A multi-factor fusion high-precision three-dimensional detection method for complex strata

Abstract

The present application relates to the field of stratum exploration, and specifically discloses a complex stratum multi-factor fusion high-precision three-dimensional detection method, comprising: S1, constructing an adaptive multi-source detection system; S2, generating a two-level detection scheme based on stratum parameters of a detection area; S3, implementing differentiated detection data collection according to the two-level detection scheme; S4, performing space-time registration on multi-source data; S5, performing preprocessing and feature extraction on the registered data; S6, constructing a multi-factor weight fusion model; S7, performing error compensation and optimization on the fusion model; S8, performing parameter adaptation on a dual-mode borehole sonar module; S9, performing parameter setting and protection on an intelligent elastic wave CT module; and S10, verifying the results. The present application solves the problems of insufficient multi-source data fusion, failure to consider the multi-factor coupling influence of strata, difficulty in adapting the detection scheme to stratum complexity, reliance on manual data processing, and incomplete detection coverage of the traditional method, and effectively improves the identification accuracy of complex strata.

Description

Technical Field

[0001] This invention relates to the field of stratigraphic exploration technology, specifically to a high-precision three-dimensional exploration method for complex stratigraphic formations that integrates multiple factors. Background Technology

[0002] In the fields of complex strata exploration and geological engineering survey, traditional three-dimensional exploration methods have long faced technical limitations, making it difficult to meet the demand for high-precision and comprehensive geological information acquisition. In particular, the technical shortcomings are particularly prominent in core application scenarios such as structural identification and spatial distribution reconstruction of complex strata.

[0003] Traditional methods for probing complex geological formations generally rely on single geophysical techniques or simple data overlay, lacking a deep fusion mechanism for multi-source data. For example, tube wave testing can only cover a limited area around the borehole, making it difficult to reflect large-scale geological structures; while elastic wave CT can acquire cross-hole profile information, the output results are mostly in two-dimensional form, failing to intuitively present the three-dimensional spatial relationships of the formations; borehole sonar and borehole radar are limited by the differences in the physical properties of the detection medium, and are prone to signal distortion in high-conductivity or high-attenuation formations, leading to biases in the judgment of undesirable geological bodies such as caves and fissures. These single-technique or simply combined detection modes not only lose key geological details and introduce interpretation errors, but also fail to fully consider the coupled effects of multiple factors such as lithological solubility, tectonic fissure distribution, and groundwater activity in complex formations, directly resulting in incomplete and inaccurate descriptions of the true geological structure, making the detection results lack sufficient reliability in engineering design and construction guidance.

[0004] Meanwhile, existing detection methods lack adaptability and intelligence. On the one hand, traditional methods often employ fixed detection parameters and procedures, failing to dynamically adjust technical solutions based on stratigraphic complexity (such as karst development intensity and lithological variations). For example, using conventional survey line spacing and excitation energy in areas with strong karst development can easily lead to insufficient data penetration or signal interference. On the other hand, data processing heavily relies on human experience. From raw data denoising and feature extraction to model construction, there is a lack of automated collaborative analysis mechanisms, resulting in low efficiency and potential inconsistencies in results due to human judgment. Furthermore, existing methods do not comprehensively consider issues such as signal attenuation caused by detection depth and measurement blind spots caused by complex terrain, making it difficult to achieve systematic detection of complex strata from shallow to deep and from local to overall, thus hindering the in-depth development of geological engineering exploration.

[0005] In summary, existing technologies have significant shortcomings in three-dimensional exploration of complex strata. There is an urgent need for an innovative exploration method that can integrate the advantages of multi-source exploration, achieve dynamic adaptation and intelligent fusion, and comprehensively cover the influence of multiple factors in complex strata. This method would break through the bottlenecks of traditional technologies and provide high-precision geological information support for fields such as geological disaster early warning and engineering construction safety assurance. Summary of the Invention

[0006] To address the aforementioned problems in existing technologies, this invention provides a high-precision three-dimensional detection method for complex strata by fusing multiple factors. This method effectively solves the problems of insufficient multi-source data fusion, difficulty in adapting to strata complexity, reliance on manual data processing, and incomplete detection coverage in traditional methods. It effectively improves the accuracy of complex strata identification and spatial reconstruction, and enhances the adaptability and intelligence of detection.

[0007] To achieve the above objectives, this invention proposes a high-precision three-dimensional detection method for complex strata by fusing multiple factors, comprising:

[0008] S1. Construct an adaptive multi-source detection system, which includes an improved pipe wave testing module, a dual-mode borehole sonar module, an intelligent elastic wave CT module, and a zoned water transient electromagnetic module.

[0009] S2. Based on the stratigraphic parameters of the detection area, a two-level detection scheme is generated. In the early stage, the survey line is laid out through the transient electromagnetic module of the zoned water area and the distribution range of soluble rock is covered. The resistivity anomaly area is extracted by interpreting the data. Combined with the geological borehole verification, a three-dimensional hierarchical model is constructed to divide the stratigraphy into micro-complex layer, medium-complex layer and high-complex layer. In the later stage, the corresponding detection module combination is matched.

[0010] S3. Implement differentiated data collection according to the two-level detection scheme. Each module is deployed according to the preset spatial layout and collects corresponding geological data.

[0011] S4. Perform spatiotemporal registration on multi-source data to ensure spatiotemporal consistency;

[0012] S5. Preprocess and extract features from the registered data to identify stratigraphic anomaly areas;

[0013] S6. Construct a multi-factor weighted fusion model, dynamically adjust the weights of each module based on the stratigraphic characteristics and correct the weight coefficients in combination with the groundwater distribution, and calculate the fused three-dimensional coordinates.

[0014] S7. Perform error compensation and optimization on the fusion model, and output a 3D point cloud model and a digital elevation model;

[0015] S8. Perform parameter adaptation on the dual-mode borehole sonar module;

[0016] S9. Configure and protect the intelligent elastic wave CT module;

[0017] S10. Select verification boreholes covering different complex geological sections, compare the actual geological information with the fusion model prediction information, and use borehole radar to assist in verification to confirm whether the detection accuracy meets the requirements.

[0018] Preferably, the improved tube wave testing module consists of a super magnetostrictive source and a directional receiving array; the dual-mode borehole sonar module is equipped with a multi-frequency sonar probe and a 360° rotating scanning mechanism; the intelligent elastic wave CT module integrates a high-frequency source and a high-sensitivity detector; and the zoned water transient electromagnetic module adopts a multi-channel receiving coil.

[0019] Preferably, in S2, the three-dimensional grading model is a three-dimensional grading model of resistivity-solubility-tectonic fracture density; the criteria for determining a slightly complex layer are that the tectonic fracture density is less than a set threshold and there is no obvious karst development; the criteria for determining a moderately complex layer are that the tectonic fracture density is between two set thresholds and there are local small caves; and the criteria for determining a highly complex layer are that the tectonic fracture density is greater than a set threshold and there are large caves or dissolution zones.

[0020] Preferably, in S3, the micro-complex layer uses a combination of an improved pipe wave testing module as the main sensor and a dual-mode borehole sonar module as a supplementary sensor; the medium-complex layer uses a combination of an improved pipe wave testing module and an intelligent elastic wave CT module for collaborative detection; and the high-complexity layer uses a combination of an intelligent elastic wave CT module as the main sensor and a dual-mode borehole sonar module for three-dimensional scanning. The spacing between adjacent survey lines and boreholes is greater in the micro-complex layer region than in the medium-complex layer region, and greater in the medium-complex layer region than in the high-complexity layer region. The modules are arranged according to a preset spatial layout, specifically with the transient electromagnetic module survey lines of the zoned water area parallel to the direction of the detection area, the improved pipe wave testing module and the dual-mode borehole sonar module arranged along the survey lines corresponding to the boreholes, and the intelligent elastic wave CT module arranged across the boreholes to form a grid-like layout. Each module collects relevant data on borehole-side reflected waves, cave cross-sections, cross-bore wave velocities, and resistivity distribution.

[0021] Preferably, in S4, the spatiotemporal registration of multi-source data specifically involves aligning the time dimension of data from each module based on timestamps, and unifying the coordinate system through a spatial coordinate transformation matrix, wherein the coordinate transformation matrix satisfies:

[0022] ;

[0023] Where (X,Y,Z) are coordinates in a unified coordinate system. For the first The original coordinates of the detection module, where M is a 4×4 spatial transformation matrix;

[0024] Preferably, in S5, the preprocessing and feature extraction of the registered data to identify stratigraphic anomaly areas specifically involves using a data fusion algorithm to repair missing data and eliminate data overlap errors, using a ranging algorithm to distinguish between target and interference echoes, and using a semantic segmentation model that incorporates a lithological feature attention mechanism to identify stratigraphic anomaly areas. The data fusion algorithm constructs a complete signal array by performing matrix operations on the echo time difference of data collected by each module, filling in missing data segments, and eliminating overlapping data parts.

[0025] Preferably, in step S6, the fused three-dimensional coordinates are calculated using a fusion formula, wherein the fusion formula satisfies:

[0026] ;

[0027] In the formula, For the fused 3D coordinates, C, S, and E are the inversion coordinates of the improved pipe wave testing module, the dual-mode borehole sonar module, and the intelligent elastic wave CT module, respectively. , , The weighting coefficients and ;

[0028] When the groundwater level is shallow, increase the weighting coefficient of the dual-mode borehole sonar module. When the groundwater depth is at a medium level, increase the weighting coefficient of the intelligent elastic wave CT module. When the groundwater depth is relatively deep, increase the weighting coefficient of the improved pipe wave testing module. .

[0029] Preferably, in step S7, the specific steps for error compensation and optimization of the fused model to output a 3D point cloud model and a digital elevation model include: constructing a 3D model through mesh partitioning; correcting height errors using a slope adaptive correction formula; dynamically segmenting and correcting the initial 3D model using a modeling platform; and outputting a 3D point cloud model and a digital elevation model. The mesh unit size of the mesh partitioning decreases as the detection accuracy requirement increases. The slope adaptive correction formula is:

[0030] ;

[0031] In the formula, As compensation value, Here, k is the original elevation, and k is the slope correction factor. This represents the actual slope angle of the strata.

[0032] Preferably, in S8, the parameter adaptation of the dual-mode borehole sonar module specifically involves selecting a sonar probe of the corresponding frequency according to the geological type of the detection area, adjusting the scanning angle and the probe coupling state to ensure good coupling between the probe and the borehole wall; wherein, the sonar probe frequency of the dual-mode borehole sonar module is selected with a high-frequency probe in dense strata and a low-frequency probe in loose or fractured strata.

[0033] Preferably, in S9, the specific steps for setting parameters and protecting the intelligent elastic wave CT module include setting core detection parameters according to the stratigraphic development section and burying protective pipes in the detection borehole; the cross-hole spacing of the intelligent elastic wave CT module is greater in the micro-complex layer region than in the medium-complex layer region, and greater in the medium-complex layer region than in the high-complex layer region, while the excitation energy is the opposite; the core detection parameters include sampling interval, filter passband, number of superpositions and excitation energy.

[0034] Therefore, this invention proposes a high-precision three-dimensional detection method for complex strata by fusing multiple factors, with the following beneficial effects:

[0035] (1) Integrating the advantages of multi-source detection modules, the two-level detection scheme with dynamic adaptation can be adapted to strata of different complexities, solving the limitations of traditional single-technology detection and improving the comprehensiveness and pertinence of stratum detection.

[0036] (2) By relying on spatiotemporal registration, multi-factor weight fusion and error compensation optimization, data errors and interpretation deviations are reduced, and the accuracy of complex stratigraphic structure identification and spatial distribution restoration is improved.

[0037] (3) Introduce intelligent algorithms to realize automated data processing and anomaly identification, reduce reliance on manual labor, improve data processing efficiency and consistency of results, and provide reliable and high-precision geological support for geological engineering.

[0038] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the adaptive multi-source detection system module composition of a high-precision three-dimensional detection method for multi-factor fusion in complex strata according to the present invention;

[0040] Figure 2 This is a flowchart of the stratigraphic zoning process of a two-stage detection scheme for a high-precision three-dimensional detection method for complex strata with multi-factor fusion according to the present invention.

[0041] Figure 3 This is a schematic diagram of the differentiated detection data acquisition points and parameters of a high-precision three-dimensional detection method for multi-factor fusion in complex strata according to the present invention;

[0042] Figure 4 This is a schematic diagram of the spatiotemporal registration coordinate transformation of multi-source data in a high-precision three-dimensional exploration method for complex strata based on multi-factor fusion according to the present invention;

[0043] Figure 5 This invention presents a multi-factor weighted fusion model and a result diagram of stratigraphic anomaly identification using a multi-factor fusion high-precision three-dimensional detection method for complex strata.

[0044] Figure 6This is a diagram showing the error compensation and three-dimensional results of a multi-factor fusion high-precision three-dimensional exploration method for complex strata according to the present invention. Detailed Implementation

[0045] To make the technical solutions, advantages, and objectives of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below. The described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention without creative effort are within the protection scope of this application.

[0046] Unless otherwise defined, the technical or scientific terms used in this invention shall have the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.

[0047] like Figures 1-6 As shown, the present invention provides a high-precision three-dimensional detection method for complex strata by fusing multiple factors, comprising:

[0048] S1. Construct an adaptive multi-source detection system, which includes an improved tube wave testing module, a dual-mode borehole sonar module, an intelligent elastic wave CT module, and a zoned water transient electromagnetic module; wherein, the improved tube wave testing module consists of a super magnetostrictive source and a directional receiving array, the dual-mode borehole sonar module is equipped with a multi-frequency sonar probe and a 360° rotating scanning mechanism, the intelligent elastic wave CT module integrates a high-frequency source and a high-sensitivity detector, and the zoned water transient electromagnetic module adopts a multi-channel receiving coil;

[0049] S2. A two-level detection scheme is generated based on the stratigraphic parameters of the detection area: In the first stage, a transmissive electromagnetic module is used to deploy survey lines in the zoned water area, covering the distribution range of soluble rocks in the detection area. The resistivity anomaly area below the overburden is extracted through data interpretation. Combined with the verification results of geological boreholes, a three-dimensional classification model of resistivity-solubleness-tectonic fracture density is constructed to divide the strata into slightly complex, medium-complex, and highly complex layers. In the later stage, corresponding detection module combinations are matched for stratigraphic sections with different complexities.

[0050] The criteria for determining a slightly complex layer are that the density of tectonic fractures is less than a set threshold and there is no obvious karst development. The criteria for determining a moderately complex layer are that the density of tectonic fractures is between two set thresholds and there are local small caves. The criteria for determining a highly complex layer are that the density of tectonic fractures is greater than a set threshold and there are large caves or dissolution zones.

[0051] S3. Implement differentiated detection data acquisition according to a two-level detection scheme: During the acquisition process, the spatial layout of each module meets the following requirements: The transient electromagnetic module for zoned water areas is laid out parallel to the direction of the detection area, and the spacing between adjacent measurement lines is set according to the complexity of the strata; the improved pipe wave test module and the dual-mode borehole sonar module are arranged along the direction of the measurement line corresponding to the boreholes, and the spacing between the boreholes is adapted to the detection requirements of different complex levels; the intelligent elastic wave CT module is deployed across the boreholes, and the cross-hole connection line is distributed perpendicularly or obliquely to the measurement line to form a grid-like detection layout; each module acquires corresponding data, the improved pipe wave test module acquires the borehole-side reflected wave signal, the dual-mode borehole sonar module acquires the karst cave cross-section data through the scanning mechanism, the intelligent elastic wave CT module acquires the cross-hole wave velocity data according to the set point spacing, and the transient electromagnetic module for zoned water areas acquires the resistivity distribution data below the overburden layer;

[0052] For micro-complex layers, a combination of an improved tube wave testing module as the main component and a dual-mode borehole sonar module as a supplementary component is used. For medium-complex layers, a combination of an improved tube wave testing module and an intelligent elastic wave CT module for collaborative detection is used. For highly complex layers, a combination of an intelligent elastic wave CT module as the main component and a dual-mode borehole sonar module for three-dimensional scanning is used.

[0053] The spacing between adjacent measuring lines of the transient electromagnetic module in the zoned water area is greater in the micro-complex layer region than in the medium-complex layer region, and greater in the medium-complex layer region than in the high-complex layer region; the corresponding borehole spacing of the improved tube wave test module and the dual-mode borehole sonar module is greater in the micro-complex layer region than in the medium-complex layer region, and greater in the medium-complex layer region than in the high-complex layer region.

[0054] S4. Spatiotemporal registration of multi-source data: Align the time dimension of data from each module based on timestamps, and unify the coordinate system through a spatial coordinate transformation matrix, wherein the coordinate transformation matrix satisfies:

[0055] ;

[0056] Where (X,Y,Z) are coordinates in a unified coordinate system. For the first The original coordinates of the detection module, where M is a 4×4 spatial transformation matrix;

[0057] S5. Preprocessing and feature extraction of the registered data: The data fusion algorithm is used to construct the echo time delay matrix to form a signal array, repair missing data and eliminate data overlap error; the ranging algorithm is used to distinguish between target echo and interference echo; the semantic segmentation model with lithological feature attention mechanism is used to identify stratigraphic anomaly areas by inputting laser point cloud parameters and wave velocity values ​​obtained by the intelligent elastic wave CT module.

[0058] The data fusion algorithm constructs a complete signal array by performing matrix operations on the echo time difference of the data collected by each module, filling in missing data segments and eliminating overlapping data.

[0059] S6. Construct a multi-factor weighted fusion model: Dynamically adjust the weights of each detection module based on formation characteristics, and correct the weight coefficients in conjunction with groundwater distribution. The fusion formula satisfies:

[0060] ;

[0061] In the formula, For the fused 3D coordinates, C, S, and E are the inversion coordinates of the improved pipe wave testing module, the dual-mode borehole sonar module, and the intelligent elastic wave CT module, respectively. , , The weighting coefficients and ;

[0062] When the groundwater depth is shallow, increase the weighting coefficient of the dual-mode borehole sonar module; when the groundwater depth is medium, increase the weighting coefficient of the intelligent elastic wave CT module; when the groundwater depth is deep, increase the weighting coefficient of the improved pipe wave testing module.

[0063] S7. Error Compensation and Optimization of the Fusion Model: A 3D model is constructed through mesh generation. An adaptive slope correction formula is used to correct height errors. The initial 3D model is dynamically sectioned and corrected using the modeling platform. A plugin enables data exchange between the modeling platform and the mesh generation software, converting the mesh file to a specified format and outputting a 3D point cloud model and a digital elevation model. The adaptive slope correction formula is:

[0064] ;

[0065] In the formula, As compensation value, Here, k is the original elevation, and k is the slope correction factor. This represents the actual slope angle of the strata.

[0066] The size of the grid cells in the mesh is set according to the required detection accuracy; the higher the required detection accuracy, the smaller the size of the grid cells.

[0067] S8. Parameter adaptation of the dual-mode borehole sonar module: Select the sonar probe of the corresponding frequency according to the geological type of the detection area, and adjust the scanning angle and probe coupling state through the 360° rotating scanning mechanism to ensure good coupling between the probe and the borehole wall.

[0068] The dual-mode borehole sonar module uses a high-frequency probe in dense formations and a low-frequency probe in loose or fractured formations.

[0069] S9. Set parameters and protect the intelligent elastic wave CT module: Set the sampling interval, filter passband, number of superpositions and excitation energy parameters according to the formation development section, and bury a protective pipe in the detection borehole to avoid the probe getting stuck;

[0070] The inter-aperture spacing of the intelligent elastic wave CT module is greater in the micro-complex layer region than in the medium-complex layer region, and greater in the medium-complex layer region than in the high-complex layer region. The excitation energy is less in the micro-complex layer region than in the medium-complex layer region, and less in the medium-complex layer region than in the high-complex layer region.

[0071] S10. Results Verification: Select verification boreholes, compare the stratigraphic information revealed by the boreholes with the predicted information of the fusion model, and use borehole radar to assist in verification to confirm whether the detection accuracy meets the requirements.

[0072] Verification boreholes should cover geological sections of varying complexity, with at least two verification boreholes set up for each complex geological section. The detection depth of the borehole radar should be consistent with the detection depth range of the fusion model.

[0073] Example

[0074] This embodiment uses a highway subgrade survey project in a karst development area in southern China as an application scenario. The lithology of this area is mainly carbonate rock, with complex geological bodies such as karst caves, dissolution fissures, and underground rivers. It is necessary to use high-precision 3D exploration to clarify the spatial distribution of adverse geological bodies, so as to provide a basis for subgrade design and construction safety. The specific implementation steps are as follows:

[0075] (1) Preliminary preparations and system setup:

[0076] An adaptive multi-source detection system is constructed, with the following module configurations and parameters:

[0077] Improved tube wave test module: equipped with a 500W super magnetostrictive source, a 16-channel directional receiving array, a receiving frequency range of 50-2000Hz, and a signal sampling rate of 10kHz to ensure accurate acquisition of reflected wave signals within a range of 5-15m from the hole;

[0078] Dual-mode borehole sonar module: Equipped with a dual-frequency sonar probe suitable for 500kHz dense formations and 1MHz loose formations, and equipped with a servo motor driven 360° rotating scanning mechanism, with a scanning step size that can be set to 1°-5° and a probe coupling pressure adjustment range of 0.2-0.5MPa.

[0079] Intelligent elastic wave CT module: integrates a high-frequency piezoelectric source with a main frequency of 1-5kHz and 24 high-sensitivity detectors, supports a maximum cross-hole distance of 50m, the sampling interval can be set to 20-100μs, and the filter passband can be adjusted to 100-2000Hz according to the formation noise characteristics.

[0080] Transient electromagnetic module for zoned water areas: Employs multi-channel receiving coils with a coil area of ​​10m² 2 The transmission frequency is 5-500Hz, the measurement delay time is 0.1-100ms, and the data acquisition density is 5m between each measurement point to ensure continuous acquisition of resistivity data within a depth of 30m below the cover layer.

[0081] Auxiliary equipment: Three GPS / BeiDou dual-mode positioning reference points are set up with a coordinate accuracy of ±3cm. One data acquisition and control terminal is set up at the edge of the detection area. Synchronous control and data transmission with each module are achieved through 4G / fiber optic links, and the synchronization error is controlled within 10ms.

[0082] (2) Formulation of a two-stage detection scheme:

[0083] Preliminary stratigraphic analysis and survey line layout:

[0084] Transient electromagnetic modules for zoned water areas were used to lay out survey lines along the highway subgrade. The total length of the survey lines was 2km, with 5 parallel survey lines laid out. The initial spacing between adjacent survey lines was set at 50m, and was subsequently dynamically adjusted according to the complexity.

[0085] The collected resistivity data were interpreted to extract anomalous areas (suspected caves, underground rivers) with resistivity <50Ω·m below the overburden. Combined with data from three existing geological boreholes (30m deep) in the area, the lithology (limestone, dolomite), fracture density (1-8 fractures / m), and karst development (small caves with diameters of 0.5-3m) revealed by the boreholes were obtained.

[0086] A three-dimensional hierarchical model based on resistivity, solubility, and tectonic fracture density was constructed to divide the probe area into three stratigraphic segments:

[0087] Micro-complex layer: structural fracture density <2 fractures / m, no obvious karst, corresponding resistivity >100Ω·m, mainly distributed in survey line 1-2 (length about 600m).

[0088] Medium-complex layer: structural fracture density 2-5 fractures / m, local development of small karst caves (diameter <2m), corresponding resistivity 50-100Ω·m, mainly distributed in survey line 2-4 (length about 800m).

[0089] Highly complex layer: structural fracture density > 5 fractures / m, large karst caves (diameter > 2m) or underground rivers exist, corresponding resistivity < 50Ω·m, mainly distributed in survey line 4-5 (length about 600m).

[0090] Post-module combination matching:

[0091] Micro-complex layer: Matching the combination of "improved tube wave test module + dual-mode borehole sonar module (supplementary test)";

[0092] Medium-complex layer: Matching the synergistic combination of "improved tube wave testing module + intelligent elastic wave CT module";

[0093] Highly complex layers: Match the combination of "intelligent elastic wave CT module + dual-mode borehole sonar module (3D scanning)" and simultaneously retain the transient electromagnetic module of the zoned water area for deep resistivity verification.

[0094] (3) Differentiated data acquisition and equipment layout:

[0095] The equipment spatial layout and acquisition parameters are adjusted according to different levels of complexity, as follows:

[0096]

[0097]

[0098] During the data acquisition process, Φ100mm PVC protective pipes were buried in all boreholes to prevent the probe from getting stuck; before sonar scanning, the borehole walls were cleaned with circulating water to ensure good coupling between the probe and the borehole walls.

[0099] (4) Spatiotemporal registration of multi-source data:

[0100] Time dimension alignment: Extract the data acquisition timestamps of each module (accurate to milliseconds), and use the system time of the data acquisition control terminal as a reference to perform time correction on the tube wave, sonar, elastic wave CT, and transient electromagnetic data. For example, align the excitation time of the elastic wave CT source with the signal start time of the tube wave receiving array to ensure that the time difference of the same geological event (such as cave reflection) is <50ms.

[0101] Spatial coordinate unification: A local coordinate system is established based on three reference points (X-axis to the east, Y-axis to the north, and Z-axis vertically downwards). The original coordinates of each module are converted into a unified coordinate system using a 4×4 spatial transformation matrix M.

[0102] ;

[0103] in, The original coordinates of each module (such as the depth of the sonar probe in the borehole and the position of the elastic wave CT detector) are converted, and the coordinate deviation is controlled within ±5cm to ensure that the spatial positions of the multi-source data correspond one-to-one.

[0104] (5) Data preprocessing and feature extraction:

[0105] Missing data repair and interference removal:

[0106] An echo time delay matrix was constructed using a data fusion algorithm. This matrix was used to process three missing data points caused by signal attenuation in the pipe wave test and two overlapping segments of cross-hole data in the elastic wave CT: missing segments were filled by interpolation of the time difference between adjacent borehole data, and interference signals with a signal-to-noise ratio <3:1 (such as ground vibration and electromagnetic noise) in the overlapping data were removed.

[0107] A ranging algorithm is used to distinguish target echoes from interference echoes using sonar data: the echo distance is calculated ( , For the speed of sound, (Echo time), eliminating interfering echoes (such as reflections from air bubbles inside the borehole) that are more than 50cm away from the borehole diameter.

[0108] Identification of stratigraphic anomaly areas:

[0109] A semantic segmentation model incorporating a lithological feature attention mechanism was constructed. The input parameters were laser point cloud parameters (point cloud density, reflection intensity) and elastic wave CT wave velocity values. The model training samples used geological information (karst caves, fissures, normal strata) revealed by known boreholes.

[0110] Based on model identification, a total of 12 abnormal karst cave areas (areas ranging from 5 to 80 m²) were extracted. 2 ), 8 densely fractured zones (100-300m in length and 5-15m in width), and 1 underground river channel (parallel to survey line 4-5, 3-5m in width and 15-25m in depth), with anomaly area boundary identification accuracy error <1m.

[0111] (6) Construction of multi-factor weight fusion model:

[0112] Weighting coefficients are determined by dynamically adjusting the weights based on the groundwater depth distribution in the detection area (20-25m for slightly complex layers, 15-20m for moderately complex layers, and 5-15m for highly complex layers).

[0113] Micro-complex layers (deep groundwater): Weighting of the improved pipe wave testing module =0.4, weight of dual-mode borehole sonar module =0.3, weight of intelligent elastic wave CT module =0.3 (elastic wave CT is only used for verification of edge regions);

[0114] Medium-complex layer (medium groundwater depth): =0.3, =0.2, =0.5 (based primarily on elastic wave CT cross-hole profile data);

[0115] Highly complex layers (with shallow groundwater depth): =0.2, =0.5, =0.3 (Sonar 3D scanning is more suitable for depicting the morphology of shallow caves); the weight coefficients of all regions satisfy the condition that the sum of the weights is 1.

[0116] 3D coordinate fusion calculation: The fusion formula is used to fuse the inverted coordinates of multi-source data.

[0117] ;

[0118] In the formula, For the fused three-dimensional coordinates, C is the stratigraphic interface coordinates inverted by the tube wave test (error ±0.5m), S is the karst cave boundary coordinates inverted by the sonar scan (error ±0.3m), and E is the center coordinates of the karst zone inverted by the elastic wave CT (error ±0.8m).

[0119] After fusion, a three-dimensional coordinate model of the entire detection area is obtained, and the spatial position deviation of the adverse geological bodies is controlled within ±0.6m.

[0120] (7) Fusion model error compensation and optimization:

[0121] Slope adaptive correction: Since the detection area has local slopes (5°-15°), the slope adaptive correction formula is used to correct the height error.

[0122] ;

[0123] In the formula, As compensation value, The original elevation is given by the inversion function of each module, and k is the slope correction factor (set according to the lithology of the strata: k=0.95 for limestone and k=0.98 for dolomite). The actual slope angle of the strata (obtained by GPS measurement);

[0124] After correction, the height error was reduced from ±0.8m to ±0.3m, ensuring the smoothness of the stratum interface on the slope section.

[0125] 3D model construction and optimization:

[0126] The fused data was divided into 1m×1m×0.5m grid cells (0.5m×0.5m×0.2m for highly complex layers) using mesh generation software to generate an initial 3D model;

[0127] The plugin enables data exchange between the mesh generation software and the Surfer modeling platform, converting mesh files to .obj format and performing dynamic sectioning correction on the model in three directions: X (east-west), Y (north-south), and Z (vertical). For example, it can section along the direction of the underground river channel area to correct the jagged error of the channel width in the model.

[0128] The final output is a 3D point cloud model (point cloud density 20 points / m). 3 It features a digital elevation model (DEM) with a resolution of 1m, which allows for cross-sectional viewing in any direction and provides clear visualization of adverse geological features.

[0129] (8) Module parameter adaptation and protection:

[0130] Dual-mode drilling sonar module adaptation:

[0131] For micro-complex layers (dense limestone strata), a 500kHz probe was used with a scanning angle of 180° and a coupling pressure of 0.3MPa.

[0132] For medium-complex layers (interbedded limestone and dolomite), switch probes according to the borehole lithology: 500kHz for dense sections and 1MHz for loose and dissoluted sections, with a scanning angle of 270° and a coupling pressure of 0.25MPa.

[0133] For highly complex layers (intense dissolution and instability within the pore), a 1MHz probe was used for 360° full-angle scanning, with the coupling pressure increased to 0.4MPa, while the scanning speed was reduced (5° / s) to avoid the probe colliding with the cave wall.

[0134] Intelligent elastic wave CT module setup and protection:

[0135] Micro-complex layer: Two sets of trans-hole pairs are arranged only in the edge region, with a sampling interval of 100μs, a filter passband of 500-1500Hz, 15 stacking times, and an excitation energy of 300J.

[0136] Medium-complex layer: sampling interval 50μs, filter passband 200-1500Hz, stacking times 20 times, excitation energy 400J, transap spacing 30m;

[0137] Highly complex layer: sampling interval 20μs, filter passband 100-2000Hz, stacking times 30 times, excitation energy 500J, transap spacing 15m;

[0138] All exploratory boreholes are equipped with 5mm thick PVC protective pipes, with the bottom of the pipes sealed and the top covered to prevent mud and sand from entering; the elastic wave CT detector and the seismic source probe are covered with rubber protective sleeves to prevent collision damage.

[0139] (9) Results verification:

[0140] Verification borehole layout and comparison: Six verification boreholes (one in a slightly complex layer, two in a moderately complex layer, and three in a highly complex layer) that were not used in the modeling were selected within the exploration area. The borehole depth was 35m, and actual geological information was obtained through core drilling.

[0141] Verification borehole 1 in a highly complex layer: revealed a karst cave with a depth of 18m and a diameter of 6m. The fusion model predicted a position deviation of 0.7m and a diameter deviation of 0.5m.

[0142] Verification borehole for underground river passage: revealed an underground river at a depth of 20m (water flow velocity 0.5m / s), the model predicted a deviation of 0.3m in the center depth and 0.4m in the width;

[0143] Verification boreholes in medium-complex strata revealed two densely fractured zones (burial depths of 10m and 15m), with model-predicted depth deviations of less than 0.5m.

[0144] Borehole radar-assisted verification: Borehole radar (frequency 100MHz) was used to assist in the detection of all verification boreholes. The radar images and the anomaly areas (cavities, fissure zones) predicted by the fusion model had a 93% agreement rate, with the agreement rate of 95% for highly complex layers. The verification results show that the detection accuracy meets the requirements of highway subgrade survey (allowable error <1m).

[0145] Using the method of this invention, the spatial distribution of adverse geological bodies within a 2km section of road in the karst development area was successfully determined. Compared with traditional single elastic wave CT detection, the identification rate of karst caves is improved, the positioning accuracy of underground river channels is improved, and the data processing efficiency is improved. The detection results are directly used for roadbed design optimization, such as adopting a bridge crossing scheme for highly complex underground river areas and a grouting reinforcement scheme for densely fractured areas in medium-complex layers, which effectively avoids construction risks and reduces engineering costs.

[0146] Therefore, this invention provides a high-precision three-dimensional detection method for complex strata by multi-factor fusion, which solves the problems of traditional methods that rely on single technologies or simple data superposition, lack deep fusion of multi-source data, fail to fully consider the coupling effects of multiple factors such as strata lithology and solubility, have fixed detection parameter processes, rely on manual data processing, and suffer from insufficient detection depth signal attenuation and terrain blind zone coverage. It effectively improves the accuracy of complex strata structure identification and spatial distribution reconstruction, enhances detection adaptability and intelligence, improves data processing efficiency and result consistency, realizes systematic detection, and provides reliable high-precision geological information support for geological disaster early warning and engineering safety assurance.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the technical solutions of the present invention, and these modifications or equivalent substitutions cannot cause the modified technical solutions to deviate from the spirit and scope of the technical solutions of the present invention.

Claims

1. A complex stratum multi-factor fusion high-precision three-dimensional detection method, characterized in that, The method comprises the following steps: S1, constructing an adaptive multi-source detection system, which comprises an improved tube wave test module, a dual-mode borehole sonar module, an intelligent elastic wave CT module, and a partitioned water transient electromagnetic module; S2, generating a two-level detection scheme based on the stratum parameters of the detection area, laying out measurement lines and covering the soluble rock distribution range through the partitioned water transient electromagnetic module in the early stage, interpreting the data to extract the resistivity anomaly area, combining with the geological borehole verification to construct a three-dimensional hierarchical model, and dividing the stratum into micro-complex layer, medium-complex layer, and high-complex layer, and matching the corresponding detection module combination in the later stage; S3, implementing differentiated data collection according to the two-level detection scheme, and each module is laid out according to the preset spatial layout and collects corresponding geological data; S4, performing space-time registration on the multi-source data to ensure space-time consistency; S5, pre-processing the registered data and extracting features to identify abnormal stratum areas; S6, constructing a multi-factor weight fusion model, dynamically adjusting the weights of each module based on the stratum characteristics, and correcting the weight coefficients combined with the underground water distribution to calculate the three-dimensional coordinates after fusion; S7, error compensation and optimization of the fusion model, outputting the three-dimensional point cloud model and digital elevation model; S8, parameter adaptation of the dual-mode borehole sonar module; S9, parameter setting and protection of the intelligent elastic wave CT module; S10, selecting verification boreholes covering different complex stratum sections, comparing the actual information of the stratum with the predicted information of the fusion model, using borehole radar for auxiliary verification, and confirming whether the detection accuracy meets the requirements; In S3, the micro-complex layer adopts the combination mode of the improved tube wave test module as the main module and the dual-mode borehole sonar module for supplementary measurement, the medium-complex layer adopts the combination mode of the improved tube wave test module and the intelligent elastic wave CT module for collaborative detection, and the high-complex layer adopts the combination mode of the intelligent elastic wave CT module as the main module and the dual-mode borehole sonar module for three-dimensional scanning; the interval between adjacent measurement lines and the interval between boreholes are all greater in the micro-complex layer region than in the medium-complex layer region, and greater in the medium-complex layer region than in the high-complex layer region; the spatial layout of each module is that the measurement lines of the partitioned water transient electromagnetic module are parallel to the strike of the detection area, the corresponding boreholes of the improved tube wave test module and the dual-mode borehole sonar module are arranged along the measurement lines, the intelligent elastic wave CT module is laid out in a grid-like manner by cross-hole arrangement, and each module collects data related to hole-side reflected waves, cave sections, cross-hole wave velocities, and resistivity distributions; In S7, the error compensation and optimization of the fusion model, the specific steps of outputting the three-dimensional point cloud model and the digital elevation model include constructing a three-dimensional model through grid subdivision, correcting the height error by using a slope adaptive correction formula, dynamically cutting and correcting the three-dimensional initial model on the modeling platform, and outputting the three-dimensional point cloud model and the digital elevation model; the grid element size of the grid subdivision decreases as the detection accuracy requirement increases, and the slope adaptive correction formula is: ； In the formula, is a compensation value, is the original elevation, k is a slope correction coefficient, is the actual slope angle of the stratum.

2. The multi-factor fusion high-precision three-dimensional exploration method for complex formations according to claim 1, characterized in that, The improved tube wave testing module is composed of an ultra-magnetic vibration source and a directional receiving array, the dual-mode borehole sonar module is configured with a multi-frequency sonar probe and a 360° rotary scanning mechanism, the intelligent elastic wave CT module integrates a high-frequency vibration source and a high-sensitivity detector, and the partitioned water area transient electromagnetic module adopts a multi-channel receiving coil.

3. The method of claim 1, wherein the method is characterized by, In S2, the three-dimensional hierarchical model is a three-dimensional hierarchical model of resistivity-solubility-tectonic fissure density; the micro-complex layer determination criterion is that the tectonic fissure density is less than a set threshold value and no obvious karst development exists; the medium-complex layer determination criterion is that the tectonic fissure density is between two set threshold values and local small-sized caves exist; and the high-complex layer determination criterion is that the tectonic fissure density is greater than a set threshold value and large-sized caves or dissolution zones exist.

4. The method of claim 1, wherein the method is characterized by, In S4, the spatio-temporal registration of the multi-source data specifically includes aligning the time dimensions of the data of each module based on timestamps and unifying the coordinate systems through a spatial coordinate conversion matrix, which satisfies: ； Wherein, (X, Y, Z) is the coordinate of the unified coordinate system, is the first Class detection module original coordinates, M is a 4 × 4 space conversion matrix.

5. The method of claim 1, wherein the method is characterized by, In S5, the pre-processing and feature extraction of the registered data to identify abnormal stratum regions specifically include repairing missing data and eliminating data overlap errors by using a data fusion algorithm, distinguishing target and interference echoes by using a ranging algorithm, and identifying abnormal stratum regions by using a semantic segmentation model with a lithology feature attention mechanism; the data fusion algorithm performs matrix operations on the echo time differences of the data collected by each module, constructs a complete signal array, fills in the missing data segments, and eliminates the data overlap parts.

6. The method of claim 1, wherein the method is characterized by, In S6, the three-dimensional coordinates after fusion are calculated by using a fusion formula, which satisfies: ； In the formula, C, S, and E are the inversion coordinates of the improved tube wave testing module, the dual-mode borehole sonar module, and the intelligent elastic wave CT module, respectively, , , is a weight coefficient and ; When the groundwater depth is shallow, the weight coefficient of the dual-mode borehole sonar module is increased When the groundwater depth is at a medium level, the weight coefficient of the intelligent elastic wave CT module is increased When the groundwater depth is deep, the weight coefficient of the improved tube wave test module is increased .

7. The method of claim 1, wherein the method is characterized by, In S8, the parameter adaptation of the dual-mode borehole sonar module specifically includes selecting a sonar probe with a corresponding frequency according to the stratum type of the detection area, adjusting the scanning angle and the probe coupling state to ensure good coupling between the probe and the borehole wall; wherein the frequency of the sonar probe of the dual-mode borehole sonar module is a high-frequency probe in dense strata and a low-frequency probe in loose or broken strata.

8. The method of claim 1, wherein the method is characterized by, In S9, the specific steps of parameter setting and protection of the intelligent elastic wave CT module include setting core detection parameters according to the stratum development section and embedding a protection pipeline in the detection borehole; the cross-hole spacing of the intelligent elastic wave CT module is greater in the micro-complex layer region than in the medium-complex layer region, and greater in the medium-complex layer region than in the high-complex layer region, and the excitation energy is the opposite; the detection core parameters include sampling interval, filter passband, stacking number and excitation energy.

Images