and the surface moisture content of the pavement layer, as measured by a surface moisture meter, is no higher than 3%, while there is no residual moisture inside the pavement layer affecting electromagnetic wave propagation. The first ground-penetrating radar (GPR) detection data is the raw data set collected by detecting the pavement layer of the target bridge deck under the aforementioned preset dry conditions. This data includes information such as the reflection signals of electromagnetic waves at different medium interfaces within the pavement layer, propagation time, amplitude intensity, and detection location coordinates, which can reflect the original structural characteristics of the pavement layer under dry conditions. According to the pre-planned survey line scheme, the high-frequency antenna of the GPR system is placed close to the bridge deck surface and moved at a constant speed along the survey line for detection. At the same time, the spatial coordinates of each detection point are recorded synchronously by a high-precision positioning device. For example, using a GPR system with a 2.0 GHz high-frequency antenna, a survey line is laid every 5 meters along the longitudinal section (parallel to the driving direction) and a survey line is laid every 8 meters along the cross section (perpendicular to the driving direction) of the target bridge deck. The moving speed is controlled at 3 km / h, and the first ground-penetrating radar detection data containing reflection waveforms and detection coordinates is collected.

[0071] Specifically, the preset humidity conditions are designed to simulate a water intrusion scenario, ensuring that the surface and internal defect areas of the target bridge deck pavement fully absorb moisture, creating a preset environmental condition with a significant difference in dielectric constant compared to the dry state. This can be simulated by artificial water spraying to mimic moderate to heavy natural rain, ensuring that moisture can penetrate into hidden defects within the pavement layer (such as cracks and voids), and that there is no significant water accumulation on the pavement surface affecting the detection operation. For example, a water truck equipped with a uniform spraying device can be used to slowly spray water along the target bridge deck at a spraying intensity of 2.5 mm / min for 45 minutes. After this, there should be no significant water accumulation on the bridge deck surface, indicating that moisture has penetrated into potential defect areas within the pavement layer, thus achieving the preset humidity conditions. The second ground-penetrating radar (GPR) detection data is the original data collected under the aforementioned preset humidity conditions, using the same GPR system, survey line scheme, and detection parameters as the first GPR detection data, to re-detect the target bridge deck pavement layer. This data reflects the difference in dielectric constant between the wet and dry cement-containing areas within the pavement layer after moisture intrusion, thus highlighting the radar signal characteristics of areas with abnormal moisture levels. After water spraying, the target bridge deck pavement layer can be re-measured immediately under the same conditions as the first ground-penetrating radar detection data, including the acquisition path, movement speed, and antenna parameters, ensuring the comparability of the two sets of data. Environmental conditions (e.g., temperature and humidity) during the detection process can be recorded simultaneously. By acquiring ground-penetrating radar detection data of the target bridge deck pavement layer under preset dry and preset wet conditions, a comparative basis for the same pavement structure under different moisture states can be established. This highlights the impact of moisture intrusion on radar response characteristics, providing data support for subsequently distinguishing between structural features and anomalies caused by moisture changes, thereby improving the targeting and effectiveness of detection and analysis from the source.

[0072] Step S102: Based on the detection data of the first and second ground-penetrating radars, determine the abnormal moisture area inside the pavement layer of the target bridge deck;

[0073] Specifically, the moisture anomaly area is the region inside the pavement layer of the target bridge deck that is formed by moisture intrusion under pre-set humid conditions, exhibiting a significant change in dielectric constant compared to the dry state, and showing obvious signal differences in ground-penetrating radar detection data. The moisture anomaly area is the core carrier of hidden water seepage channels, usually corresponding to hidden defects such as cracks, voids, and interlayer separation inside the pavement layer. These defects are filled with moisture under humid conditions, causing changes in electromagnetic wave propagation characteristics and forming identifiable radar signal features. By systematically comparing and analyzing the first ground-penetrating radar detection data (dry state baseline data) and the second ground-penetrating radar detection data (humid state data containing moisture), the signal difference areas caused by moisture intrusion can be screened out, and the signal interference corresponding to non-moisture-related original defects (such as inherent aggregate voids, slight structural inhomogeneities, etc.) existing under the dry state can be eliminated. Finally, the anomaly areas that only appear under humid conditions and are related to the dielectric properties of moisture can be identified.

[0074] Specifically, the process of identifying abnormal moisture areas within the pavement layer of a target bridge deck based on data from both the first and second ground-penetrating radars (GPR) utilizes the influence of moisture on electromagnetic wave propagation. Moisture significantly increases the dielectric constant of the pavement medium, causing significant differences in the intensity, phase, and propagation time of the reflected high-frequency electromagnetic waves emitted by the GPR in the cement-containing wet area compared to the dry area. By using the first GPR data in a dry state as a benchmark and simultaneously comparing it with the second GPR data in a wet state, the signal changes caused by moisture intrusion are accurately captured, thereby locating abnormal areas of moisture accumulation. This overcomes the limitation of single-state detection in distinguishing between primary defects and seepage defects. Comparative analysis of the first and second GPR data can identify abnormal response areas caused by moisture infiltration, thus determining abnormal moisture areas within the pavement layer of the target bridge deck. This reduces the ambiguity in interpreting single GPR data and improves the accuracy of identifying the distribution of moisture and the location of anomalies within the pavement layer.

[0075] Step S103: Use the preset hydrological simulation to simulate and analyze the target bridge deck pavement layer, obtain the corresponding hydrological simulation data, and determine the seepage path information of the target bridge deck pavement layer based on the hydrological simulation data.

[0076] Specifically, the pre-set hydrological simulation is a pre-set experimental operation designed to recreate the seepage process of the target bridge deck pavement under natural rainfall conditions. It simulates the dynamic scenarios of water replenishment, infiltration, and confluence in a controlled manner to induce the hydrological response of hidden seepage channels and capture relevant data. The hydrological simulation data is a collection of raw data collected during the pre-set hydrological simulation using various observation devices. This data reflects the flow state of water on and within the target bridge deck pavement, including key information such as the water replenishment process, surface runoff characteristics, and seepage point responses. It is the core basis for inferring the seepage path, specifically covering the runoff direction, velocity, and flow rate; the location, occurrence time, and flow rate changes of seepage points; and the dynamic distribution of water on the bridge deck. After the pre-set hydrological simulation is initiated, high-definition cameras can be deployed to capture the overall dynamics of surface runoff on the bridge deck in real time, recording the runoff confluence path; flow meters can be used to measure runoff velocity at key cross-sections of the bridge deck; and stopwatches can be used to record the initial flow at each seepage point. The time of occurrence was used to collect the water flow at the seepage points at regular intervals using a measuring cup and calculate the instantaneous flow rate; at the same time, the spatial location of all seepage points was manually marked (coordinates were recorded using a differential GPS system); for example, in a preset hydrological simulation of the pavement layer of a certain ring expressway, the runoff was captured by a high-definition camera and flowed to both sides along the centerline of the bridge deck. The flow velocity at the key section was measured to be 0.15 m / s using a flow meter, and three seepage points were recorded with coordinates (X1, Y1), (X2, Y2), and (X3, Y3), respectively. The first occurrence times were 12 minutes, 18 minutes, and 25 minutes after watering, respectively, and the instantaneous flow rates were 5 ml / s, 3 ml / s, and 4 ml / s, respectively, forming a complete set of hydrological simulation data.

[0077] Specifically, seepage path information is derived from hydrological simulation data. By analyzing and integrating runoff characteristics and the response relationship of seepage points, it deduces the complete path network information of water infiltrating from the surface of the target bridge deck pavement layer, migrating internally, and finally seeping out. This includes the approximate range of infiltration points, the dominant direction of internal seepage, the connectivity of different segments, and the corresponding relationships of seepage points. This information is key to reflecting the macroscopic distribution of seepage channels. Analysis of runoff direction and confluence areas in hydrological simulation data can identify surface areas where water is easily infiltrated (candidate areas for infiltration points). By combining the occurrence time and flow rate variation patterns of each seepage point, the correspondence between different infiltration points and seepage points can be determined (e.g., early-appearing seepage points correspond to near-infiltration areas, and high-flow-rate seepage points correspond to main seepage channels). Integrating the above information, a coherent path is constructed from the candidate infiltration point area to the internal migration path to the seepage point, forming the seepage path information.

[0078] Specifically, the process of determining the seepage path information of the target bridge deck pavement layer based on the hydrological simulation data generated by the pre-set hydrological simulation involves using the controllable water input conditions created by the pre-set hydrological simulation to capture the dynamic response data of the seepage process. Through data correlation analysis, the complete flow trajectory of the seepage is reconstructed, overcoming the limitations of traditional static observation in obtaining internal seepage paths. This provides a macroscopic and dynamic path benchmark for subsequent matching with water anomaly areas identified by ground-penetrating radar, ensuring that the location of hidden seepage channels conforms to both internal physical anomaly characteristics and actual seepage patterns. Performing a pre-set hydrological simulation on the target bridge deck pavement layer and obtaining the corresponding hydrological simulation data reflects the seepage process and migration path of water in the pavement layer, transforming the originally unobservable seepage behavior into analyzable and traceable seepage path information, providing a reliable hydrological basis for subsequent identification of hidden seepage channels.

[0079] Step S104: Based on the matching relationship between the abnormal moisture area and the seepage path information, determine the hidden seepage channels of the target bridge deck pavement layer to realize the location of the hidden seepage channels of the bridge deck pavement layer.

[0080] Specifically, the matching relationship is the degree of fit between the abnormal moisture area inside the target bridge deck pavement layer and the seepage path information in terms of spatial location and morphological characteristics. The core is reflected in whether the distribution range of the abnormal moisture area is consistent with the direction of the seepage path, whether the spatial coordinates overlap, and whether the shape (linear or sheet-like) matches the macroscopic characteristics of the seepage channel. It is the basis for judging whether the abnormal moisture area is a real seepage channel. Hidden seepage channels are continuous channels existing within the pavement layer of a target bridge deck, without obvious signs of surface damage, but simultaneously satisfying the conditions of internal moisture accumulation (corresponding to abnormal moisture areas) and actual moisture seepage (corresponding to seepage path information). Essentially, they are connected paths formed by hidden defects such as internal cracks, interlayer separation, and voids in the pavement layer, which can guide water flow migration. They are the core hidden dangers leading to damage to the bridge deck pavement layer and the main structure of the bridge. Based on the above matching relationship, abnormal moisture areas that have a high degree of matching with the seepage path information can be identified as hidden seepage channels, while non-seepage abnormal areas with low matching (such as simply loose aggregate or isolated voids without seepage) can be excluded. This ensures that the identified channels have both the physical characteristics of internal moisture accumulation and conform to the dynamic laws of actual seepage. By matching and analyzing the identified areas of abnormal moisture with the information on seepage paths, and by verifying the consistency between spatial location and seepage behavior, areas that simultaneously exhibit abnormal moisture characteristics and actual seepage characteristics can be screened out. This allows for more reliable identification of hidden seepage channels within the target bridge deck pavement layer, reducing misjudgments and omissions, and achieving precise positioning without damaging the pavement structure.

[0081] In this optional embodiment, when determining the abnormal moisture region inside the target bridge deck pavement layer based on the first and second ground-penetrating radar detection data, the first and second ground-penetrating radar detection data can be imaged to obtain a first radar image sequence and a second radar image sequence. The second radar image sequence and the first radar image sequence are then subjected to a differential operation (i.e., pixel-level or feature-level differential operation) to obtain a radar response difference map. Connected regions with signal strength exceeding a preset strength threshold and spatial continuity satisfying a preset continuity threshold are extracted from the radar response difference map. These connected regions are then connected into a seepage channel network based on their spatial topology, and the channels in the seepage channel network that are linear or sheet-like are identified as abnormal moisture regions inside the target bridge deck pavement layer.

[0082] Specifically, the first and second radar image sequences are continuous, visualized image sets formed after professional imaging processing of the first and second ground-penetrating radar (GPR) data (dry conditions and wet conditions, respectively). Both image sequences visually present the reflection characteristics of GPR electromagnetic waves within the pavement layer in two-dimensional or three-dimensional image form. The first radar image sequence reflects the original structural characteristics of the pavement layer under dry conditions, while the second radar image sequence reflects the changes in reflection signals in the cement-containing wet area after moisture intrusion. Ground-penetrating radar data processing software can be used to sequentially filter the original detection data (mean filtering, median filtering) and adjust the gain (self-weighting). Imaging processing techniques such as dynamic gain control, background removal, and gather overlay convert discrete radar signals into a sequence of recognizable profiles or planar images, ensuring the clarity of medium interfaces and anomalous areas in the images. For example, for the left-hand pavement layer of a ring expressway from K12+300 to K12+500, imaging processing of the first ground-penetrating radar data under dry conditions yielded a first radar image sequence consisting of 80 consecutive profiles, clearly showing the interface reflection waves between the asphalt surface layer and the cement leveling layer. The same imaging processing was applied to the second ground-penetrating radar data under humid conditions, resulting in a corresponding second radar image sequence, with some areas showing newly added strong reflection wave characteristics.

[0083] Specifically, pixel-level or feature-level differential operations quantify the differences between the second and first radar image sequences through mathematical calculations, highlighting the changes in radar response caused by water intrusion. Pixel-level differential operations subtract (or calculate ratios) the grayscale values ​​or signal amplitudes of the two images pixel by pixel, directly reflecting the signal differences of individual pixels. Feature-level differential operations first extract key feature parameters such as reflected wave amplitude, phase, and waveform morphology from the two images, and then calculate the differences or similarities of these feature parameters, focusing on changes in core features. Both types of radar image sequences can be imported into professional image processing software. For pixel-level operations, a point-by-point calculation logic is used between the pixel values ​​of the second and first images. For feature-level operations, reflected wave feature points in the images are first extracted through edge detection and threshold segmentation, and then the amplitude differences and phase shifts of these feature points are calculated to generate differential result data.

[0084] Specifically, the radar response difference map is a visualization result of pixel-level or feature-level differential operations. It is presented in image form and can intuitively reflect the signal differences between the second radar image sequence and the first radar image sequence. The radar response difference map uses different colors or grayscale gradients to indicate the degree of difference. The high difference area corresponds to the area inside the paving layer where the dielectric constant changes due to moisture intrusion, which is the core basis for locating moisture anomalies. The pixel-level or feature-level difference data obtained from the differential operation can be converted into a visualization image through pseudo-color encoding or grayscale mapping. The larger the difference value, the more vivid the corresponding color (e.g., red represents the maximum difference), generating a continuous map file to ensure that the boundaries and ranges of the difference areas can be clearly identified. For example, by pseudo-color encoding the above pixel-level difference data to generate a radar response difference map, the segments K12+380-K12+410 and K12+450-K12+470 in the map show obvious red high difference areas, indicating that the radar signal changes significantly after the area becomes wet, suggesting the presence of moisture accumulation.

[0085] Specifically, connected regions whose signal strength exceeds a preset strength threshold and whose spatial continuity meets a preset continuity threshold are selected from the radar response difference map. These regions simultaneously meet the criteria for signal difference and spatial continuity, eliminating isolated noise interference and are highly likely to be valid anomaly regions formed by moisture intrusion. The preset strength threshold is determined statistically based on the difference signal of normal regions without moisture interference, used to distinguish between valid differences and noise. The preset continuity threshold is a set standard for the continuous length or area of ​​the region, used to exclude isolated false anomalies. The preset strength threshold (usually the average of the normal region) can be determined first by analyzing the difference signal strength of normal regions in the radar response difference map and using statistical methods (e.g., mean + 2 standard deviation). The difference value is set at 2-3 times. Specifically, the difference map region corresponding to the radar response data of the target bridge deck pavement in a dry state is selected as the normal region sample. The difference signal intensity values ​​of all pixels in the sample are counted, and the sample mean μ and standard deviation σ are calculated. μ+2σ is used as the preset intensity threshold to ensure that the effective difference signal caused by water intrusion can be effectively distinguished from the invalid signal caused by environmental noise and equipment error, which is suitable for the scene characteristics of radar detection of bridge deck pavement. Then, according to the common size of the hidden water seepage channel in the pavement, a preset continuous threshold is set (for example, the continuous length is greater than or equal to 3 meters or the continuous area is greater than or equal to 5 square meters). Finally, through image segmentation and connected component analysis algorithms, continuous regions that simultaneously meet the above two threshold conditions are selected.

[0086] Specifically, the application architecture of the connected component analysis algorithm in the field of ground-penetrating radar detection technology is as follows: First, input a binary radar response difference map filtered by a preset intensity threshold (pixel value 1 represents signal strength meeting the standard, 0 represents not meeting the standard). Second, traverse the map using an eight-neighbor connected component marking algorithm, marking adjacent (including horizontal, vertical, and diagonal directions) pixels with a pixel value of 1 as the same connected component. Third, introduce morphological closing operations (dilation followed by erosion) to process the connected components, filling in small pores and broken areas within the components, adapting to the discontinuous signal characteristics that may exist in the hidden seepage channels of the pavement layer during ground-penetrating radar detection. Fourth, output the minimum circumference of each connected component. The algorithm takes parameters such as rectangular coordinates, continuous length, and area as input. The input data is directly related to the radar response difference map, and the output parameters provide a quantitative basis for subsequent comparison with the preset continuous threshold, ensuring that the algorithm is accurately combined with the bridge deck seepage detection scenario. For example, for the radar response difference map, by statistically analyzing the difference signal intensity in normal areas, the preset intensity threshold is determined to be 35 (grayscale difference). Combined with the common size of seepage channels, the preset continuous threshold is set to a continuous length ≥ 3 meters. After the algorithm is used for screening, two connected regions that meet the conditions are obtained, namely the K12+380-K12+410 segment (continuous length 30 meters) and the K12+450-K12+470 segment (continuous length 20 meters).

[0087] In this optional embodiment, when connecting the spatial topological relationships of connected regions to form a seepage channel network, connected region parameters can be obtained, and these parameters can be verified to eliminate connected regions with abnormal parameters. The range of values ​​for the orientation angles can be standardized, and depth intervals can be divided based on the thickness of the target bridge deck pavement layer, ensuring that all connected region parameters meet the spatial characteristic requirements for bridge deck seepage detection. The planar distance between the center points of any two connected regions can be calculated to quantify the orientation similarity between connected regions. The depth compatibility between regions can be determined by the overlap of depth intervals, thus completing a multi-dimensional quantitative analysis of the spatial topological relationships of connected regions. The system sets scenario-based topological association judgment conditions based on the actual distribution pattern of seepage channels in the target bridge deck pavement layer. The topological association between connected regions is determined based on plane spacing, direction similarity, and depth adaptability. At the same time, the system uses structural features in the design drawings of the target bridge deck pavement layer for auxiliary verification. The topological association of connected regions that meet the scenario-based association characteristics is further judged. For connected regions that are determined to be topologically associated, connecting line segments are drawn with the nearest point of the region boundary as the endpoint to form a preliminary network structure of seepage channels. Redundant connections are eliminated and the boundaries are smoothed in the preliminary network structure. Finally, the structured seepage channel network data is output.

[0088] Specifically, the seepage channel network is a network structure that reflects the distribution of potential seepage channels by connecting multiple connected regions selected from radar response difference maps according to their spatial topological relationships (e.g., adjacent locations, consistent orientations, and similar depths). This network integrates scattered connected regions, restores the possible connectivity relationships of seepage channels, and avoids the loss of channel integrity caused by isolated analysis of a single region. Spatial topology analysis algorithms can be used in geographic information systems or professional image analysis software to perform correlation analysis on the location, orientation, and depth of the selected connected regions, connecting connected regions that are close in distance (e.g., spacing less than or equal to 1 meter), have consistent orientations, and are within the same depth range to form a continuous network structure.

[0089] Specifically, when adapting the spatial topology analysis algorithm to the construction scenario of bridge deck seepage channel network in the field of ground-penetrating radar detection technology, the core architecture and application logic are as follows: The algorithm input data is explicitly defined as the core parameters of each connected region selected after radar detection, including the three-dimensional coordinates (x, y, z) of the region center point based on the bridge deck global coordinate system (with the bridge starting point as the origin, x-axis along the longitudinal direction, y-axis along the transverse direction, and z-axis along the pavement layer depth direction), the coordinates of the vertices of the region boundary polygon, the orientation angle (angle with the longitudinal direction of the bridge deck), and the depth range (z-axis value range). The input data is directly related to the connected regions obtained by ground-penetrating radar detection, and the parameter format is adapted to the spatial scale of the bridge deck pavement layer (units are unified as follows). The algorithm architecture consists of four core steps: The first step is data preprocessing and standardization, which verifies the parameters of the input connected regions, removes abnormal regions with coordinate deviations exceeding 0.5 meters, unifies the orientation angles to a range of 0-180°, and divides the depth intervals according to the pavement thickness (e.g., 0-10 cm, 10-20 cm) to ensure that the parameters conform to the spatial characteristics of bridge deck seepage detection; The second step is the quantitative calculation of spatial topological relationships, which calculates the planar distance between the center points of any two connected regions using the Euclidean distance formula (ignoring the horizontal / vertical straight-line distances of depth differences), calculates the orientation similarity using the vector angle formula (the smaller the angle, the higher the similarity), and calculates the interval overlap formula. The first step is to determine depth compatibility (depth interval overlap rate greater than or equal to 70% is considered the same depth interval); the second step is to match association rules, based on the actual distribution pattern of seepage channels on the bridge deck, setting scenario-based association thresholds: plane spacing less than or equal to 1 meter, directional angle less than or equal to 15°, and depth interval overlap rate greater than or equal to 70%. Connected regions that meet these three conditions are judged to have topological association. At the same time, bridge deck structural features are introduced to assist in verification. If the distance between two regions is 1-2 meters but there is an expansion joint or structural joint in between (based on the bridge deck design drawings), it is still judged as associated (structural joints are easy to become the connecting carrier of seepage channels). The third step is network construction and optimization, which involves matching the relationships determined to be associated. The network structure is formed by drawing connecting lines from the nearest point on the boundary of each region to create a preliminary network structure. A topology cleaning algorithm is then used to remove redundant connecting lines and to add smooth transition curves to the boundaries of each region (adapting to possible curvatures of the seepage channels). The final output is structured seepage channel network data, including the 3D coordinates of network nodes (center points and endpoints of connected regions), the direction, length, and depth information of channel segments, a network topology table (recording associated objects in each region), and a visualized network graph. This output data not only restores the connectivity integrity of potential seepage channels in the bridge deck pavement layer but also provides accurate spatial topology data for subsequent seepage path analysis and protection scheme development.

[0090] When adapting the topology cleaning algorithm to the construction scenario of bridge deck seepage channel network in the field of ground-penetrating radar detection, its core lies in its deep integration with the spatial characteristics of the bridge deck pavement layer and the distribution pattern of seepage channels. Its input data is the seepage channel network structure data initially constructed in the fourth step, specifically including the three-dimensional coordinates of the connecting line segments at the nearest points of the associated connected regions (adapted to the global coordinate system of the bridge deck, x-axis longitudinal, y-axis lateral, z-axis depth direction, in meters), the direction angle of each line segment, the depth range, and the spatial positional relationship data between line segments. The input data directly inherits the network construction results after the association rule matching in the preceding steps, ensuring consistency with the initial topological form of the bridge deck seepage channels. The core operation logic of the algorithm is as follows: First, based on the intersection judgment rules of the three-dimensional coordinates of the line segments, identify the intersecting line segments and redundant line segments (such as multiple overlapping connecting line segments corresponding to the same connection relationship) generated by the connection of connected regions in the initial network. This is combined with the thin spatial scale of the bridge deck pavement layer (usually 0- The algorithm first removes redundant line segments (less than 0.3 meters in length) and abnormal intersections (intersecting with other line segments at angles greater than 90°) by eliminating invalid and redundant line segments (20 cm in length). Then, for the remaining valid connecting line segments, a Bezier curve fitting method is used to supplement smooth transition curves based on the difference in direction angles between adjacent connected regions and the boundary contour characteristics. This ensures that the curve curvature matches the possible natural bending shape of the seepage channel within the pavement layer (avoiding rigid, broken-line connections). The algorithm output data is the core component of the optimized structured seepage channel network, including the three-dimensional coordinates of the non-intersecting and redundant channel line segments, smooth transition curve parameters, and precise coordinates of network nodes (center points of connected regions and connection endpoints). This ensures that the output data not only conforms to the actual connection logic of the bridge deck seepage channels but also provides topologically correct spatial data support for subsequent seepage path analysis and protection scheme formulation. Based on this input / output setting and operation logic, the algorithm can be directly applied in the construction of the bridge deck seepage channel network.

[0091] Specifically, the essence of hidden seepage channels is a continuous defect within the pavement layer, such as cracks and interlayer voids. In radar response difference maps, they typically appear as a continuous distribution pattern of lines (corresponding to crack-type channels) or sheets (corresponding to interlayer voids and large-area void-type channels). Isolated, irregularly shaped areas are mostly noise or non-seepage defects. By screening channels of this specific shape, abnormal areas truly caused by water intrusion can be accurately located, eliminating false interference. Morphological analysis of the seepage channel network can be performed to extract the morphological parameters of each component in the network, such as aspect ratio and outline shape. Criteria for linear channels (aspect ratio greater than or equal to 5:1) and criteria for sheet-like channels (aspect ratio less than or equal to 3:1 and area greater than or equal to 5 square meters) can be set. Channel areas that meet the criteria are identified as abnormal water areas. Imaging processing was performed on ground-penetrating radar (GPR) data under dry and wet conditions, and pixel-level or feature-level differential operations were used to generate radar response difference maps, making the radar response changes caused by water intrusion explicit. Then, by combining signal strength thresholds and spatial continuity constraints to extract connected regions and construct a seepage channel network, the interference of random noise and isolated anomalies can be effectively suppressed, making the identified water anomaly areas more consistent with the spatial morphological characteristics of actual seepage channels, thereby improving the accuracy of GPR in identifying water anomalies inside bridge deck pavement layers.

[0092] In this optional embodiment, when performing simulation analysis on the target bridge deck pavement layer using a preset hydrological simulation to obtain corresponding hydrological simulation data, and determining the seepage path information of the target bridge deck pavement layer based on the hydrological simulation data, dynamic response data collected by the structural health monitoring system of the target bridge deck pavement layer within a preset historical period can be obtained. The dynamic response data may include beam strain data, vibration frequency data, and deflection change data. Structural response anomaly patterns with temporal correlation to rainfall events or bridge surface water conditions are extracted from the dynamic response data. Based on the structural response anomaly patterns, the distribution of seepage-sensitive areas in the target bridge deck pavement layer is determined. The surface of the target bridge deck pavement layer is subjected to zoned water spraying, and the spatiotemporal variation data of humidity and the dynamic image sequence of runoff on the surface of the target bridge deck pavement layer are obtained respectively. Spatiotemporal correlation analysis is performed on the spatiotemporal variation data of humidity and the dynamic image sequence of runoff to obtain the seepage path information of the target bridge deck pavement layer.

[0093] Specifically, a structural health monitoring system is a specialized monitoring system deployed on bridge structures to collect, transmit, and store bridge structure operation status data in real time, enabling real-time perception and recording of structural damage and anomalies. A structural health monitoring system can consist of various sensors, data acquisition modules, transmission modules, and a data storage platform. Its core function is to continuously capture the structural response of the bridge under load and environmental influences, providing data support for analyzing the impact of water seepage on the structure. For example, the structural health monitoring system installed on a section of a ring expressway from K12+300 to K12+500 includes distributed strain gauges, accelerometers, displacement gauges, and data acquisition terminals, which can collect dynamic response data of the beam structure in real time and store it on a cloud platform. The preset historical period is a pre-defined time range for retrieving data from the structural health monitoring system to ensure sufficient sample size for extracting effective structural response anomaly patterns. The preset historical period covers multiple rainfall events and different environmental conditions to ensure that seepage-related structural response patterns can be captured, avoiding misjudgments of anomaly patterns due to insufficient data or a single scenario. It can be determined based on the rainfall frequency of the target area, the service life of the bridge, and the structural stability, and can be set to 6 months to 1 year. For example, for the bridge mentioned above, combined with the local climate characteristics of approximately 80 rainy days per year, the preset historical period is set to 6 months (from January 1 to June 30 of the current year), which includes 12 rainfall events.

[0094] Specifically, dynamic response data is a set of raw data continuously collected by the structural health monitoring system within a preset historical period, reflecting the stress, vibration, and deformation state of the target bridge deck pavement and related beam structures. Dynamic response data directly reflects the structural state changes under different operating conditions such as normal operation, rainfall, and water accumulation, and is the core data source for analyzing the impact of water seepage on the structure. For example, exporting six months of monitoring data from the aforementioned structural health monitoring system includes beam strain records every 5 minutes, vibration frequency data every 10 minutes, and deflection change data every hour, totaling 120,000 data entries, fully covering various operating conditions within the preset historical period. Beam strain data is collected by strain gauges deployed at key sections of the beam, reflecting the quantitative data on the degree of internal stress deformation caused by load, temperature changes, or moisture intrusion. Its numerical changes are directly related to the stress state of the structural materials; when water seepage leads to structural material deterioration or uneven local stress, abnormal fluctuations in strain data will occur. Vibration frequency data, collected by accelerometers, is quantitative data reflecting the inherent vibration characteristics of the bridge structure. The vibration frequency is closely related to material stiffness and structural integrity. When water seepage causes bond failure between the pavement layer and the beam, or when the beam concrete deteriorates, the structural stiffness decreases, leading to an abnormal drop in vibration frequency. Deflection change data, collected by displacement gauges, reflects the vertical displacement changes of the beam under load or structural deformation, demonstrating the overall stiffness of the beam. When water seepage causes a decrease in the strength of the beam concrete or corrosion of the internal reinforcing steel, the beam's load-bearing capacity is affected, resulting in an abnormally large increase in deflection.

[0095] Structural response anomaly patterns are specific data change patterns selected from dynamic response data that are highly correlated with rainfall events or bridge water accumulation in time and deviate from the normal structural response range. These anomaly patterns directly reflect the impact of water seepage on the structural state, typically manifesting as abrupt changes in beam strain, a decrease in vibration frequency, and an abnormal increase in deflection after rainfall. They are repeatable and temporally correlated. By statistically analyzing the temporal correspondence between dynamic response data and rainfall / water accumulation events, a normal response threshold range can be set, and continuous data sequences exceeding the threshold and matching the rainfall / water accumulation time sequence can be selected to form standardized anomaly patterns. For example, analyzing the aforementioned 6 months of dynamic response data revealed a specific change sequence where, within 6-12 hours after the start of rainfall, beam strain increased by more than 20 με, vibration frequency decreased by more than 0.3 Hz, and deflection increased by more than 0.8 mm, and this state persisted for more than 24 hours. This sequence highly matched 8 rainfall events and was identified as a structural response anomaly pattern.

[0096] Specifically, extracting the abnormal structural response pattern requires preprocessing the bridge's dynamic response data (including parameters such as beam strain, vibration frequency, and deflection). This includes sensor zero-point calibration, eliminating outliers caused by equipment errors based on the 3σ principle, and completing the sampling gaps using linear interpolation to form a standardized time-series dataset with unified timestamps. Simultaneously, concurrent meteorological rainfall data (rainfall start and end times, cumulative rainfall) and bridge water accumulation monitoring data (water accumulation formation time, water depth) are collected to establish a three-dimensional data association library for rainfall-water accumulation-structural response. A time correlation window is defined, using the rainfall start time or water accumulation formation time as a baseline, setting response observation periods of 24 hours before and after. The structural dynamic response data within this period is compared and analyzed with historical normal response data from periods without rainfall or water accumulation, using statistical methods (such as the mean-standard deviation method) to calculate... The normal fluctuation range of each response parameter (e.g., normal range of beam strain is ±5με, normal range of vibration frequency is ±0.1Hz of the design value, and normal range of deflection is ±0.3mm) is used to determine the abnormal threshold of each parameter; continuous data sequences exceeding the abnormal threshold within the observation period are screened out, and the temporal consistency of the sequence with the rainfall / water accumulation event is verified, for example, the response mutation occurs within 6-12 hours after the start of rainfall, and the duration of the abnormal state matches the duration of rainfall / water accumulation; the repeatability of the screened sequences corresponding to multiple rainfall / water accumulation events is verified, and the consistent characteristic patterns in each event are extracted, such as the common indicators such as the magnitude of strain increase, the value of frequency decrease, the amount of deflection increase, and the duration, etc., and occasional isolated abnormal data are eliminated, finally forming a standardized structural response anomaly pattern, ensuring the temporal correlation and repeatability of this pattern with the rainfall / water accumulation state.

[0097] Specifically, the distribution of water-sensitive areas is determined based on structural response anomaly patterns. This is achieved by combining the spatial correspondence between sensor deployment locations and the bridge deck pavement layer. The identified areas within the target bridge deck pavement layer that are prone to water seepage and will trigger structural response anomalies after seepage are pinpointed. This process clarifies high-risk areas requiring close monitoring, providing a precise basis for subsequent differentiated zonal water spraying. First, a mapping relationship can be established between the spatial coordinates of each sensor in the structural health monitoring system and the planar coordinates of the target bridge deck pavement layer. Then, based on the sensor locations corresponding to the structural response anomaly patterns, the areas triggering the anomalies are determined. Structural regions; finally, this structural region is mapped in reverse to the bridge deck pavement layer to delineate the corresponding area range and form a distribution of water-sensitive areas; for example, in the above bridge, the sensors corresponding to the three types of abnormal structural response modes are located on the left side of the mid-span of the beam, the right side near the support, and around the expansion joint, respectively. Through spatial mapping, the left lane of the K12+350-K12+380 section, the right lane of the K12+420-K12+450 section, and the area around the expansion joint of the K12+390-K12+410 section in the bridge deck pavement layer are identified as water-sensitive areas, forming a distribution map containing three key sensitive areas.

[0098] Specifically, the surface of the target bridge deck pavement is subjected to zoned water spraying. The water spraying intensity or start time of different zones is differentiated based on the preset bridge structural characteristics and the distribution of seepage-sensitive areas. During the water spraying process, multiple humidity sensors deployed on the surface of the target bridge deck pavement acquire spatiotemporal humidity variation data, and an image acquisition device acquires a dynamic image sequence of runoff on the surface of the target bridge deck pavement. Spatiotemporal correlation analysis is performed on the spatiotemporal humidity variation data and the dynamic runoff image sequence to obtain seepage path information.

[0099] Specifically, zoned water spraying is a hydrological simulation method that divides the surface of the target bridge deck pavement into multiple independent and continuous detection areas based on the structural distribution and potential seepage risk differences. Controllable water spraying equipment is used to spray each area separately. This avoids the problems of insufficient water supply to critical areas or data redundancy in non-critical areas caused by uniform water spraying. It enables precise capture of seepage responses in different areas, providing more targeted data support for subsequent seepage path analysis. In the process, based on the target bridge deck's planar layout and pavement layer structural zoning (e.g., the junction area between the asphalt surface layer and the cement leveling layer, the boundary area between the lane and the emergency lane), combined with the suspected seepage areas identified in the preliminary survey, the bridge deck is divided into several independent rectangular areas with sides of 5-10 meters using marking lines or positioning stakes. Then, an independent and controllable water spraying unit (e.g., zoned sprinkler heads, movable water spraying devices) is configured for each area to ensure that the water spraying operations in each area do not interfere with each other. The preset bridge structural characteristics are the inherent structural attributes of the target bridge deck pavement, such as the distribution of joints (e.g., expansion joints, construction joints), pavement thickness, and interlayer connection methods. These characteristics directly affect the water infiltration efficiency. The distribution of water-sensitive areas is determined through preliminary visual surveys and analysis of historical maintenance records, identifying areas prone to water seepage (e.g., previously repaired areas, areas with concentrated cracks). The core logic of differentiating the water spray intensity or start time of different zones based on the preset bridge structural characteristics and the distribution of water-sensitive areas is to allocate water spray resources differently according to the regional water seepage risk level, ensuring that high-risk areas receive sufficient water replenishment to induce a water seepage response, while avoiding data interference caused by ineffective water spraying in low-risk areas.

[0100] Specifically, the spatiotemporal humidity variation data consists of a dataset continuously collected at preset time intervals by multiple humidity sensors deployed on the surface of the target bridge deck pavement during the water spraying process. This dataset includes the humidity value at the location of each sensor and the collection timestamp. This data can accurately reflect the infiltration progress and humidity change rate of water at different locations, and intuitively show the diffusion and infiltration trajectory of water on the bridge deck surface. Humidity sensors can be evenly deployed at 2-3 meter intervals in each zone, with the sensor probes close to and fixed to the pavement surface to ensure full contact with the bridge deck. The data collection interval is set to 1 minute, and the collection accuracy is ±1%RH. Sensor data recording is started synchronously after the water spraying begins, and the location coordinates, collection time, and corresponding humidity value of each sensor are stored in real time.

[0101] Specifically, the runoff dynamic image sequence is a dynamic image set formed by continuously capturing images of the target bridge deck pavement surface using an image acquisition device and arranging them in chronological order. It can intuitively present the macroscopic flow trajectory of water on the bridge deck surface, capture the runoff concentration area (potential infiltration point) and flow path, and complement the microscopic viewpoint data and macroscopic dynamics with the humidity spatiotemporal variation data. High-definition cameras can be installed at both ends and sides of the target bridge deck to ensure that the camera shooting angle can completely cover all areas, and the lens focal length is adjusted to clearly present the runoff traces on the bridge deck surface. The shooting frame rate is set to 10 frames / second and the image resolution is 1920×1080 pixels. The shooting starts synchronously after the water spraying begins, and the image files are automatically stored according to the timestamp to form a continuous runoff dynamic image sequence.

[0102] Specifically, the process of obtaining seepage path information by performing spatiotemporal correlation analysis on humidity variation data and runoff dynamic image sequences involves matching humidity changes at microscopic locations with runoff dynamics on macroscopic surfaces, using time as a unified dimension. Sensor locations where humidity rises first and increases significantly correspond to concentrated runoff confluence areas in the image sequence, which are potential infiltration points. The temporal differences in humidity changes reflect the order of water infiltration. Combined with the runoff direction in the image sequence, the migration path of water on and inside the pavement layer can be deduced, ultimately integrating to form complete seepage path information. By introducing historical dynamic response data from the bridge deck pavement structural health monitoring system and extracting structural response anomaly patterns that are temporally correlated with rainfall events or water accumulation states, key areas more sensitive to seepage can be identified before hydrological simulation. Based on the distribution of these seepage-sensitive areas, targeted deployment of zoned water spray schemes can improve the effectiveness and relevance of hydrological simulation, avoid interference from invalid or redundant test areas, thereby further improving the accuracy of seepage path information acquisition and indirectly improving the overall positioning accuracy of hidden seepage channels.

[0103] In this optional embodiment, when determining the distribution of water-sensitive areas in the target bridge deck pavement based on structural response anomaly modes, a global planar coordinate system for the bridge deck can be established using the design drawings of the target bridge deck pavement as a reference. The installation coordinates of various sensors are retrieved, and deviations are eliminated through coordinate transformation to establish a precise mapping relationship between the sensor spatial coordinates and the bridge deck planar coordinates, ensuring that the sensor position projection accuracy meets requirements. The trigger sensors corresponding to each structural response anomaly mode are organized, and the bridge deck projection coordinates of various sensors are extracted. Combined with the trigger frequency characteristics of the anomaly modes, a set of key sensors strongly correlated with structural anomalies caused by water seepage is selected. Based on the sensor type and monitoring accuracy parameters, the bridge deck monitoring coverage range (i.e., strain) of each key sensor (i.e., each core sensor) is set. The monitoring area (circular monitoring area) is delineated with the sensor coverage radius of 5 meters and the vibration sensor coverage radius of 8 meters, centered on the sensor's bridge deck projection coordinates. Overlapping monitoring areas of adjacent key sensors are merged to form a joint monitoring area. Combining the preset bridge structural mechanics model, the load transfer range of the bridge deck pavement layer corresponding to the key sensor monitoring area is analyzed to obtain the intersection of the sensor monitoring area and the load transfer range, resulting in preliminary candidate areas sensitive to seepage. Isolated candidate areas with an area smaller than the preset value (i.e., less than 10 square meters) are eliminated. Based on the bridge deck structural characteristics, the boundaries of the candidate areas are adjusted to align with the structural characteristics. Adjacent candidate areas with consistent characteristics are merged to finally generate a distribution map of seepage-sensitive areas and a visual annotation of the bridge deck.

[0104] The pre-defined bridge structural mechanics model is a specialized mechanical analysis model pre-constructed based on the design drawings of the target bridge deck pavement layer, material mechanical parameters (such as the elastic modulus of asphalt concrete surface layer, the compressive strength of cement concrete leveling layer, and the interlayer bond strength), and the overall structural form of the bridge (such as beam bridge and slab bridge). Its core is to adapt to the scene depth determined by the water seepage sensitive area of ​​the bridge deck, and to connect the sensor monitoring area with the actual water-seepage-prone area of ​​the pavement layer by quantifying the structural force transmission law. The model's input data consists of core information about the structural regions corresponding to key sensors, including the three-dimensional coordinates of the sensor-monitored structural regions (based on the transformation of the bridge deck's global planar coordinate system), sensor type (strain sensor / vibration sensor), and design parameters of the corresponding structural parts (such as pavement layer thickness, beam support type, and reinforcement density). The input data is directly related to the sensor monitoring coverage and bridge deck structural characteristics. The model's core application logic is: based on the load transfer theory in elasticity, it simulates the stress and strain transfer paths of the sensor-monitored structural regions under vehicle loads and environmental effects, and calculates the load diffusion range of the bridge deck pavement layer corresponding to that structural region (e.g., in the beam mid-span region monitored by strain sensors, the load is transferred in the pavement layer at a diffusion angle of 30°-45°, and the vibration...). The support area monitored by the dynamic sensor is transmitted at a diffusion angle of 25°-35°, clearly defining the lateral and longitudinal boundaries and depth range of load transmission. The model output data is a set of spatial coordinates of the load transmission range of the bridge deck pavement layer corresponding to the sensor-monitored structural area (adapted to the global plane coordinate system of the bridge deck, in meters). This output data can be directly intersected with the monitoring area of ​​the key sensor to ensure that the initially obtained seepage-sensitive candidate area not only covers the abnormal response range monitored by the sensor, but also conforms to the actual mechanical transmission law of the bridge deck structure. This avoids the one-sidedness of relying solely on the sensor coverage range to define the area, making the determination of seepage-sensitive areas more scientific and accurate. The model can be reproduced based on design drawings and material parameters, and the application can be adapted to the scenario according to the above input-output logic.

[0105] Specifically, the implementation steps for forming the distribution of seepage-sensitive areas include: First, coordinate system calibration and mapping relationship construction. Using the target bridge deck pavement design drawings as a reference, establish a global planar coordinate system for the bridge deck pavement (with the bridge starting point as the origin, the x-axis along the bridge's longitudinal direction, and the y-axis along the transverse direction, in meters). Simultaneously, retrieve the installation design coordinates (including longitudinal, transverse, and vertical height) of each sensor in the structural health monitoring system. Eliminate coordinate system deviations using the Affine transformation algorithm to establish a precise mapping relationship between the sensor spatial coordinates and the bridge deck pavement planar coordinates, ensuring that the projection error of the sensor position on the bridge deck plane does not exceed 3 centimeters. Second, conduct abnormal mode-sensor position correlation matching. Identify the trigger sensor numbers corresponding to each structural response abnormal mode, extract the bridge deck projection coordinates of these sensors, and combine this with the trigger frequency of the abnormal mode (e.g., if a sensor is triggered ≥5 times in an abnormal mode, it is considered a highly correlated sensor) to screen out the core sensor set strongly correlated with structural anomalies caused by seepage. Third, determine the sensor monitoring coverage range. Based on sensor type, such as strain sensors, vibration sensors, etc., and monitoring accuracy parameters, set the bridge deck monitoring coverage half-range for each core sensor. For example, the coverage radius of strain sensors is 5 meters and the coverage radius of vibration sensors is 8 meters. A circular monitoring area is delineated with the sensor's bridge deck projection coordinates as the center. If the monitoring areas of adjacent core sensors overlap, they are merged into a joint monitoring area. Then, the structural response area is mapped inversely to the bridge deck pavement layer. Combined with the bridge structural mechanics model, the load transfer range of the bridge deck pavement layer corresponding to the structural area monitored by the core sensors (e.g., mid-span of the beam, near the support) is analyzed. The intersection of the sensor monitoring area and the load transfer range is taken to obtain the preliminary candidate areas for water seepage sensitivity. The area is integrated and the boundary is calibrated. Isolated candidate areas with an area of ​​less than 10 square meters are eliminated. Based on the structural features such as bridge deck lane division, expansion joint location, and pavement layer joints, the boundaries of the candidate areas are adjusted to ensure that the boundaries are aligned with the structural features with a deviation of no more than 2 centimeters. Adjacent candidate areas with the same features are merged to form a distribution map of water seepage sensitivity areas that includes the sensitive area number, longitudinal / lateral coordinate range, area size, associated anomaly mode type, and risk level. Visual annotations on the bridge deck plane are also attached to ensure that the distribution accurately corresponds to the actual bridge deck location, providing a clear basis for differentiated water spray settings.

[0106] In the scenario of constructing a distribution of water seepage sensitive areas on the bridge deck, the core of the Affine transformation algorithm is to correct coordinate deviations through linear transformation, thereby achieving accurate mapping between the sensor spatial coordinates and the global planar coordinate system of the bridge deck pavement layer. Its application depth is adapted to the scenario requirements of coordinate system calibration and mapping relationship construction. The algorithm's input data includes two core types of information: first, the installation design coordinates of each sensor in the structural health monitoring system (including longitudinal, lateral, and vertical heights, in meters); second, the reference parameters of the global plane coordinate system of the bridge deck pavement (with the bridge starting point as the origin, x-axis along the longitudinal direction and y-axis along the lateral direction, in meters) and the correspondence between the design coordinates and actual measured coordinates of fixed marker points on the bridge deck (such as the endpoints of expansion joints and the bases of guardrail posts) (used to solve for transformation parameters). The input data comes directly from design drawings and on-site measurements, ensuring consistency with the actual spatial characteristics of the bridge deck. The core application logic of the algorithm is as follows: based on the linear characteristics of affine transformation, by matching the design coordinates with the measured coordinates of the fixed marker points on the bridge deck, transformation parameters such as translation, rotation angle, scaling ratio, and shear coefficient are solved. The sensor installation design coordinates are then substituted into the affine transformation formula (x'=a1x+ The algorithm, a2y+a3, y'=b1x+b2y+b3, where x and y are the original design coordinates of the sensors, x' and y' are the projected coordinates of the bridge deck, and a1-a3 and b1-b3 are the transformation parameters obtained by solving, completes the mapping from the spatial coordinates of the sensors to the plane coordinates of the bridge deck. Simultaneously, through multiple iterations to optimize the transformation parameters, it ensures that the projection error of the mapped sensors does not exceed 3 cm. The algorithm's output data is the precise projected coordinates of each sensor in the global plane coordinate system of the bridge deck. This output data directly supports subsequent steps such as abnormal mode-sensor location association matching and delineation of sensor monitoring coverage areas. It realizes the logical connection between sensor coordinate correction, precise positioning, and sensitive area delineation. By obtaining the design and measured coordinates of fixed marker points on the bridge deck, solving for the transformation parameters, and substituting them into the algorithm to complete the coordinate mapping, it ensures the accuracy of the entire seepage sensitive area distribution construction.

[0107] In this optional embodiment, when performing spatiotemporal correlation analysis on humidity spatiotemporal variation data and runoff dynamic image sequences to obtain seepage path information of the target bridge deck pavement layer, the humidity spatiotemporal variation data and runoff dynamic image sequences can be standardized and preprocessed separately to establish a coordinate system mapping relationship between the humidity spatiotemporal variation data and runoff dynamic image sequences, and achieve accurate spatiotemporal matching based on timestamps; the humidity change rate and humidity peak occurrence time of each humidity sensor are calculated, and humidity sensors that meet the preset screening conditions are selected; runoff regions are extracted from the runoff dynamic image sequences, the runoff direction vector is analyzed, and the confluence concentration area is identified; the physical locations of humidity sensors that meet the preset screening conditions are mapped to the runoff dynamic images, and are marked to simultaneously meet the humidity change index and peak... Potential infiltration points are identified based on their early occurrence time, location in concentrated runoff areas, and the time difference between humidity rise and runoff occurrence meeting preset requirements. A sequence is constructed based on the humidity peak time series of humidity sensors meeting preset screening criteria. This sequence is then combined with the runoff direction vector to trace the seepage source path. The migration order is determined by correlating the humidity time series differences of surrounding humidity sensors. Furthermore, the internal seepage channels of the target bridge deck pavement are supplemented based on the humidity gradient between adjacent humidity sensors. Information on potential infiltration points, seepage migration direction angles, key migration nodes, and migration speeds is integrated and verified and corrected using structural joint and porosity data from the target bridge deck pavement design drawings. Paths exceeding preset deviation ranges have their migration directions adjusted. Finally, standardized seepage path information, including text descriptions and visual path diagrams, is output.

[0108] Specifically, spatiotemporal correlation analysis of humidity variation data and runoff dynamic image sequences is performed to obtain infiltration path information, including:

[0109] Data preprocessing and spatiotemporal alignment: The humidity data is zero-point calibrated, outliers are removed based on the 3σ principle, and the sampling gap is filled by linear interpolation to obtain a standardized time-series dataset containing sensor number, sampling timestamp, and humidity value; Gaussian filtering is used to reduce noise in the runoff dynamic image sequence, and the 0.5-second inter-frame interval is unified by the inter-frame difference method to establish the mapping relationship between the image coordinate system and the physical coordinate system of the paving layer (1 pixel corresponds to 2 cm), and then the accurate correspondence between the two types of data is achieved based on the timestamp.

[0110] Key features were extracted: the humidity change rate (ΔH / Δt) and peak occurrence time of each sensor were calculated, and key sensors with a change rate greater than 0.5%RH / second and peak time in the top 30 percentile (preset screening conditions) were selected; the runoff region of each frame image was extracted by U-Net semantic segmentation algorithm, and the runoff direction vector and confluence concentration region (density greater than 50 pixels / square meter) were analyzed by Lucas-Kanade optical flow method.

[0111] The U-Net semantic segmentation algorithm, in collaboration with the Lucas-Kanade optical flow method, is adapted to the scenario of extracting hydrological features of bridge runoff. Its core lies in accurately capturing the core features of runoff through the logical connection between region segmentation and flow direction analysis to support the inference of seepage paths. The input data for the U-Net semantic segmentation algorithm is a sequence of dynamic runoff images (1920×1080 pixels, 10 frames / second). These images have been pre-processed (grayscale conversion, Gaussian filtering for noise reduction) to highlight the grayscale difference between the bridge surface background and the runoff. The model is pre-trained based on sample annotations of bridge surface runoff (light gray) and pavement background (dark gray). During training, the input image patch size is set to 512×512 pixels, the learning rate to 0.001, and the number of iterations to 500, adapting to the morphological characteristics of thin-layer runoff on the bridge surface. The algorithm extracts the edge and texture features of the runoff region through an encoder-decoder network structure, outputting a binary segmented image of runoff and background (runoff region pixel value is 255, background is 0), clearly defining the spatial range of runoff in each frame. The input data for the Lucas-Kanade optical flow method is the continuous frame binarized segmented image output by U-Net. Combined with preset parameters (window size 15×15 pixels, number of iterations 10, number of pyramid layers 3), the flow direction vector (including direction angle and velocity) is obtained by calculating the displacement vector of runoff pixels between adjacent frames. Then, the runoff pixel density per unit area (square meters) is counted to filter out the confluence concentration area with a density greater than 50 pixels / square meter. The algorithm output is the runoff direction vector matrix and the bridge surface coordinate set of the confluence concentration area. It is linked with the humidity feature data of key sensors to provide core feature support for determining the seepage path information. Based on the above input and output settings, model training parameters and algorithm parameters, the two algorithms can be directly applied in this scenario.

[0112] Spatiotemporal correlation matching: Key sensor locations are mapped to images, marking potential infiltration points with early humidity peak times and located in confluence areas, ensuring the time difference between humidity rise and runoff occurrence is less than 1 second. Then, the infiltration migration path is traced backward: A sequence is constructed based on humidity peak time, and the source path is traced using runoff direction vectors. The migration order is determined by correlating humidity time differences with surrounding sensors, and internal infiltration channels are supplemented based on humidity gradients (greater than 1%RH / cm) from adjacent sensors. Information such as potential infiltration points, migration direction angles, key nodes, and migration velocities is integrated, and verified and corrected using structural joint and porosity data from pavement design drawings (adjusting the migration direction when deviation exceeds 5 cm). Finally, standardized infiltration path information including text descriptions and visualized path diagrams is output. By zoning and spraying water onto different sections of the bridge deck pavement, and setting different spraying intensities or start times for each section, the hydrological simulation process can more closely resemble actual rainfall conditions and differences in bridge deck structure, thus stimulating the real seepage response of potential seepage channels. By combining the spatiotemporal variation data of humidity obtained from humidity sensors with the dynamic image sequence of runoff for spatiotemporal correlation analysis, the characteristics of water infiltration and surface runoff can be reflected simultaneously. This allows for a more comprehensive and intuitive determination of seepage path information, improving the precision of seepage path identification and providing reliable hydrological basis for the accurate location of hidden seepage channels.

[0113] In this optional embodiment, when determining the hidden seepage channels of the target bridge deck pavement layer based on the matching relationship between the moisture anomaly area and the seepage path information, in order to locate the hidden seepage channels of the bridge deck pavement layer, the spatial coordinate set of the moisture anomaly area and the path spatial coordinate set of the seepage path information can be spatially superimposed to determine the spatial overlap, direction consistency index, and depth correlation index between the moisture anomaly area and the seepage path information. Among them, the depth correlation index is obtained by comparing the burial depth of the moisture anomaly area with the seepage depth inferred from the seepage path information. Based on the spatial overlap, direction consistency index, and depth correlation index, the comprehensive score of each path segment in the seepage path information is determined. All path segments with comprehensive scores exceeding the preset matching threshold are combined to obtain the hidden seepage channels, so as to locate the hidden seepage channels of the bridge deck pavement layer.

[0114] Specifically, the process of spatially overlaying the spatial coordinate set of the water anomaly area with the spatial coordinate set of the path indicated by the seepage path information includes: unifying the coordinates of the two types of data (i.e., the spatial coordinate set of the water anomaly area and the spatial coordinate set of the path indicated by the seepage path information) to the same spatial reference system (e.g., the WGS84 geodetic coordinate system); and using a geographic information system platform to overlay the three-dimensional coordinate data (including planar coordinates and burial depth) of the water anomaly area with the planar coordinates of the seepage path and the inferred seepage depth data in layers, establishing the correlation between the two in terms of spatial location, extension direction, and depth dimension, so as to break down the independent storage barrier of the two types of data, provide a unified spatial benchmark for subsequent quantitative analysis of the matching degree, and avoid matching deviations caused by coordinate system differences.

[0115] Specifically, spatial overlap is the proportion of overlap between the information of the water anomaly area and the seepage path in planar space. It quantitatively reflects the degree of fit between their physical locations and is one of the core indicators for judging the matching relationship. The larger the value, the more the actual distribution of the water anomaly area matches the spatial location of the seepage path, and the more likely it is to be a real seepage channel. In the GIS layer after spatial overlay, the total length (or total area) of the water anomaly area and the length (or area) of the overlapping part can be measured respectively. The spatial overlap can be calculated by the formula "spatial overlap = overlapping length (or area) / total length of the corresponding segment of the seepage path (or total area of ​​the water anomaly area) × 100%".

[0116] Specifically, the orientation consistency index is the degree of alignment between the extension direction of the water anomaly area and the extension direction of the corresponding segment of the seepage path. It is quantified by the azimuth difference, reflecting the matching of the two in the macroscopic distribution direction. The closer the value is to 100%, the more consistent the orientations are, which conforms to the physical characteristic of the seepage channel extending along a fixed direction. In the GIS platform, the azimuth angles (i.e., the angles with due north, ranging from 0° to 360°) of the center line of the water anomaly area and the center line of the corresponding segment of the seepage path can be extracted separately, and the absolute difference between the two azimuth angles can be calculated. When the azimuth angle difference is ≤10°, the orientation consistency is set to 100%. For every 1° increase in the difference, the consistency decreases by 10%. The orientation consistency index is calculated using the formula: orientation consistency index = 100% - (absolute difference in azimuth angle × 10%) (the minimum value is 0%).

[0117] Specifically, the depth correlation index is the degree of agreement between the actual burial depth of the water anomaly area and the inferred seepage depth from the seepage path information. It quantitatively reflects the matching between the two in the vertical dimension, avoiding misjudgments caused by planar overlap but inconsistent depths. The larger the value, the more closely the burial location of the water anomaly area matches the depth range of seepage, which conforms to the law of water migration at a specific depth within the pavement layer. The burial depth range of the water anomaly area (2-5 cm) can be extracted from ground-penetrating radar data, and the seepage depth range of the seepage path (3-6 cm) can be inferred by combining hydrological simulation data and the pavement layer structural thickness. The ratio of the overlap interval length of the two depth ranges to the total length of the seepage depth range is calculated, i.e., depth correlation index = overlap interval length / total length of seepage depth range × 100%.

[0118] Specifically, the comprehensive score for each path segment is a quantitative score obtained by weighting and summing spatial overlap, directional consistency, and depth correlation indicators through preset weights. This comprehensively reflects the overall matching degree between each segment of the seepage path and the area with abnormal moisture, avoiding the one-sidedness of a single indicator and achieving a comprehensive evaluation of the matching relationship. The weights of the three indicators can be preset (based on industry experience and detection accuracy requirements, such as 40% for spatial overlap, 30% for directional consistency, and 30% for depth correlation), and the score is calculated using the formula: Comprehensive Score = Spatial Overlap × 40% + directional consistency × 30% + depth correlation × 30%, with a score range of 0-100.

[0119] Specifically, the preset matching threshold is the critical value that distinguishes between valid and invalid matches. Path segments with a comprehensive score exceeding this threshold indicate that they have a high degree of fit with the water anomaly area in terms of spatial location, direction, and depth, and are components of the real seepage channel. By combining these highly matched path segments according to spatial topological relationships (such as adjacency and connectivity), the complete direction and range of the hidden seepage channel can be restored, avoiding channel breaks or omissions caused by single segment analysis.

[0120] Specifically, for example, regarding the asphalt concrete pavement layer on the left side of a ring expressway from K12+300 to K12+500, the DGPS coordinate set of the moisture anomaly area and the coordinate set of the seepage path were first unified into the WGS84 coordinate system, and spatial overlay was completed in the geographic information system. Then, the spatial overlap, orientation consistency index, and depth correlation index of each path segment were calculated, and the comprehensive scores of the two core segments were obtained according to the preset weights, which were 86.21 and 86.71 points respectively. The preset matching threshold was set to 70 points, and these two high-scoring segments were selected. Combining their spatial adjacency and orientation consistency, they were integrated with the intermediate weak matching segments, and finally a hidden seepage channel with a total length of 59 meters and a burial depth of 2-6 centimeters was formed, starting from K12+380 in the north and ending at K12+470 in the south. This provided a clear positioning basis for the precise maintenance of the bridge. By overlaying and analyzing information on water anomaly areas and seepage paths in a unified spatial coordinate system, and comprehensively considering multi-dimensional indicators such as spatial overlap, consistency of direction, and depth correlation, the seepage path segments are quantitatively scored. This effectively avoids misjudgments caused by relying solely on a single spatial overlap relationship. Path segments with comprehensive scores exceeding a threshold are combined into hidden seepage channels, ensuring that the finally determined seepage channels simultaneously meet both geophysical anomaly characteristics and hydrological seepage characteristics. This improves the reliability and stability of the location results of hidden seepage channels in the bridge deck pavement layer.

[0121] Specifically, after step S104, the method further includes: acquiring at least one auxiliary detection data synchronously collected on the target bridge deck pavement layer, wherein the auxiliary detection data may include at least one of infrared thermal imaging data, acoustic detection data, and high-density resistivity data; performing multi-physics field coupling verification on the hidden seepage channels and the abnormal areas in the auxiliary detection data to obtain a report on the hidden seepage channels including confidence level.

[0122] Specifically, auxiliary detection data is supplementary detection data collected simultaneously on the target bridge deck pavement layer during the collection of first and second ground-penetrating radar (GPR) data and the conduct of preset hydrological simulations. This data is used to cross-verify the authenticity of hidden seepage channels. Auxiliary detection data is acquired based on different physical detection principles, forming a multi-dimensional verification with GPR data and hydrological simulation data. This helps eliminate potential misjudgments from single technical methods and improves the reliability of hidden seepage channel location results. Data can be collected in the same detection area and at the same time (ensuring a consistent detection environment) on the target bridge deck pavement layer using corresponding auxiliary detection equipment. The collection range completely overlaps with the GPR survey line and the hydrological simulation area. The collection parameters are set according to the pavement layer thickness and detection accuracy requirements. For example, for the left pavement layer of a certain ring expressway from K12+300 to K12+500, infrared thermal imaging data and high-density resistivity data were collected simultaneously while collecting GPR data and conducting water spray simulations. The collection area completely covers the identified candidate area for hidden seepage channels.

[0123] Specifically, infrared thermal imaging data is a data set that captures the infrared radiation energy of the surface and shallow layers of the pavement layer of the target bridge deck using an infrared thermal imager and converts it into a temperature distribution image. The physical principle is that the specific heat capacity of the wet cement area is greater than that of the dry area. Under the same environmental conditions, such as sunlight and ventilation, areas with abnormal moisture will show lower temperature characteristics, forming obvious temperature anomaly zones. An infrared thermal imager with a resolution of not less than 640×480 pixels can be selected. During sunny weather and periods without strong direct sunlight, such as early morning or evening, thermal imaging images are collected from the side or top of the bridge deck at an angle of 30°-45°, with a density of one image every 5 meters, to ensure coverage of the entire detection area. Acoustic wave detection data is a data set obtained by transmitting elastic sound waves into the pavement layer using acoustic wave detection equipment, receiving reflected or transmitted wave signals, and recording parameters such as sound wave propagation speed and amplitude attenuation. The physical principle is that moisture reduces the elastic modulus of the pavement medium, causing the sound wave propagation speed in the cement-containing wet channel to be lower than in the dry area, and the amplitude attenuation to be more significant, resulting in an abnormal sound wave response. A portable acoustic wave detector can be used, with the sound wave transmission frequency set to 20-50kHz. Detection points are simultaneously deployed along the ground-penetrating radar survey line at a distance of 2 meters. The transmitting probe and receiving probe are placed on both sides of the detection point to collect sound wave propagation time and amplitude data. High-density resistivity data is obtained by applying a weak current to the interior of the pavement layer using a high-density resistivity meter to measure the resistivity distribution data at different depths. The physical principle is that moisture increases the conductivity of the pavement medium, and the resistivity of the wet cement-containing area is much lower than that of the dry area, forming an abnormally low resistivity zone. A high-density resistivity meter can be used, with the electrode spacing set to 1 meter. An electrode array is arranged along the detection area, a stable weak current (5mA) is applied, the potential difference between the electrodes is measured, and the resistivity values ​​at different depths are calculated.

[0124] Specifically, the core logic of multi-physics field coupling verification of hidden seepage channels and abnormal areas in auxiliary detection data is that hidden seepage channels, as cement-containing wet interconnected structures within the pavement layer, will produce characteristic responses in different physical fields. In the electromagnetic wave field (ground-penetrating radar), they exhibit dielectric constant anomalies; in the temperature field (infrared thermal imaging), they exhibit temperature anomalies; in the elastic wave field (acoustic wave detection), they exhibit propagation velocity anomalies; and in the electric field (high-density resistivity), they exhibit resistivity anomalies. Coupling verification involves spatial matching and physical parameter correlation analysis between the spatial distribution of hidden seepage channels and abnormal areas in various auxiliary detection data. If different physical fields show anomalies... If the anomalous areas of the physical field are highly concentrated in the same spatial range and the physical parameter responses conform to the common laws of moisture presence, then the authenticity of the hidden seepage channels can be verified; otherwise, the possibility of misjudgment needs to be ruled out. The spatial coordinates (planar coordinates plus burial depth) of the hidden seepage channels and the coordinates of the anomalous areas of each auxiliary detection data can be unified into the same coordinate system, such as the WGS84 geodetic coordinate system, and spatially superimposed. Then, the spatial overlap ratio and depth consistency between the hidden seepage channels and each anomalous area should be analyzed. Finally, it should be verified whether the anomalous parameters of each auxiliary detection data conform to the physical characteristics of moisture, such as decreased temperature, decreased wave velocity, and decreased resistivity, and the verification results should be comprehensively judged.

[0125] Specifically, the confidence level is a grading index that quantifies the reliability of the location results of hidden seepage channels after multi-physics field coupling verification. It is used to intuitively reflect the credibility of the location results and provide a reference for subsequent maintenance decisions. The confidence level can be determined based on the types of auxiliary detection data involved in the verification, the spatial overlap of abnormal areas, and the correlation of physical parameters. It is generally divided into three levels: high confidence (complete verification), medium confidence (partial verification), and low confidence (unverified). Confidence judgment rules can be preset. For example, if the abnormal areas of two or more auxiliary detection data highly overlap with the hidden seepage channels (overlap greater than or equal to 80%), and the physical parameters meet the moisture content characteristics, it is judged as high confidence; if one auxiliary detection data meets the above conditions, it is judged as medium confidence; if no auxiliary detection data meets the conditions, it is judged as low confidence. The level is determined according to the rules of the coupling verification results.

[0126] Specifically, the hidden seepage channel report is a standardized and actionable technical document that integrates the location results, multiphysics coupling verification results, and confidence levels from steps S101 to S104. The report clearly defines the core information of the hidden seepage channel (location, extent, depth), verification process, confidence level, and targeted maintenance recommendations, providing bridge maintenance units with clear and accurate decision-making support. It can also organize all data from the detection process (ground radar data, hydrological simulation data, auxiliary detection data), analyze the results (abnormal moisture areas, seepage paths, matching relationships, verification...), and verify... The report should be written according to the structure of detection overview - location process - verification results - confidence level - maintenance recommendations, and should include relevant images (radar images, anomaly maps, overlay maps) and data tables. For example, the report on hidden seepage channels should clearly record the starting and ending coordinates of the channel (K12+380-K12+470), length of 50 meters, width of 1.2-1.5 meters, and burial depth of 2-6 centimeters. It should also describe in detail the matching process of ground-penetrating radar and hydrological simulation, the verification process of infrared and high-density resistivity, mark the confidence level as high confidence, and propose maintenance recommendations of grouting sealing and local milling and repaving. After the initial location of hidden seepage channels is completed, at least one auxiliary detection data such as infrared thermal imaging, acoustic wave detection, or high-density resistivity is introduced to verify the identified hidden seepage channels through multi-physics field coupling. This allows for cross-verification of seepage channels from different physical response perspectives, reducing the uncertainty brought by a single technology. By generating a report on hidden seepage channels that includes confidence levels, the location results are made more intuitive and credible, providing a reliable basis for subsequent maintenance decisions of the bridge deck pavement layer.

[0127] Specifically, multiphysics coupling verification is performed on hidden seepage channels and abnormal areas in auxiliary detection data to obtain a hidden seepage channel report including confidence levels. This includes: spatially registering the spatial distribution model of the hidden seepage channels with the abnormal distribution model generated from the auxiliary detection data in the same coordinate system to obtain the spatial consistency and physical parameter correlation between the hidden seepage channels and each auxiliary detection abnormal area; determining the confidence level of the hidden seepage channels based on the spatial consistency and physical parameter correlation; and generating a hidden seepage channel report based on the confidence level.

[0128] Both the spatial distribution model of the hidden seepage channels and the anomaly distribution model generated from auxiliary detection data are digital 3D models adapted to multi-physics field coupling verification scenarios and supporting the determination of the confidence level of hidden seepage channels. They achieve coupled verification logic through a unified coordinate system and precise correlation. The spatial distribution model of the hidden seepage channels is based on the design drawings of the target bridge deck pavement layer. The input data consists of the core parameters of the hidden seepage channels determined in step S104 (including planar coordinates measured based on differential GPS, burial depth range, length, and orientation angle detected by ground-penetrating radar, all in meters). Geographic Information System (GIS) modeling software is used for polygon drawing and depth assignment, outputting a 3D model reflecting the spatial location and shape of the channels within the pavement layer, accurately reproducing linear or sheet-like distribution characteristics. The anomaly distribution model generated from auxiliary detection data targets auxiliary data such as infrared thermal imaging, acoustic wave detection, and high-density resistivity. The input data consists of raw data collected by various auxiliary detection devices (such as the temperature matrix of infrared thermal imaging, wave velocity data from acoustic wave detection, and resistance value data from resistivity meters). Anomalies related to moisture are extracted through threshold segmentation (e.g., infrared low temperature threshold below 3°C and acoustic low wave velocity threshold below 2500m / s). After obtaining the planar coordinates, range, and depth information of the anomaly areas, they are imported into the same modeling platform to construct anomaly 3D models of the corresponding physical fields. Both models are unified to the WGS84 geodetic coordinate system and spatially registered through fixed marker points on the bridge deck (e.g., expansion joint endpoints). The accurate spatial coordinates and morphological data output directly support the subsequent calculation of spatial consistency (overlap, alignment) and physical parameter correlation (the degree of fit between anomaly parameters and moisture characteristics). Based on the above input parameters, modeling logic, and coordinate system requirements, two types of models can be accurately constructed and coupled verification can be carried out to ensure the scientific nature of confidence level determination and report generation.

[0129] Specifically, the spatial distribution model of the hidden seepage channels is a digital and visualized three-dimensional spatial model constructed based on the core information of the hidden seepage channels (including planar coordinates, burial depth range, length, width, and orientation) determined in step S104. The spatial distribution model accurately restores the spatial location and morphological characteristics of the hidden seepage channels within the target bridge deck pavement layer, serving as a benchmark for subsequent coupling verification with auxiliary detection of abnormal areas. The boundary coordinates (based on measured data from the differential GPS) and burial depth data (based on ground-penetrating radar detection results) of the hidden seepage channels can be imported into a geographic information system or 3D modeling software. Through operations such as polygon drawing, depth assignment, and morphological fitting, a spatial distribution model containing planar distribution and vertical depth information can be generated.

[0130] Specifically, the anomaly distribution model generated from auxiliary detection data is constructed by processing infrared thermal imaging data, acoustic wave detection data, and high-density resistivity data respectively, extracting anomalous areas of physical parameters related to moisture anomalies, and building corresponding digital spatial models. Based on its detection principle, the anomaly distribution model reflects the anomalous distribution of specific physical fields, forming a multi-dimensional comparison with the spatial distribution model of hidden seepage channels. For various types of auxiliary detection data, threshold segmentation and anomaly recognition algorithms can be used to extract anomalous areas that conform to moisture-containing characteristics, such as the low-temperature region of infrared thermal imaging, the low-wave velocity region of acoustic wave detection, and the low-resistivity region of high-density resistivity, to obtain anomalies. The planar coordinates, extent, and depth information of the normal area (if detectable) are then imported into the same modeling platform to construct corresponding anomaly distribution models. For example, for the infrared thermal imaging data synchronously collected from the bridge deck, low-temperature anomaly areas with temperatures more than 3°C lower than the surrounding area are extracted through threshold segmentation, and their planar coordinates (K12+378-K12+472) and width of 1.3-1.6 meters are obtained to construct an infrared anomaly distribution model. For high-density resistivity data, low-resistivity anomaly areas with resistivity below 1000Ω・m are extracted, and corresponding resistivity anomaly distribution models are constructed. Both types of models are highly correlated with the spatial extent of hidden seepage channels.

[0131] Both the infrared anomaly distribution model and the resistivity anomaly distribution model are adapted for multi-physics field coupling verification scenarios of hidden water seepage channels in bridge deck pavement. Based on corresponding auxiliary detection data, they are digital 3D spatial models constructed, primarily used to verify the authenticity of hidden water seepage channels from different physical dimensions. The input data for the infrared anomaly distribution model consists of synchronously acquired infrared thermal imaging data (resolution no less than 640×480 pixels, acquired during early morning or evening without strong sunlight). Combining this with the physical characteristic that the specific heat capacity of the cement-containing wet area of ​​the bridge deck pavement is greater than that of the dry area, a threshold segmentation algorithm is used to extract low-temperature anomaly areas with temperatures 3°C or higher below the surrounding area. The planar coordinates, range, and depth mapping information of this area based on the bridge deck's global coordinate system (inferred from the pavement thickness) are obtained. This information is then imported into a geographic information system modeling platform to construct a 3D model, outputting spatial distribution characteristics that intuitively reflect the temperature anomalies in areas of moisture accumulation. The input data for the resistivity anomaly distribution model consists of raw resistivity data collected by a high-density resistivity meter (electrode spacing 1 meter, using a Wenner device for electrode placement). The model utilizes moisture... The physical laws that enhance the conductivity of the pavement medium are investigated. An anomaly identification algorithm is used to extract low-resistivity anomaly areas with resistivity below 1000 Ω·m, accurately obtaining their planar coordinates, extension range, and corresponding pavement layer depth information. A three-dimensional model is constructed in the same modeling platform, outputting anomaly spatial distribution characteristics that reflect the enhanced conductivity caused by moisture. Both types of models are unified to a coordinate system consistent with the spatial distribution model of hidden seepage channels (such as the WGS84 geodetic coordinate system). Spatial calibration is completed through fixed marker points on the bridge deck. The spatial data of the anomaly areas output by the model accurately compares with the spatial distribution of hidden seepage channels. Based on the above input parameters, processing algorithms, modeling logic, and coordinate system requirements, the two types of models can be quickly constructed and coupled for verification, providing reliable support for the determination of the confidence level of hidden seepage channels.

[0132] Specifically, the threshold segmentation algorithm of this invention is a dual-threshold segmentation algorithm adapted to the auxiliary detection scenario of bridge deck seepage, and the anomaly identification algorithm is a multi-feature anomaly detection algorithm based on K-means clustering. The two algorithms work together to accurately extract the moisture anomaly area in the auxiliary detection data. The core architecture, the integration method with the scenario, and the input and output settings of its application in the bridge deck seepage detection scenario in the field of ground-penetrating radar detection technology are as follows: The dual-threshold segmentation algorithm matches the moisture anomaly feature threshold corresponding to bridge deck seepage to the differences in physical characteristics of three types of data: infrared thermal imaging, acoustic detection, and high-density resistivity. The input of this algorithm is the original detection data matrix of various types after preprocessing (denoising of infrared thermal imaging data, wave velocity calibration of acoustic detection data, and electrode error correction of high-density resistivity data). Each element in the matrix is ​​bound to the global planar coordinates of the bridge deck (x-axis along the longitudinal direction of the bridge, y-axis along the transverse direction, in meters), and corresponds to physical parameters such as infrared temperature value, acoustic wave velocity value, and high-density resistivity value. The algorithm architecture consists of four core processes. The first step is to statistically analyze the physical parameters based on the various detection data of the dry area of ​​the bridge deck. The algorithm first sets a baseline value, such as the baseline temperature value obtained by statistically analyzing the average temperature of the dry area in infrared thermal imaging, the baseline wave velocity value obtained by statistically analyzing the average wave velocity of the dry area in acoustic detection, and the baseline resistance value obtained by statistically analyzing the average resistance value of the dry area in high-density resistivity. The second step combines the physical characteristics of moisture anomalies to set upper and lower thresholds for dual thresholds. For example, infrared thermal imaging uses the baseline temperature value of -3℃ as the lower threshold and the baseline temperature value as the upper threshold; acoustic detection uses the baseline wave velocity value of -200m / s as the lower threshold and the baseline wave velocity value as the upper threshold; and high-density resistivity uses the baseline resistance value of -500Ω・m as the lower threshold and the baseline resistance value as the upper threshold. This accurately defines the low temperature, low wave velocity, and low resistance anomaly range corresponding to water seepage on the bridge deck. The third step performs dual threshold segmentation on the detection data matrix, marking pixels / data points within the threshold range as suspected anomalies. The fourth step introduces morphological closing operations to fill small holes in the suspected anomaly area and opening operations to remove isolated micro-anomalies, adapting to the local discontinuities that may exist in the water seepage area of ​​the bridge deck pavement. The algorithm finally outputs the plane coordinate matrix and parameter range of the initial screening suspected anomaly area for various detection data.A multi-feature anomaly detection algorithm based on K-means clustering is used as the fine screening step. The input is the initial screening data of suspected anomaly areas from a dual-threshold segmentation algorithm, and structural feature data (locations of expansion joints, structural joints, and lane joints) from the bridge deck design drawings are imported as auxiliary verification criteria. The core architecture of this algorithm includes: First, extracting multi-dimensional features of the initially screened anomaly areas, including the mean of physical parameters, such as average infrared temperature, average acoustic velocity, area, morphological contour, and spatial distance from the bridge deck structural joints. Second, setting the number of clusters to 2 (normal cluster and moisture anomaly cluster), and performing cluster analysis on the extracted features using the K-means algorithm to calculate the clustering of each initially screened area with the values ​​in each cluster. The algorithm first calculates the Euclidean distance from the center of the water anomaly cluster. The third step quantifies the anomaly degree, classifying areas closer to the center as high-anomaly regions. The fourth step combines bridge deck structural features for auxiliary verification, eliminating non-water-related anomaly regions caused by structural joints or material differences. For example, low resistivity areas at structural joints are not necessarily caused by seepage. The algorithm ultimately outputs precise data on water anomaly regions after careful screening, including planar coordinate range, physical parameter anomalies, and region depth information (when detectable). This output data is fully compatible with the coordinate system of the modeling platform and can be directly imported to construct anomaly distribution models corresponding to various auxiliary detection methods, achieving accurate extraction and model construction of anomaly regions in bridge deck seepage auxiliary detection data.

[0133] Specifically, the process of spatially registering the spatial distribution model of the hidden seepage channel with the anomaly distribution model generated from auxiliary detection data in the same coordinate system includes: aligning the coordinate systems of all models to the same reference, and then using coordinate calibration and position fine-tuning to ensure precise spatial alignment between the hidden seepage channel model and each auxiliary anomaly distribution model, eliminating spatial deviations caused by positioning errors of the detection equipment. This lays the foundation for consistency and correlation in subsequent quantitative analysis, achieving spatial unification of data from different physical fields, and ensuring the accuracy and effectiveness of comparing channel and anomaly locations. The calibration coordinates of the differential global positioning system can be used as a reference to adjust the origin and azimuth of the hidden seepage channel model and each auxiliary anomaly distribution model. By selecting fixed marker points on the bridge deck, such as the endpoints of expansion joints and the bases of guardrail posts, as registration control points, the coordinate deviation between each model and the control points is calculated, and translation and rotation corrections are performed to ensure that the coordinate error of all models at the control points is less than 0.1 meters.

[0134] Specifically, spatial consistency refers to the degree of fit between the spatial distribution model of the hidden seepage channel and the distribution models of each auxiliary detection anomaly after spatial registration, in terms of planar range, extension direction, and depth range. It is the core indicator for quantifying the spatial relationship between the two. The higher the value, the more the physical location of the hidden seepage channel overlaps with the anomaly area identified by the auxiliary detection, indirectly confirming the authenticity of the channel. In the registered modeling platform, the planar overlap (overlap area / total area of ​​the hidden seepage channel × 100%) and the direction matching (azimuth angle of the center lines of the two) between the hidden seepage channel model and each auxiliary anomaly model can be calculated respectively. A difference of less than or equal to 10° is considered a perfect match, and the match rate decreases by 10% for every 1° increase in the difference. The spatial consistency score (out of 100 points) is obtained by combining the three indicators: the planar overlap of the hidden seepage channel with the infrared anomaly model is 92%, the orientation match is 98%, and the depth overlap is 83%. For example, in the above registered model analysis: the planar overlap of the hidden seepage channel with the infrared anomaly model is 92%, the orientation match is 98%, and the depth overlap is 83%, with a spatial consistency score of 91 points; the planar overlap with the resistivity anomaly model is 90%, the orientation match is 97%, and the depth overlap is 87%, with a spatial consistency score of 92 points.

[0135] Specifically, the correlation of physical parameters refers to the degree of agreement between the physical characteristics of moisture content corresponding to hidden seepage channels and the physical parameter change patterns reflected by various auxiliary detection anomaly distribution models. The core logic is that moisture intrusion leads to specific changes in the physical parameters of the pavement medium, such as decreased temperature, slower sound wave velocity, and decreased resistivity. If the parameter changes in the auxiliary anomaly models are consistent with this pattern, a strong correlation is formed; otherwise, the correlation is weak. The correlation patterns of moisture parameters for various auxiliary detection methods can be clearly identified. For example, infrared thermal imaging shows a decrease in temperature, acoustic detection shows a decrease in propagation velocity, and high-density resistivity shows a decrease in resistivity. This can be compared with the core parameters of the auxiliary anomaly models, such as the temperature in the infrared anomaly area. The correlation between temperature differences, wave velocity differences in the acoustic anomaly zone, and resistance differences in the resistivity anomaly zone is evaluated to determine if they conform to this pattern. The significance of these parameter changes is then quantified. For example, a temperature difference greater than or equal to 3℃ indicates a strong correlation, 2-3℃ a moderate correlation, and less than 2℃ a weak correlation. This yields a physical parameter correlation score (out of 100). For instance, in the analysis of the aforementioned auxiliary anomaly model: the low-temperature zone of the infrared anomaly model is 3-5℃ lower than the surrounding temperature, consistent with the temperature decrease caused by moisture, resulting in a parameter correlation score of 95. The low-resistivity zone of the resistivity anomaly model is 700-1200 Ω·m lower than the surrounding resistivity, consistent with the increased conductivity (decreased resistivity) caused by moisture, resulting in a parameter correlation score of 93.

[0136] Specifically, the core logic for determining the confidence level of hidden seepage channels based on spatial consistency and physical parameter correlation is that the confidence level is a comprehensive quantitative classification of the reliability of the location results of hidden seepage channels. Spatial consistency reflects location matching, while physical parameter correlation reflects principle consistency. Combining the two can comprehensively eliminate misjudgments based on a single technology, ensuring the scientific nature of the level determination. The weights of the two indicators can be preset, for example, spatial consistency weight 50% and physical parameter correlation weight 50%, and the comprehensive verification score corresponding to each type of auxiliary anomaly model can be calculated (comprehensive score = spatial consistency score × 50% + physical parameter correlation score × 50%). Then, based on... The confidence level is determined based on the types of auxiliary models used in the verification and the overall scores (e.g., a score of 85 or higher for two or more auxiliary models indicates high confidence, a score of 85 or higher for one auxiliary model indicates medium confidence, and no auxiliary model indicates low confidence). The final confidence level is then determined according to these rules. For example, in the above verification analysis: the overall score of the infrared anomaly model is (91×50%+95×50%)=93, and the overall score of the resistivity anomaly model is (92×50%+93×50%)=92.5. Both models score 85 or higher, therefore the confidence level of the hidden seepage channel is determined to be high confidence.

[0137] Specifically, the core logic for generating a report on hidden seepage channels based on confidence levels is to integrate key data, analysis results, and verification conclusions from the entire detection process, using the confidence level as the core support, to form a standardized and implementable technical document. This provides a clear and reliable decision-making basis for bridge maintenance. The document can be written according to a structured framework of detection overview - location process - coupled verification results - confidence level - maintenance recommendations. It integrates the target bridge deck foundation information (location, pavement structure), location data from ground-penetrating radar and hydrological simulation, auxiliary detection data, and model registration analysis results, along with relevant images (radar difference maps, model registration data, etc.). The report includes reference maps, abnormal parameter curves, and data tables, clearly indicating the confidence level and judgment criteria, and finally forming a complete report. For example, the report on the hidden seepage channels on the bridge deck mentioned above records the channel location (K12+380-K12+470), dimensions (length 50 meters, width 1.2-1.5 meters, burial depth 2-6 centimeters), positioning process (dry / wet GPR comparison plus hydrological path matching), coupling verification details (infrared plus resistivity dual model registration analysis), clearly indicating the confidence level as high confidence, and proposing maintenance suggestions such as grouting and sealing along the channel direction and local milling and repaving of key areas.

[0138] Figure 2 An embodiment of the bridge deck pavement concealed seepage channel positioning system of the present invention is shown.

[0139] In this optional embodiment, the bridge deck pavement layer concealed seepage channel locating system includes:

[0140] Data acquisition unit 201 is used to acquire first ground-penetrating radar detection data of the target bridge deck pavement layer under preset dry conditions and second ground-penetrating radar detection data under preset wet conditions;

[0141] The area determination unit 202 is used to determine the abnormal moisture area inside the pavement layer of the target bridge deck based on the detection data of the first ground-penetrating radar and the detection data of the second ground-penetrating radar.

[0142] The path determination unit 203 is used to perform simulation analysis on the target bridge deck pavement layer using a preset hydrological simulation, obtain the corresponding hydrological simulation data, and determine the seepage path information of the target bridge deck pavement layer based on the hydrological simulation data.

[0143] The channel determination unit 204 is used to determine the hidden seepage channels of the target bridge deck pavement layer based on the matching relationship between the abnormal moisture area and the seepage path information, so as to realize the positioning of the hidden seepage channels of the bridge deck pavement layer.

[0144] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database stores static and dynamic information data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements the steps in the above method embodiments.

[0145] Those skilled in the art will understand that Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the computer device to which the present invention is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0146] In addition, the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0147] In addition, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in the above method embodiments.

[0148] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage, etc. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc.

[0149] This invention is not limited to the structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this invention is limited only by the appended claims.

Claims

1. A method for locating hidden seepage channels in bridge deck pavement, characterized in that, include: Acquire first ground-penetrating radar detection data of the target bridge deck pavement layer under preset dry conditions and second ground-penetrating radar detection data under preset wet conditions; Based on the detection data from the first and second ground-penetrating radars, the area of ​​abnormal moisture inside the pavement layer of the target bridge deck was identified. The target bridge deck pavement layer is simulated and analyzed using a preset hydrological simulation to obtain the corresponding hydrological simulation data, and the seepage path information of the target bridge deck pavement layer is determined based on the hydrological simulation data. Based on the matching relationship between areas of abnormal moisture and seepage path information, the hidden seepage channels of the target bridge deck pavement layer are identified, so as to realize the location of the hidden seepage channels of the bridge deck pavement layer.

2. The method for locating hidden seepage channels in bridge deck pavement layer according to claim 1, characterized in that, The determination of abnormal moisture areas within the target bridge deck pavement layer based on the first and second ground-penetrating radar detection data includes: Imaging processing was performed on the first and second ground-penetrating radar detection data respectively to obtain the first radar image sequence and the second radar image sequence. The second radar image sequence is compared with the first radar image sequence to obtain a radar response difference map. Extract connected regions from radar response difference maps where the signal strength exceeds a preset strength threshold and the spatial continuity satisfies a preset continuity threshold. Based on the spatial topological relationship of the connected regions, a seepage channel network is formed, and the channels that are distributed in a linear or sheet-like manner in the seepage channel network are identified as abnormal moisture areas inside the target bridge deck pavement layer.

3. The method for locating hidden seepage channels in bridge deck pavement layer according to claim 2, characterized in that, The process of connecting the seepage channels into a network based on the spatial topology of the connected regions includes: Obtain the parameters of the connected regions, verify the parameters of the connected regions, remove connected regions with abnormal parameters, unify the range of values ​​for the direction angle, and divide the depth interval according to the thickness of the target bridge deck pavement layer, so that all connected region parameters meet the spatial characteristic requirements of bridge deck seepage detection. Calculate the planar distance between the center points of any two connected regions, quantify the orientation similarity between each connected region, and determine the depth fit between regions by the overlap of depth intervals, so as to complete the multi-dimensional quantitative analysis of the spatial topological relationship of connected regions. Based on the actual distribution pattern of seepage channels in the target bridge deck pavement, scenario-based topological association judgment conditions are set, and the topological association between connected regions is judged according to the plane spacing, direction similarity, and depth adaptability. At the same time, the structural features in the design drawings of the target bridge deck pavement are used for auxiliary verification, and the topological association of connected regions that meet the scenario-based association characteristics is further judged. For connected regions that are determined to be topologically related, draw connecting line segments with the nearest point on the region boundary as the endpoint to form a preliminary network structure of seepage channels; Redundant connections are removed and boundaries are smoothed in the initial network structure, and finally structured seepage channel network data is output.

4. The method for locating hidden seepage channels in bridge deck pavement layer according to claim 1, characterized in that, The step of using a preset hydrological simulation to simulate and analyze the target bridge deck pavement layer, obtaining corresponding hydrological simulation data, and determining the seepage path information of the target bridge deck pavement layer based on the hydrological simulation data includes: Acquire dynamic response data of the target bridge deck pavement layer within a preset historical period; Extract structural response anomaly patterns from dynamic response data that are temporally correlated with rainfall events or bridge water accumulation. Based on the structural response anomaly pattern, the distribution of water-sensitive areas in the target bridge deck pavement layer was determined; The surface of the target bridge deck pavement was sprayed with water in different areas, and the spatiotemporal variation data of humidity and the dynamic image sequence of runoff on the surface of the target bridge deck pavement were obtained respectively. Spatiotemporal correlation analysis was performed on humidity spatiotemporal variation data and runoff dynamic image sequences to obtain the seepage path information of the target bridge deck pavement layer.

5. The method for locating hidden seepage channels in bridge deck pavement layer according to claim 4, characterized in that, The method of determining the distribution of water-sensitive areas in the target bridge deck pavement layer based on structural response anomaly patterns includes: Based on the design drawings of the target bridge deck pavement layer, a global plane coordinate system for the bridge deck is established, the installation coordinates of various sensors are retrieved, and the deviation is eliminated through coordinate transformation. A precise mapping relationship between the spatial coordinates of the sensors and the plane coordinates of the bridge deck is established to ensure that the accuracy of the sensor position projection meets the requirements. We organized the triggering sensors corresponding to the abnormal response modes of each structure, extracted the bridge deck projection coordinates of various sensors, and combined them with the triggering frequency characteristics of the abnormal modes to screen out the set of key sensors that are strongly correlated with structural anomalies caused by water seepage. Based on the sensor type and monitoring accuracy parameters, the bridge deck monitoring coverage range of each key sensor is set, the monitoring area is delineated with the sensor bridge deck projection coordinates as the center, and the overlapping monitoring areas of adjacent key sensors are merged to form a joint monitoring area. Based on the pre-set bridge structural mechanics model, the load transfer range of the bridge deck pavement layer corresponding to the key sensor monitoring area is analyzed, and the intersection of the sensor monitoring area and the load transfer range is obtained to obtain preliminary candidate areas for water seepage sensitivity. Isolated candidate regions with areas smaller than a preset value are removed. Based on the structural features of the bridge deck, the boundaries of the candidate regions are adjusted and aligned with the structural features. Adjacent candidate regions with consistent features are merged, and finally, a distribution map of water-sensitive areas and visual annotations of the bridge deck are generated.

6. The method for locating hidden seepage channels in bridge deck pavement layer according to claim 4, characterized in that, The spatiotemporal correlation analysis of humidity spatiotemporal variation data and runoff dynamic image sequences yields the following information on the seepage path of the target bridge deck pavement layer: Standardized preprocessing was performed on the spatiotemporal variation data of humidity and the dynamic image sequence of runoff, respectively, to establish the coordinate system mapping relationship between the spatiotemporal variation data of humidity and the dynamic image sequence of runoff, and to achieve accurate spatiotemporal matching based on timestamps; Calculate the humidity change rate and peak humidity occurrence time of each humidity sensor, and select humidity sensors that meet the preset screening conditions; extract runoff regions from the runoff dynamic image sequence, analyze the runoff direction vector, and identify the confluence concentration area; The physical locations of humidity sensors that meet the preset screening criteria are mapped onto the runoff dynamic image, and potential infiltration points that simultaneously meet the following criteria are marked: humidity change index meets the standard, peak occurrence time is earlier, they are located in the confluence area, and the time difference between humidity increase and runoff occurrence meets the preset requirements. A sequence is constructed based on the humidity peak time series of humidity sensors that meet the preset screening conditions. The seepage source path is traced by combining the runoff direction vector. The migration order is determined by associating the humidity time series differences of surrounding humidity sensors. The seepage channels inside the target bridge deck pavement layer are supplemented based on the humidity gradient between adjacent humidity sensors. By integrating information on potential infiltration points, seepage migration direction angles, key migration nodes, and migration speeds, and verifying and correcting the data on structural joints and porosity in the design drawings of the target bridge deck pavement layer, the migration direction of paths exceeding the preset deviation range is adjusted, and finally, standardized seepage path information containing text descriptions and visual path diagrams is output.

7. The method for locating hidden seepage channels in bridge deck pavement layer according to claim 1, characterized in that, The method of determining the hidden seepage channels in the target bridge deck pavement layer based on the matching relationship between abnormal moisture areas and seepage path information, in order to locate the hidden seepage channels in the bridge deck pavement layer, includes: The spatial coordinate set of the water anomaly area is spatially superimposed with the path spatial coordinate set of the seepage path information to determine the spatial overlap, direction consistency index, and depth correlation index between the water anomaly area and the seepage path information. Based on spatial overlap, directional consistency index, and depth correlation index, the comprehensive score of each path segment in the seepage path information is determined. All paths with comprehensive scores exceeding the preset matching threshold are segmented and combined to obtain hidden seepage channels, thereby enabling the location of hidden seepage channels in the bridge deck pavement layer.

8. A bridge deck pavement layer concealed seepage channel positioning system, characterized in that, include: The data acquisition unit is used to acquire first ground-penetrating radar detection data of the target bridge deck pavement layer under preset dry conditions and second ground-penetrating radar detection data under preset wet conditions. The area determination unit is used to determine the area of ​​abnormal moisture inside the pavement layer of the target bridge deck based on the detection data of the first and second ground-penetrating radars. The path determination unit is used to simulate and analyze the target bridge deck pavement layer using a preset hydrological simulation, obtain the corresponding hydrological simulation data, and determine the seepage path information of the target bridge deck pavement layer based on the hydrological simulation data. The channel determination unit is used to determine the hidden seepage channels of the target bridge deck pavement layer based on the matching relationship between the abnormal moisture area and the seepage path information, so as to realize the positioning of the hidden seepage channels of the bridge deck pavement layer.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

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

Images