Geological surveying and mapping image processing method based on real-time dynamic positioning

Abstract

The invention discloses a geological surveying and mapping image processing method based on real-time dynamic positioning, and relates to the technical field of tunnel lining image data processing, and the method comprises the steps: obtaining the cavity, leakage and crack confidence data of a target lining point, and screening a first to-be-measured point through a set threshold value; calculating a confidence coefficient difference value between the cavity and the leakage, delimiting a first region and verifying crack confidence coefficient distribution; adjusting the confidence coefficient according to the area ratio of the second region to the first region, and generating a second to-be-measured point; optimizing confidence data of the third to-be-measured point through Euclidean distance judgment and a vector adjustment coefficient; and finally outputting a defect diagnosis report. And meanwhile, a crack-cavity symbiont is identified through morphological symbiosis analysis, the confidence coefficient is enhanced, and leakage path tracing is realized by selecting a path tracking mode based on a spatial distribution form. According to the method, the accuracy of tunnel lining defect diagnosis and the reliability of cause analysis are remarkably improved through multiple rounds of space verification and confidence iterative optimization.

Description

Technical Field

[0001] This invention relates to the field of tunnel lining image data processing technology, and more specifically, this application relates to a geological mapping image processing method based on real-time dynamic positioning. Background Technology

[0002] In the field of structural health monitoring in tunnel engineering, especially for the identification and diagnosis of typical defects such as voids, leaks, and cracks behind the lining, existing technologies mainly rely on threshold judgments based on single sensor data or simple fusion data. These methods classify defects based solely on the characteristic intensity of isolated data points, ignoring the inherent correlation and evolutionary patterns of defects in spatial distribution. Consequently, the reliability and accuracy of diagnostic results in real-world complex engineering scenarios cannot meet the requirements of precise maintenance.

[0003] Specifically, existing technologies typically analyze each detection point independently when processing multi-source detection data, directly determining the defect type using preset fixed thresholds. This method has significant limitations: firstly, it cannot effectively identify local anomalies caused by sensor noise, surface contamination, or temporary environmental factors, easily misclassifying isolated noise points as structural defects—for example, incorrectly identifying local surface depressions as underlying cavities, or misclassifying condensation marks as continuous leakage; secondly, due to the lack of comprehensive analysis of the spatial distribution patterns of defects, existing technologies struggle to reveal the intrinsic connections between different defect types. For instance, they cannot distinguish between true underlying cavities and erosion areas formed by long-term leakage, nor can they identify the replenishment relationship between surface leakage points and underlying cavities, let alone determine the symbiotic characteristics of cracks and cavities in terms of mechanical mechanisms. These limitations of isolated analysis directly affect the engineering guidance value of the diagnostic conclusions.

[0004] To address the aforementioned issues, there is an urgent need in this field for a diagnostic method that can fully consider the spatial correlation characteristics of defects and possess a multi-round self-verification mechanism, in order to overcome the inherent defects of traditional single-point threshold judgment. Traditional tunnel lining defect diagnosis methods suffer from high misjudgment rates and inaccurate causal diagnosis when dealing with complex working conditions due to the lack of comprehensive analysis of the spatial correlation mechanism of multiple defects. Summary of the Invention

[0005] To address the aforementioned technical problems, a geological mapping image processing method based on real-time dynamic positioning is provided. This technical solution solves the problems mentioned in the background section.

[0006] In a first aspect, embodiments of this application provide a geological mapping image processing method based on real-time dynamic positioning, characterized by the following steps: S1, acquiring confidence data of target lining points, including void confidence, leakage confidence, and crack confidence; when any confidence data is greater than a preset confidence threshold, the corresponding target lining point is recorded as a first test point; S2, calculating the difference between the void confidence and leakage confidence of the first test point; if the difference is between the preset first threshold and the preset second threshold, the first test point is marked and a first region is delineated by a preset first radius; S3, acquiring several discrete target lining points in the first region, acquiring their crack confidence and determining whether it is greater than a preset confidence threshold; if so, the corresponding target lining point is recorded as a second test point and a second region is delineated by a preset second radius; if the ratio of the sum of all second regions to the first region is less than a preset third threshold, then according to the... S4. If the ratio is greater than or equal to a preset third threshold, the confidence data of all target lining points in the first region is increased according to the ratio, and the corresponding second test point is recorded as the third test point; S5. If the Euclidean distance between any two adjacent third test points is greater than or equal to a preset fourth threshold, the confidence data of all target lining points in the second region is decreased according to the distance, and S1 to S5 are repeated; S6. If there is at least one pair of third test points whose Euclidean distance is less than the preset fourth threshold, for each pair of points, a unit vector is generated in the direction from the point with the smaller crack confidence value to the point with the larger crack confidence value. An adjustment coefficient is calculated based on all the unit vectors, and the confidence data of the third test point is adjusted accordingly; S7. A defect diagnosis report is output based on the confidence data of all target lining points.

[0007] Secondly, this application provides a geological mapping image processing system based on real-time dynamic positioning, comprising: a data acquisition module: used to acquire confidence data of target lining points, including void confidence, leakage confidence, and crack confidence; when any confidence data is greater than a preset confidence threshold, the corresponding target lining point is recorded as a first test point; a data calculation module: used to calculate the difference between the void confidence and leakage confidence of the first test point; if the difference is between the preset first threshold and the preset second threshold, the first test point is marked and a first region is delineated by a preset first radius; a data judgment first module: used to acquire several discrete target lining points in the first region, acquire their crack confidence and determine whether it is greater than a preset confidence threshold; if so, the corresponding target lining point is recorded as a second test point and a second region is delineated by a preset second radius; if the ratio of the sum of all second regions to the first region is less than a preset third threshold, the confidence number of all target lining points in the first region is reduced according to the ratio. The data acquisition module is executed again to the data judgment first module; the data judgment second module is used to increase the confidence data of all target lining points in the first region according to the ratio if the ratio is greater than or equal to a preset third threshold, and record the corresponding second test point as the third test point; the data judgment third module is used to decrease the confidence data of all target lining points in the second region according to the distance if the Euclidean distance between each pair of adjacent third test points is greater than or equal to a preset fourth threshold, and execute the data acquisition module again to the data judgment third module; the data judgment fourth module is used to generate a unit vector for each pair of points, with the direction from the point with the smaller crack confidence value to the point with the larger crack confidence value, calculate the adjustment coefficient according to all the unit vectors, and adjust the confidence data of the third test point accordingly; the data output module is used to output a defect diagnosis report according to the confidence data of all target lining points.

[0008] Thirdly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned geological mapping image processing method based on real-time dynamic positioning.

[0009] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages:

[0010] 1. The crack distribution density is verified through regional ratio analysis, the spatial clustering characteristics of defect points are confirmed through Euclidean distance detection, and finally, the confidence assessment is optimized through vector calibration. This progressive verification mechanism effectively eliminates the interference of isolated noise points and surface anomalies on the diagnostic results, ensuring that the final defect diagnosis report has higher reliability. It shows a significant advantage, particularly in distinguishing between apparent leakage and structural defects, establishing a multi-level spatial verification mechanism that significantly improves the accuracy of defect identification.

[0011] 2. The confidence level data is dynamically adjusted based on the spatial verification results. When the regional ratio or spatial distance does not meet the preset threshold, the confidence level of the corresponding region is automatically reduced and the detection and analysis are repeated. When the verification is passed, the confidence level is increased and the process proceeds to the next stage. This iterative optimization mechanism enables the diagnostic process to have self-correction capabilities, effectively overcoming the problem of misjudgment accumulation caused by a single judgment in the existing technology. It significantly improves the diagnostic stability under complex working conditions and realizes the dynamic optimization and self-correction of defect features.

[0012] 3. Through area delineation, spatial distance detection, and vector analysis, a complete defect spatial correlation assessment system was established. It not only considers the characteristic intensity of individual detection points, but also focuses on the inherent connection and regularity of defect points in spatial distribution. This spatial correlation-based analysis method is more in line with the objective laws of defect development in actual engineering, can accurately identify risk areas with development potential, provide a reliable basis for preventive maintenance, and improve the engineering practicality of defect diagnosis. Attached Figure Description

[0013] Figure 1 A schematic diagram of a geological mapping image processing method based on real-time dynamic positioning provided in an embodiment of this application;

[0014] Figure 2 A schematic diagram of the logic flow of a geological mapping image processing method based on real-time dynamic positioning provided in an embodiment of this application;

[0015] Figure 3 A schematic diagram of the logic flow for morphological symbiosis analysis provided in an embodiment of this application;

[0016] Figure 4 Figure 3 illustrates the logical flow of leakage path tracing as provided in an embodiment of this application;

[0017] Figure 5 This is a schematic diagram of the structure of a geological mapping image processing system based on real-time dynamic positioning, provided in an embodiment of this application. Detailed Implementation

[0018] This application embodiment solves the technical problem of insufficient accuracy in the identification and risk assessment of voids, leakage and cracks in the health monitoring of linings of large transportation tunnel projects by using a geological mapping image processing method based on real-time dynamic positioning.

[0019] In existing technologies, tunnel lining defect diagnosis mainly relies on threshold judgments based on single sensor data, lacking comprehensive analysis of the spatial correlation of defects. Existing technologies analyze data from individual monitoring points in isolation, easily misjudging local noise as structural defects and failing to reveal the intrinsic connections between cracks, leaks, and cavities. In one tunnel project, surface depressions were misjudged as cavities behind the tunnel, leading to unnecessary excavation and repair; in another case, a persistent leakage area was broken down into multiple isolated points, failing to identify its recharge path and delaying critical remediation efforts.

[0020] To address these issues, the inventors discovered that defects exhibit dynamic correlations in their spatial distribution, such as the correlation between leakage paths and crack orientations, and the clustering of cracks at the edges of cavities. Traditional single-point analysis cannot capture these characteristics, necessitating a multi-round spatial verification mechanism. By studying the symbiotic relationship of defects, the inventors proposed triggering regional expansion analysis using confidence level differences, and dynamically revising the initial judgment using the data distribution characteristics of adjacent regions. This approach overcomes the limitations of traditional static threshold methods, introducing spatial topological relationships into the defect diagnosis logic.

[0021] Therefore, this application proposes a method including the following steps: acquiring multiple confidence level data of target lining points, and marking any confidence level exceeding a threshold as a test point; calculating the difference between the confidence levels of voids and leaks, and delineating the analysis area based on the range of the difference; verifying the confidence level of cracks within the area, and dynamically adjusting the confidence level through the area ratio; performing Euclidean distance analysis on the areas that meet the conditions, and generating vector adjustment coefficients based on the relationship between neighboring points; and finally outputting a comprehensive diagnostic report.

[0022] Existing technologies determine defects based on a single point of data only, while this solution triggers multiple rounds of spatial verification through difference values, establishing a dynamic self-correcting mechanism. Existing technologies process different defect types independently; this solution introduces confidence difference comparison for the first time, enabling the capture of spatial correlation characteristics between different defect types. Existing technologies use fixed thresholds; this solution achieves adaptive parameter adjustment through region ratios and distance thresholds, significantly improving diagnostic robustness under complex operating conditions.

[0023] Through the above technical solutions, this application effectively distinguishes between real defects and sensor noise, avoiding misjudging local surface damage as structural defects. Spatial correlation analysis accurately identifies the leakage path and the supply relationship with the underlying cavity, providing a reliable basis for developing a comprehensive remediation plan. A dynamic calibration mechanism eliminates the influence of isolated anomalies, reducing the false alarm rate while maintaining high sensitivity.

[0024] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0025] like Figure 1 The diagram shows a structural schematic of a geological mapping image processing method based on real-time dynamic positioning provided in an embodiment of this application. The method includes the following steps: S1. Obtaining confidence data for target lining points, including void confidence, leakage confidence, and crack confidence. When any confidence data exceeds a preset confidence threshold, the corresponding target lining point is recorded as the first test point. S2. Calculating the difference between the void confidence and leakage confidence of the first test point. If the difference is between the preset first threshold and the preset second threshold, the first test point is marked, and a first region is defined by a preset first radius. S3. Obtaining several discrete target lining points within the first region, acquiring their crack confidence, and determining whether it exceeds a preset confidence threshold. If so, the corresponding target lining point is recorded as the second test point, and a second region is defined by a preset second radius. If the ratio of the sum of all second regions to the first region is less than a preset third threshold, then... S4. If the ratio is greater than or equal to a preset third threshold, the confidence data of all target lining points in the first region is increased according to the ratio, and the corresponding second test point is recorded as the third test point; S5. If the Euclidean distance between any two adjacent third test points is greater than or equal to a preset fourth threshold, the confidence data of all target lining points in the second region is decreased according to the distance, and S1 to S5 are repeated; S6. If there is at least one pair of third test points whose Euclidean distance is less than the preset fourth threshold, for each pair of points, a unit vector is generated in the direction from the point with the smaller crack confidence value to the point with the larger crack confidence value. An adjustment coefficient is calculated based on all the unit vectors, and the confidence data of the third test point is adjusted accordingly; S7. A defect diagnosis report is output based on the confidence data of all target lining points.

[0026] Figure 2This is a schematic diagram of the logical flow of a geological mapping image processing method based on real-time dynamic positioning provided in an embodiment of this application. The confidence data refers to the quantitative index of defect probability obtained by multi-source sensor fusion calculation. Specifically, it can be achieved by combining color texture analysis, temperature distribution model and three-dimensional point cloud curvature calculation, and is used to characterize the probability of occurrence of different defect types.

[0027] The difference value refers to the numerical difference in confidence levels of different defects at the same location. It can be achieved through subtraction or ratio calculation and is used to trigger the spatial verification mechanism.

[0028] The region ratio refers to the ratio of the area of ​​a sub-region to the area of ​​the parent region. It can be achieved using spatial overlay analysis and is used to determine whether a local anomaly is a characteristic of a sexual defect.

[0029] The adjustment factor is a correction factor generated based on the confidence gradient of neighboring points. It can be implemented through vector composition operations and is used to eliminate the influence of isolated noise points.

[0030] When the confidence level at a certain point exceeds a threshold, a defect is not immediately identified. Instead, the confidence level distribution relationship between that point and the surrounding area is calculated. The difference value is used to determine whether the analysis range needs to be expanded, and the distribution density of other types of defects is verified within the expanded area.

[0031] If the region ratio is below the threshold, it indicates that the initial anomaly may be caused by local disturbance. The confidence level of that region will be automatically reduced and the analysis will be repeated.

[0032] If the regional ratio consistently meets the standard after multiple verifications, spatial topology analysis is initiated to detect the distance relationship between neighboring anomalies.

[0033] For densely distributed anomaly clusters, the confidence level is calculated and the direction is adjusted through vector synthesis to enhance the confidence level of the real defects while suppressing random noise.

[0034] Furthermore, the specific process for obtaining the confidence data of the target lining point is as follows: Acquire tunnel lining surface image data collected by a multispectral camera, thermal infrared image data collected by a thermal infrared camera, and three-dimensional point cloud data collected by a laser scanner; based on the tunnel lining surface image data, generate confidence data C_wet for the wetted area using a preset color texture analysis model; based on the thermal infrared image data, generate confidence data C_cold for the temperature anomaly using a preset temperature distribution analysis model; based on the three-dimensional point cloud data, generate confidence data C_dent for the concave area using a preset curvature calculation analysis model; perform fusion calculation on the confidence data C_wet for the wetted area, C_cold for the temperature anomaly, and C_dent for the concave area at the same spatial location to generate confidence scores P_hollow for voids, P_leak for leaks, and P_crack for cracks; the fusion calculation formula for the void confidence score is as follows: The formula for calculating the leakage confidence level is as follows: Where k is a preset weighting coefficient extracted from the geological survey image database; the fusion calculation formula for crack confidence is as follows: , where sqrt represents the square root operation.

[0035] In this embodiment, a multispectral camera refers to a device capable of capturing images in the visible and near-infrared bands. Specifically, it can be implemented using an imaging system equipped with multiple narrowband filters to extract color difference features on the lining surface caused by leakage.

[0036] A thermal infrared camera is an imaging device that detects the heat radiated by an object itself. Specifically, it can be implemented using an uncooled focal plane array detector to capture areas of abnormal temperature caused by cavities or leaks.

[0037] A laser scanner is a device that generates three-dimensional coordinates by emitting laser pulses and receiving reflected signals. Specifically, it can be implemented using scanning equipment based on phase or pulse ranging principles, and is used to construct the geometric deformation characteristics of the lining surface.

[0038] The preset color texture analysis model refers to an algorithm model based on color space conversion and texture feature extraction. Specifically, it can be implemented by combining the HSV color space with the gray-level co-occurrence matrix to quantify the probability distribution of wet stain areas on the surface.

[0039] The preset temperature distribution analysis model refers to a mathematical model based on the heat conduction equation and temperature gradient calculation. Specifically, it can be implemented using the finite difference method combined with Gaussian filtering to identify the confidence level of abnormally low temperature regions.

[0040] The preset curvature calculation and analysis model refers to a calculation method based on the differential geometric properties of point clouds. Specifically, it can be implemented by using the moving least squares method to fit the local surface and calculate the principal curvature, which is used to detect the geometric features of concave regions.

[0041] Multi-band reflectance data of the tunnel lining surface were acquired using a multispectral camera. The RGB image was converted to the HSV color space using a color texture analysis model, and the local variance of the saturation channel was extracted as confidence data for the wet stain area.

[0042] After the radiation temperature data collected by the thermal infrared camera is processed by the temperature distribution analysis model, the average temperature deviation between each pixel and the surrounding area is calculated, and temperature anomaly confidence data is generated by combining the preset temperature gradient threshold.

[0043] The 3D point cloud data generated by the laser scanner is processed by the curvature calculation and analysis model. The local surface is fitted with the neighborhood point cloud and the Gaussian curvature is calculated. The negative curvature region is mapped to the confidence data of the concave region.

[0044] The three confidence levels are fused and calculated after spatial registration. The void confidence level is generated by suppressing the influence of moisture and strengthening the correlation between geometric deformation and temperature anomaly. The leakage confidence level is generated by temperature anomaly as the main factor and superimposed with geometric deformation as an auxiliary correction. The crack confidence level is generated by balancing the combined effects of humidity and temperature and strengthening geometric deformation characteristics.

[0045] By establishing a confidence fusion formula with clear physical meaning, the discrimination boundary between defect types is strengthened while preserving the characteristics of each sensor data.

[0046] For example, the inverse correction of the wet stain confidence score is introduced into the void confidence score calculation, which can effectively distinguish between dry void areas and wet erosion areas; the crack confidence score uses the geometric mean to balance the influence of humidity and temperature, avoiding misidentification caused by a single factor.

[0047] This multi-source data collaborative verification mechanism provides a reliable data foundation for subsequent defect classification and spatial correlation analysis.

[0048] Furthermore, step S4 also includes a morphological symbiosis analysis step: based on the first region to which the third test point belongs obtained in step S4, linear crack features are extracted through cluster analysis, and key points of the linear crack features are identified, including endpoints, intersections, and turning points; based on the void confidence obtained in step S1, if the void confidence is greater than a preset void threshold, the corresponding target lining point is recorded as the fourth test point and a preset third radius is used to delineate the third region, and the third region is analyzed through a boundary tracking algorithm to generate the void region boundary; the shortest Euclidean distance from each key point to the adjacent void region boundary is calculated, and when at least one key point has a shortest Euclidean distance from the void boundary that is less than a preset proximity threshold, it is determined that a crack-void symbiosis exists in the region where the key point is located; the symbiosis reinforcement coefficient is calculated based on the symbiosis reinforcement formula. Where R_s represents the symbiotic reinforcement coefficient, d represents the minimum distance from the key point to the cavity boundary, D represents the preset proximity threshold, and δ is the preset reinforcement amplitude coefficient; the crack confidence and cavity confidence of the third test point in the region where the key point is located are multiplied by the symbiotic reinforcement coefficient, the crack confidence and cavity confidence are updated according to the multiplication result, and the updated crack confidence and cavity confidence are used in subsequent steps S6 and S7.

[0049] In this embodiment, Figure 3 This is a schematic diagram of the logical flow of morphological symbiosis analysis provided in an embodiment of this application; cluster analysis refers to the method of grouping crack features in a first region to which crack confidence data points with similar spatial distribution characteristics belong. Specifically, it can be implemented using the DBSCAN algorithm or the K-means algorithm to extract continuous linear crack features from the first region.

[0050] Key points refer to important nodes in the geometric shape of linear cracks. Specifically, they can be implemented using corner detection algorithms or curvature extremum detection algorithms. They are used to characterize the extension direction and morphological changes of cracks. It should be noted that in actual data processing, key points must include third test points, but not all third test points are identified as key points. Key points include endpoints, intersections, and turning points. The third test points contained in the key points are identified and classified as endpoints, intersections, and turning points.

[0051] Boundary tracing algorithms refer to methods for extracting the contours of regions with a hole confidence level higher than a threshold. Specifically, they can be implemented using Moore's neighborhood tracing method or chain code encoding method to generate accurate hole region boundaries.

[0052] The symbiotic strengthening coefficient is a parameter that dynamically adjusts the confidence value based on the spatial proximity of cracks and cavities. Specifically, it can be calculated by multiplying the normalized distance difference with the strengthening amplitude coefficient, and is used to reflect the mechanical correlation strength between cracks and cavities.

[0053] By analyzing the distance between key crack points and void boundaries, the symbiotic relationship between cracks and voids can be accurately identified. A dynamic confidence enhancement mechanism can improve the characterization intensity of composite defects, avoiding diagnostic omissions caused by isolated analysis. This approach is particularly suitable for diagnosing tunnel lining defects with complex mechanical relationships, providing a more reliable basis for engineering maintenance decisions.

[0054] Furthermore, step S6 also includes a leakage path tracing step: if there is at least one pair of third test points whose Euclidean distance is less than a preset fourth threshold, then the spatial distribution pattern of the leakage confidence of the third test points is analyzed, and the spatial distribution pattern analysis result is a planar distribution or a linear distribution; based on the distribution pattern analysis result, the corresponding path tracing mode is selected to generate leakage path vector data; the leakage path vector data is used to generate a defect diagnosis report.

[0055] In this embodiment, Figure 4 Figure 3 illustrates the logical flow of a leakage path tracing method provided in this application embodiment; the spatial distribution pattern of leakage confidence refers to the spatial aggregation characteristics of leakage defects, which can be achieved by using connected domain shape parameter analysis. By calculating the area, perimeter, and aspect ratio of the high confidence region, the leakage patterns of planar diffusion and linear extension can be distinguished.

[0056] The path tracing mode refers to selecting different search strategies based on the leakage pattern. Specifically, it can be implemented using the gradient ascent method or the directional constraint method. For area distribution, regional peak tracing is used, and for linear distribution, directional guided tracing is used, thereby adapting to the source tracing needs of different leakage types.

[0057] Leakage path vector data refers to geometric data that describes the direction of leakage development. Specifically, it can be implemented using ordered point sets or directional line segment sets. By recording the migration path of the point with the highest leakage confidence, a quantifiable leakage development model can be established.

[0058] Current technologies rely solely on single-point confidence levels for leak detection, failing to distinguish between surface wetting and structural leaks, and even less able to identify the direction of leak spread. By introducing spatial distribution morphology analysis, it is possible to effectively differentiate between condensate accumulation and actual leaks. Path tracing models can accurately reconstruct the leak development process, avoiding the misjudgment of temporary surface wetting as structural defects.

[0059] Through the above technical solution, this application can accurately identify the propagation path and diffusion pattern of leakage defects, distinguish between surface wetting and structural leakage, and accurately locate the source of leakage. For planar leakage, multiple potential seepage points can be identified, and for linear leakage, the water flow migration path can be traced, providing accurate location basis for leak sealing and repair, and reducing the risk of ineffective excavation due to misjudgment.

[0060] Furthermore, the spatial distribution pattern of the leakage confidence at the third test point is analyzed, specifically including: binarizing the leakage confidence at the third test point according to a preset leakage pattern threshold to obtain a high-confidence region mask; extracting the connected components of the high-confidence region mask; calculating the shape feature parameters of each connected component, including area, perimeter, and aspect ratio of the minimum bounding rectangle; and classifying the distribution pattern into planar or linear distribution based on the shape feature parameters using preset classification rules. The preset classification rules are as follows: when the area of ​​the connected component is greater than the area threshold and the aspect ratio is less than the aspect ratio threshold, it is determined to be a planar distribution; when the area of ​​the connected component is greater than the area threshold and the aspect ratio is greater than or equal to the aspect ratio threshold, it is determined to be a linear distribution.

[0061] In this embodiment, the preset leakage morphology threshold is a critical value used to distinguish between high-confidence and low-confidence regions. Specifically, it can be implemented using empirical values ​​derived from historical data, such as 0.7. Its function is to filter noise interference and focus on effective leakage areas. A connected region refers to an independent region composed of spatially continuous high-confidence pixels. Specifically, it can be implemented using region growing algorithms or morphological operations to identify leakage regions with complete morphological features.

[0062] In the shape feature parameters, the area is used to quantify the region size, the perimeter is used to describe the boundary complexity, and the aspect ratio of the minimum bounding rectangle is used to characterize the region's extension direction.

[0063] Specifically, after binarization, the leakage confidence data is used to form a high-confidence region mask, and independent region units are extracted through connected component analysis.

[0064] The area of ​​each connected region is obtained by pixel counting; for example, an area exceeding a set threshold is considered a valid leakage area. The minimum bounding rectangle is calculated using a rotating caliper algorithm, and its aspect ratio is used to determine the region's shape: when the aspect ratio is below the set threshold, it indicates that the region is approximately circular or elliptical and is determined to be a planar distribution; when the aspect ratio is above the threshold, it indicates that the region extends linearly and is determined to be a linear distribution.

[0065] The classification results provide a basis for subsequent path tracing mode selection. For example, a planar distribution corresponds to a point diffusion mode, and a linear distribution corresponds to a direction tracing mode.

[0066] By accurately distinguishing between planar and linear leaks through morphological feature analysis, a reliable basis is provided for selecting a matching path tracing algorithm, thereby improving the accuracy of leak source identification and the credibility of defect diagnosis reports, and effectively solving the problem of source location deviation caused by misjudgment of leak path morphology.

[0067] Furthermore, the corresponding path tracing mode is selected to generate leakage path vector data, specifically including: when the distribution pattern analysis result is a planar distribution, the third test point with the highest leakage confidence is obtained and recorded as the undetermined water outlet point, and the first operation of water outlet point search is performed; the first operation of water outlet point search is as follows: taking the undetermined water outlet point and taking a fourth region defined by a preset fourth radius, several discrete target lining points are obtained, and the point with the highest leakage confidence of the target lining points is taken as the new undetermined water outlet point; the first operation of water outlet point search is repeated and the position of the undetermined water outlet point in each operation is recorded until the leakage confidence of all target lining points in the fourth region is lower than the preset path confidence threshold, all recorded undetermined water outlet points are obtained and merged to obtain the first set of undetermined water outlet points; the first set of undetermined water outlet points is arranged in descending order, and the undetermined water outlet points in the preset ranking are determined as discrete leakage sources, and a defect diagnosis report is output accordingly.

[0068] like Figure 2 The diagram shown is a logical flow diagram of a geological mapping image processing system based on real-time dynamic positioning provided in an embodiment of this application.

[0069] In this embodiment, linear distribution refers to the leakage confidence level exhibiting a narrow and elongated shape in space. Specifically, this can be achieved by ensuring that the aspect ratio of the connected domain is greater than or equal to a threshold. This feature is used to identify linear leakage paths associated with structural joints or cracks.

[0070] The preset fourth radius refers to a dynamic search range set based on engineering experience. Specifically, it can be determined using a proportional coefficient related to the lining thickness. This parameter controls the tracking accuracy of the leakage path. The preset path confidence threshold refers to the critical condition for determining the termination of the leakage path. Specifically, it can be set using historical data statistics. This threshold is used to distinguish between valid leakage paths and random noise.

[0071] Preset ranking refers to the limit on the number of major leakage sources to be screened. This can be determined by a fixed number or a percentage. This parameter ensures that the diagnostic report focuses on key risk points.

[0072] When the spatial distribution of leakage confidence is determined to be linear, starting from the current highest confidence point, the system continuously searches for neighboring points with higher confidence within a preset radius, forming a progressive path tracing.

[0073] In each iteration, if a higher confidence point is found, the current position is updated and the trajectory is recorded until the overall confidence of the region falls below the threshold or further optimization is not possible.

[0074] After sorting the final set of recorded trajectory points by confidence level, the preceding points are selected as the leakage source, effectively distinguishing between the main path and branch interference. For example, in the scenario of leakage in the longitudinal structural joint of a tunnel, this method can trace back step by step along the joint length to the highest risk point, avoiding misjudging intermediate seepage points as the source.

[0075] In situations with multiple parallel leakage paths, it can effectively distinguish between the main channels and secondary branches, providing a precise location basis for leak sealing construction, avoiding resource waste caused by path misjudgment, accurately identifying the evolution direction and core source of linear leakage paths, and overcoming the tracking failure problem of existing technologies in complex situations.

[0076] Furthermore, the process of generating leakage path vector data by selecting the corresponding path tracing mode also includes: when the distribution pattern analysis result is a linear distribution, obtaining the third test point with the highest leakage confidence and recording it as the undetermined outlet point, and performing the second outlet point search operation; the second outlet point search operation is as follows: obtaining several discrete target lining points within a fourth region defined by a preset fourth radius based on the undetermined outlet point, and using the target lining point with a higher leakage confidence than the undetermined outlet point as the new undetermined outlet point; repeating the second outlet point search operation and recording the position of the undetermined outlet point in each operation until the leakage confidence of all target lining points in the fourth region is lower than the preset path confidence threshold or no target lining point with a higher leakage confidence can be found, obtaining all recorded undetermined outlet points and merging them to obtain a second set of undetermined outlet points; arranging the second set of undetermined outlet points in descending order, taking the undetermined outlet points within the preset ranking as discrete leakage sources, and outputting a defect diagnosis report accordingly.

[0077] In this embodiment, linear distribution refers to the spatial distribution of high-confidence leakage areas as long strips or continuous lines. Specifically, it can be determined by calculating the aspect ratio parameter of the connected domain. When the aspect ratio exceeds a set threshold, it is determined to be a linear distribution. This feature is used to identify leakage paths with extension characteristics.

[0078] The second operation of the outlet point search refers to the process of iteratively searching within the linear leakage area. Each iteration expands the search range with the current highest leakage confidence point as the center. This operation can dynamically track the extension direction of the leakage path.

[0079] The preset fourth radius refers to the spatial range parameter used to define each search operation. It can be set according to engineering experience or lining structure characteristics. This parameter controls the search step size to ensure the continuity of path tracing.

[0080] The preset path confidence threshold is a critical value used to terminate the search process. The search stops when the leakage confidence of all points in the area is lower than this value. This threshold setting can prevent invalid tracking in low-confidence areas.

[0081] By establishing an ordered set of undetermined water outlets, clear leakage path data is provided for engineering maintenance, significantly improving the spatial guidance value of defect diagnosis reports. It can accurately identify the extension path and core source of leakage water flow in linear leakage scenarios, solving the problem of leakage source location deviation caused by the lack of spatial correlation analysis in existing technologies.

[0082] Furthermore, using the leakage path vector data to generate a defect diagnosis report also includes: connecting all the undetermined water outlets recorded in the first set or the second set of undetermined water outlets in the original recording order to form a first sequence path or a second sequence path; establishing a search area within a preset width range on both sides of the first sequence path or the second sequence path as a reference to generate a first search area or a second search area; obtaining a number of discrete target lining points within the first search area or the second search area; if the void confidence of more than a preset number of target lining points is higher than a preset void threshold, then marking the corresponding area as a leakage void associated area and outputting a defect diagnosis report accordingly.

[0083] In this embodiment, the first set or the second set of undetermined water outlets refers to the set of spatial locations of potential leakage sources determined through different search modes. Specifically, it can be implemented by using an iterative search algorithm combined with confidence threshold judgment, and is used to record candidate water outlets found during the leakage path tracing process.

[0084] The first or second sequence path refers to the vector trajectory formed by spatially connecting the undetermined water outlets in the order of discovery. Specifically, it can be implemented using linear interpolation algorithms or minimum spanning tree algorithms, and is used to characterize the spatial direction of the leakage path.

[0085] The preset width range refers to the boundary parameter of the strip area set around the leakage path. It can be set according to the thickness of the lining structure or the seepage diffusion law to limit the search range of the associated area.

[0086] The search area refers to the spatial analysis range constructed based on the leakage path and its preset width. Specifically, it can be implemented using a buffer generation algorithm or rasterization processing to locate potentially related defect areas.

[0087] The preset number of points refers to the threshold for the number of abnormal points that trigger the marking of the region. Specifically, it can be dynamically adjusted according to the region area and sampling density to determine whether there are significant hole correlation features.

[0088] This technology addresses the lack of correlation analysis between leakage and voids in existing technologies, improves the diagnostic accuracy of complex defects, and provides reliable data support for developing collaborative remediation solutions. For example, in cases of long-term leakage in tunnel linings, this solution can accurately distinguish between simple surface seepage and persistent leakage caused by voids behind the lining, guiding engineers to take targeted grouting or drainage measures.

[0089] Figure 5 This application provides a schematic diagram of a geological mapping image processing system based on real-time dynamic positioning, comprising: a data acquisition module for acquiring confidence data of target lining points, including void confidence, leakage confidence, and crack confidence; when any confidence data exceeds a preset confidence threshold, the corresponding target lining point is recorded as a first test point; a data calculation module for calculating the difference between the void confidence and leakage confidence of the first test point; if the difference is between the preset first threshold and the preset second threshold, the first test point is marked and a first region is defined by a preset first radius; and a data judgment module for acquiring several discrete target lining points in the first region, acquiring their crack confidence, and determining whether it exceeds a preset confidence threshold; if so, the corresponding target lining point is recorded as a second test point and a second region is defined by a preset second radius; if the ratio of the sum of all second regions to the first region is less than a preset third threshold, the confidence level of all target lining points in the first region is reduced according to the ratio. The data acquisition module is re-executed to the data judgment first module; the data judgment second module is used to increase the confidence data of all target lining points in the first region according to the ratio if the ratio is greater than or equal to a preset third threshold, and record the corresponding second test point as the third test point; the data judgment third module is used to decrease the confidence data of all target lining points in the second region according to the distance if the Euclidean distance between each pair of adjacent third test points is greater than or equal to a preset fourth threshold, and re-execute the data acquisition module to the data judgment third module; the data judgment fourth module is used to generate a unit vector for each pair of points if there is at least one pair of third test points whose Euclidean distance is less than the preset fourth threshold, and calculate the adjustment coefficient according to all the unit vectors and adjust the confidence data of the third test point accordingly; the data output module is used to output a defect diagnosis report based on the confidence data of all target lining points.

[0090] This application also provides a computer-readable storage medium for storing a computer program, which, when executed by a processor, implements the steps of a geological mapping image processing method based on real-time dynamic positioning.

[0091] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0092] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0093] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0094] These computer program instructions can also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0095] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0096] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A geological mapping image processing method based on real-time dynamic positioning, characterized in that, Includes the following steps: S1. Obtain the confidence data of the target lining point, including the confidence of voids, leakage and cracks. When any confidence data is greater than the preset confidence threshold, the corresponding target lining point is recorded as the first test point. S2. Calculate the difference between the void confidence and the leakage confidence of the first test point. If the difference is between the preset first threshold and the preset second threshold, mark the first test point and delineate the first region with a preset first radius. S3. Obtain several discrete target lining points in the first region, obtain their crack confidence and determine whether it is greater than a preset confidence threshold. If so, record the corresponding target lining point as the second test point and take a preset second radius to delineate the second region. If the ratio of the sum of all second regions to the first region is less than a preset third threshold, reduce the confidence data of all target lining points in the first region according to the ratio, and re-execute S1 to S3. S4. If the ratio is greater than or equal to a preset third threshold, the confidence data of all target lining points in the first region is increased according to the ratio, and the corresponding second test point is recorded as the third test point. S5. If the Euclidean distance between any two adjacent third test points is greater than or equal to the preset fourth threshold, then reduce the confidence data of all target lining points in the second region according to the distance, and re-execute S1 to S5. S6. If there is at least one pair of third test points whose Euclidean distance is less than a preset fourth threshold, then for each pair of points, a unit vector is generated in the direction from the point with the smaller crack confidence value to the point with the larger crack confidence value. The adjustment coefficient is calculated based on all the unit vectors and the confidence data of the third test point is adjusted accordingly. S7. Output a defect diagnosis report based on the confidence data of all target lining points.

2. The geological mapping image processing method based on real-time dynamic positioning according to claim 1, characterized in that, The specific process for obtaining the confidence data of the target lining points is as follows: Acquire surface image data of tunnel lining acquired by a multispectral camera, thermal infrared image data acquired by a thermal infrared camera, and three-dimensional point cloud data acquired by a laser scanner; Based on the image data of the tunnel lining surface, confidence data C_wet of the wet stain area is generated through a preset color texture analysis model; Based on the thermal infrared image data, temperature anomaly confidence data C_cold is generated through a preset temperature distribution analysis model. Based on the three-dimensional point cloud data, confidence data C_dent of the concave region is generated through a preset curvature calculation and analysis model. The confidence data C_wet of the wet stain area, the confidence data C_cold of the temperature anomaly, and the confidence data C_dent of the depression area at the same spatial location are fused and calculated to generate the confidence scores P_hollow for voids, P_leak for leaks, and P_crack for cracks. The fusion calculation formula for hole confidence is as follows: ; The formula for calculating the leakage confidence level is as follows: Where k is a preset weighting coefficient extracted from a geological mapping image database; The formula for calculating the crack confidence level is as follows: , where sqrt represents the square root operation.

3. The geological mapping image processing method based on real-time dynamic positioning according to claim 2, characterized in that, Step S4 further includes a morphological symbiosis analysis step: Based on the first region to which the third test point belongs obtained in step S4, linear crack features are extracted through cluster analysis, and key points of the linear crack features are identified, including endpoints, intersections, and turning points. Based on the void confidence obtained in step S1, if the void confidence is greater than the preset void threshold, the corresponding target lining point is recorded as the fourth test point and the third region is defined by the preset third radius. The third region is analyzed by the boundary tracking algorithm and the void region boundary is generated. Calculate the shortest Euclidean distance from each key point to the boundary of the adjacent cavity region. When the shortest Euclidean distance between at least one key point and the cavity boundary is less than a preset proximity threshold, it is determined that a crack-cavity symbiosis exists in the region where the key point is located. Calculation of symbiote enhancement coefficient based on symbiote enhancement formula: Where R_s represents the symbiotic enhancement coefficient, d represents the minimum distance from the key point to the cavity boundary, D represents the preset proximity threshold, and δ is the preset enhancement amplitude coefficient. The crack confidence and void confidence of the third test point in the region where the key point is located are multiplied by the symbiotic strengthening coefficient. The crack confidence and void confidence are updated according to the multiplication result. The updated crack confidence and void confidence are then used in subsequent steps S6 and S7.

4. The geological mapping image processing method based on real-time dynamic positioning according to claim 1, characterized in that, Step S6 further includes a leakage path tracing step: If there is at least one pair of third test points whose Euclidean distance is less than a preset fourth threshold, then the spatial distribution pattern of the leakage confidence of the third test points is analyzed. The spatial distribution pattern analysis result is a planar distribution or a linear distribution. Based on the distribution pattern analysis results, select the corresponding path tracing mode to generate leakage path vector data; The leakage path vector data is used to generate a defect diagnosis report.

5. A geological mapping image processing method based on real-time dynamic positioning according to claim 4, characterized in that, The analysis of the spatial distribution pattern of the leakage confidence level at the third test point specifically includes: The leakage confidence of the third test point is binarized according to the preset leakage morphology threshold to obtain a high-confidence area mask. Extract the connected components of the high-confidence region mask; Calculate the shape feature parameters of each connected component, including area, perimeter, and aspect ratio of the minimum bounding rectangle; Based on the shape feature parameters, the distribution pattern is classified into planar distribution or linear distribution according to the preset classification rules; The preset classification rules are as follows: When the area of ​​a connected component is greater than the area threshold and the aspect ratio is less than the aspect ratio threshold, it is determined to be a planar distribution. When the area of ​​a connected region is greater than the area threshold and the aspect ratio is greater than or equal to the aspect ratio threshold, it is determined to be a linear distribution.

6. The geological mapping image processing method based on real-time dynamic positioning according to claim 4, characterized in that, The step of selecting the corresponding path tracing mode to generate leakage path vector data specifically includes: When the distribution pattern analysis result is a planar distribution, the third test point with the highest leakage confidence is obtained and recorded as the undetermined outlet point, and the first operation of outlet point search is performed. The first operation for searching the water outlet point is as follows: A number of discrete target lining points are obtained within a fourth region defined by a preset fourth radius from the undetermined water outlet point, and the point with the highest leakage confidence of the target lining point is taken as the new undetermined water outlet point. Repeat the first operation of searching for water outlets and record the position of the undetermined water outlets in each operation until the leakage confidence of all target lining points in the fourth region is lower than the preset path confidence threshold. Obtain all recorded undetermined water outlets and merge them to obtain the first set of undetermined water outlets. Arrange the first set of undetermined water outlets in descending order, select the undetermined water outlets within the preset ranking as discrete leakage sources, and output a defect diagnosis report accordingly.

7. A geological mapping image processing method based on real-time dynamic positioning according to claim 6, characterized in that, The step of selecting the corresponding path tracing mode to generate leakage path vector data also includes: When the distribution pattern analysis result is a linear distribution, the third test point with the highest leakage confidence is obtained and recorded as the undetermined water outlet point, and the second operation of water outlet point search is performed. The second operation for searching the water outlet point is as follows: A number of discrete target lining points are obtained within a fourth region defined by a preset fourth radius from the undetermined water outlet point. The new undetermined water outlet point is the one with a higher leakage confidence level than the undetermined water outlet point. Repeat the second operation of searching for water outlets and record the position of the undetermined water outlets in each operation until the leakage confidence of all target lining points in the fourth region is lower than the preset path confidence threshold or no target lining point with higher leakage confidence can be found. Obtain all recorded undetermined water outlets and merge them to obtain the second set of undetermined water outlets. Arrange the second set of undetermined water outlets in descending order, select the undetermined water outlets within the preset ranking as discrete leakage sources, and output a defect diagnosis report accordingly.

8. A geological mapping image processing method based on real-time dynamic positioning according to claim 6 or 7, characterized in that, The step of using the leakage path vector data to generate a defect diagnosis report further includes: Connect all the undetermined water outlets recorded in the first set or the second set of undetermined water outlets in the original recording order to form the first sequence path or the second sequence path. Based on the first sequence path or the second sequence path, a search region is established within a preset width range on both sides to generate a first search region or a second search region. Several discrete target lining points are obtained in the first search area or the second search area. If the void confidence of more than a preset number of target lining points is higher than a preset void threshold, the corresponding area is marked as a leakage void associated area and a defect diagnosis report is output accordingly.

9. A geological mapping image processing system based on real-time dynamic positioning, characterized in that, include: Data acquisition module: used to acquire confidence data of target lining points, including confidence of voids, leakage and cracks. When any confidence data is greater than the preset confidence threshold, the corresponding target lining point is recorded as the first test point. Data calculation module: used to calculate the difference between the void confidence and the leakage confidence of the first test point. If the difference is between the preset first threshold and the preset second threshold, the first test point is marked and a first region is delineated by a preset first radius. The first data judgment module is used to acquire several discrete target lining points in the first region, acquire their crack confidence and determine whether it is greater than a preset confidence threshold. If it is, the corresponding target lining point is recorded as the second test point and the second region is defined by a preset second radius. If the ratio of the sum of all second regions to the first region is less than a preset third threshold, the confidence data of all target lining points in the first region is reduced according to the ratio, and the data acquisition module is re-executed to the first data judgment module. The second data judgment module is used to increase the confidence data of all target lining points in the first region according to the ratio if the ratio is greater than or equal to a preset third threshold, and record the corresponding second test point as the third test point. The third data judgment module is used to reduce the confidence data of all target lining points in the second region according to the distance if the Euclidean distance between each pair of adjacent third test points is greater than or equal to the preset fourth threshold, and then re-execute the data acquisition module to the third data judgment module. The fourth data judgment module is used to generate a unit vector for each pair of points if there is at least one pair of third test points whose Euclidean distance is less than a preset fourth threshold. The vector is then used to calculate the adjustment coefficient based on all the unit vectors and adjust the confidence data of the third test points accordingly. Data output module: Used to output a defect diagnosis report based on the confidence level data of all target lining points.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1-8.

Images