3D geological radar AI processing and interpretation platform

Abstract

The present invention relates to the technical field of geological radar signal processing, and discloses a three-dimensional geological radar AI processing and interpretation platform, comprising: a geological causal structure diagram construction module, which obtains a geological causal structure diagram that characterizes the causal dependency relationship between geological features and radar signals; a causal-driven adaptive filtering module, which obtains an adaptive filter that can selectively retain geological feature information; a signal-geology bidirectional mapping module, which realizes the mutual conversion between signal features and geological features; a dual-loop iterative optimization module, which obtains an optimized clean geological radar signal; and a causal-physics dual-loop framework is constructed. Through geological causal diagrams, adaptive filtering, bidirectional mapping and iterative optimization, the recovery rate of geological features in a high-noise environment is improved, and the interpretability and consistency are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of geological radar signal processing technology, and more specifically, to a three-dimensional geological radar AI processing and interpretation platform. Background Art

[0002] There are multiple technical challenges in processing geological radar signals in complex geological environments. There is a contradiction between signal fidelity and noise removal: traditional filtering methods are difficult to effectively remove noise while retaining key geological information, especially in weak signal areas, where valuable geological features are often misjudged as noise; the interpretation of geological radar signals is highly dependent on the experience of experts and lacks objective and unified interpretation standards, resulting in inconsistencies in interpretation results, and interpretation results between different experts may vary greatly;

[0003] Existing methods have difficulty effectively integrating physical laws and geological knowledge into the signal processing process, resulting in the separation of signal processing and geological interpretation, and the inability to form effective synergy; in extreme noise environments (signal-to-noise ratio less than 5dB), conventional methods have difficulty extracting effective geological features from severely noise-contaminated data; automated signal processing methods usually lack the explainability of processing decisions, making it difficult for experts to evaluate and trust the processing results. Summary of the Invention

[0004] The present invention provides a three-dimensional geological radar AI processing and interpretation platform to solve the technical problem of difficulty in geological radar signal processing and extraction in complex geological environments in related technologies.

[0005] The present invention provides a 3D geological radar AI processing and interpretation platform, which is characterized by including:

[0006] A geological causal structure diagram construction module is used to construct a geological causal structure diagram. Based on the constraint-based causal discovery algorithm, a geological causal structure diagram is obtained to characterize the causal dependency relationship between geological features and radar signals.

[0007] A causal-driven adaptive filtering module constructs a causal-driven adaptive filter and uses a causal-weighted adaptive threshold algorithm to obtain an adaptive filter that can selectively retain geological feature information.

[0008] The signal-geology bidirectional mapping module builds a signal-geology bidirectional mapping model, establishes a bidirectional mapping relationship between the signal domain and the geological domain, and realizes the mutual conversion between signal characteristics and geological characteristics;

[0009] The dual-loop iterative optimization module, based on the dual-loop iterative optimization, uses a causal-driven dual-domain alternating optimization algorithm to perform iterative optimization between the signal domain and the geological domain to obtain the optimized clean geological radar signal;

[0010] The causal interpretation module receives the geological causal structure diagram and the iterative optimization process record as input, and outputs a causal interpretation report containing textual explanations, visual charts, and interactive exploration components;

[0011] The feedback learning module takes the functional modules constructed above as input and outputs an integrated system with continuous learning capabilities and a continuously updated causal knowledge base.

[0012] Furthermore, the construction of the geological causal structure diagram includes:

[0013] Standardize, detrend and detect outliers on historical geological radar data to obtain preprocessed datasets;

[0014] Extract physical constraints from electromagnetic wave propagation theory and geological knowledge to form a constraint set;

[0015] The causal graph skeleton is learned using a constraint-based PC algorithm that determines the conditional independence between nodes through the conditional independence test statistic:

[0016] ;

[0017] in is the conditional independence test statistic, For node sets For any two nodes in is the conditional independence test statistic, is the set of physical constraints, is the basic condition independence test value, Respectively and nodes, is the condition set, It is The influence factor of a physical constraint on the independence judgment of the condition, Indicates from arrive The continuous multiplication operation, is the total number of physical constraints;

[0018] Based on the V-shaped structure, acyclic constraints and physical prior knowledge, the directions of the edges in the graph are determined to obtain a directed acyclic graph.

[0019] Furthermore, the construction of the causal-driven adaptive filter includes:

[0020] Wavelet transform is used to decompose the original geological radar signal into coefficients of multiple frequency scales;

[0021] Establish the mapping relationship between wavelet coefficients and signal feature nodes in the causal structure diagram;

[0022] Calculate the causal protection coefficient of the signal feature node based on the causal structure diagram:

[0023] ;

[0024] in Signal feature node The causal protection coefficient, which represents the importance of signal characteristics to the inference of geological characteristics; is a subset of geological feature nodes; is a subset of signal feature nodes; For slave nodes To Node the causal strength of Indicates that for all geological feature nodes sum; is a set of geological feature nodes, and is the node index, For slave nodes To Node the causal strength of

[0025] Constructing an adaptive threshold function for fusion causal protection in the wavelet domain:

[0026] ;

[0027] in For the Layer wavelet detail coefficients are at positions The threshold value at For the The base threshold of the layer, is the causal protection strength parameter, is the causal protection coefficient of the corresponding signal characteristic node, is a local geological feature significance function, quantifying the location the strength of geological features that may be present at the location;

[0028] An improved soft threshold function is used to process the wavelet coefficients and reconstruct the filtered signal.

[0029] Furthermore, the construction of the signal-geology bidirectional mapping model includes:

[0030] Construct a forward mapping model to achieve the conversion from signal features to geological features;

[0031] Construct a reverse mapping model to achieve the conversion from geological features to signal features;

[0032] Construct a bidirectional consistency measurement function to ensure the self-consistency of the bidirectional mapping;

[0033] Set a joint optimization objective that integrates forward mapping loss, reverse mapping loss, bidirectional consistency loss, physical consistency loss, and causal consistency regularization term;

[0034] The bidirectional mapping model is trained through an alternating optimization strategy to achieve mutual conversion between the signal domain and the geological domain.

[0035] Furthermore, the double-loop iterative optimization includes:

[0036] Initial processing of the raw signal using a causal-driven adaptive filter;

[0037] inferring geological features using forward mapping models;

[0038] Use the reverse mapping model to generate the corresponding theoretical signal;

[0039] Adaptively fuse the filtered signal with the theoretical signal;

[0040] Adjust the parameters of the adaptive filter according to the geological characteristics and signal quality of the current iteration;

[0041] Re-filter the original signal using the updated filter parameters;

[0042] Calculate the changes in signals and geological characteristics between two adjacent iterations, and stop the iteration when the convergence condition is met.

[0043] Furthermore, the causal explanation module includes:

[0044] Compare the original signal and the processed signal to identify areas where significant changes have occurred;

[0045] Analyze the cause of each key change area to determine whether it is noise removal or signal enhancement;

[0046] For each key area of ​​change, trace the causal paths that influenced treatment decisions;

[0047] Mapping the set of interpretation pathways to geophysical laws;

[0048] Quantitative assessment of uncertainty in interpretation of results;

[0049] Generate multi-level explanatory information based on user needs;

[0050] Generate intuitive visual explanation components;

[0051] Generate causal explanation reports that include textual explanations, visual charts, and interactive exploration capabilities.

[0052] Furthermore, the feedback learning module includes:

[0053] Integrate various functional modules into a unified processing platform;

[0054] Build an expert feedback collection component to obtain evaluations of processing results and correction suggestions;

[0055] Structuring and categorizing expert feedback;

[0056] Update the causal structure diagram based on expert feedback;

[0057] Optimize algorithm parameters of each module based on expert parameter feedback;

[0058] Improve interpretation criteria based on interpretation feedback;

[0059] Use incremental learning to continuously optimize the system;

[0060] Implement continuous monitoring and optimization mechanism of system performance.

[0061] Furthermore, the tracing of the causal paths that influence the processing decision includes:

[0062] Identify corresponding nodes in the signal feature space;

[0063] Identify relevant nodes in geological feature space;

[0064] Find all paths from signal feature nodes to geological feature nodes in the causal graph;

[0065] Calculate the importance score for each path:

[0066] ;

[0067] in represents the importance score of each path, For the path, is the edge on the path, is the corresponding causal strength value;

[0068] Several paths with the highest scores are selected as the explanation path set.

[0069] Furthermore, the updating of the causal structure diagram based on expert feedback includes:

[0070] For causal feedback, update the weight of the causal edge:

[0071] ;

[0072] in and are the causal edge weights before and after the update, is the smoothing factor, For experts The recommended weight value of

[0073] Add, remove, or modify causal edges to align with expert knowledge:

[0074] ;

[0075] in represents the updated causal edge set, represents the causal edge set before the update, is the edge set to be added, is the set of edges that need to be removed.

[0076] A computer-readable storage medium is used to store computer-readable instructions, which, when read by a computer, can run the three-dimensional geological radar AI processing and interpretation platform as described above.

[0077] The beneficial effects of the present invention are as follows: by constructing a causal physical double-loop framework, most key geological features can be restored under extreme conditions with a signal-to-noise ratio of less than 8 dB; through the guidance of causal relationships, a balanced optimization of signal fidelity and noise removal is achieved, resolving the contradiction between signal fidelity and noise removal in traditional methods; by introducing geological causal structures and bidirectional mapping models, the objectivity and consistency of geological interpretation are improved, and the dependence of interpretation results on personal experience is reduced; by tracing geological basis through causal paths, the interpretability of processing results is enhanced; and by continuously learning from expert feedback, adaptive learning and knowledge accumulation are achieved, thereby improving system performance and processing accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0078] Figure 1 It is a module diagram of the 3D geological radar AI processing and interpretation platform in the present invention. DETAILED DESCRIPTION

[0079] Example 1

[0080] At least one embodiment of the present invention discloses a 3D geological radar AI processing and interpretation platform, such as Figure 1 Shown, including:

[0081] Geological causal structure diagram construction module: construct geological causal structure diagram;

[0082] In at least one embodiment of the present invention, a geological causal structure diagram is obtained by a constraint-based causal discovery algorithm, which receives a historical geological radar data set (including geological radar data and geological feature data), a geological expert knowledge base and a set of physical constraints as input, and outputs a geological causal structure diagram (the geological causal structure diagram represents the causal dependency relationship between geological features and radar signals).

[0083] Data preprocessing: The historical geological radar data are standardized, detrended and processed for outlier detection to obtain the preprocessed data set.

[0084] Physical constraint extraction: Physical constraints are extracted from electromagnetic wave propagation theory and geological knowledge to form a constraint set.

[0085] Causal Skeleton Learning: Apply the constraint-based PC algorithm (PCPhyK) to learn the causal graph skeleton. The PC algorithm performs the following operations:

[0086] Initialize a fully connected undirected graph:

[0087] ;

[0088] in is an undirected graph, is a node set, which includes all geological features and signal features; is an edge set, initially an empty set;

[0089] For every pair of nodes in the node set and the condition set , calculate the conditional independence test statistic:

[0090] ;

[0091] in is the conditional independence test statistic, For node sets For any two nodes in is the conditional independence test statistic, is the set of physical constraints, is the basic condition independence test value, Respectively and nodes, is the condition set, It is The influence factor of a physical constraint on the independence judgment of the condition, Indicates from arrive The continuous multiplication operation, is the total number of physical constraints;

[0092] when hour( is the significance threshold), remove and The edge between.

[0093] Edge direction determination: The direction of the edges in the graph is determined based on the V-shaped structure, acyclic constraints, and physical prior knowledge to obtain a directed acyclic graph.

[0094] Causal strength quantification: Calculate the strength of each causal edge and use the causal effect quantification indicator:

[0095] ;

[0096] in Represents a slave node To Node The causal strength of the node Changes to Nodes The average impact magnitude of Representation node The amount of change, Representation node The amount of change; Express expectations.

[0097] Expert knowledge fusion: Integrate the causal relationship knowledge provided by experts with the results of data-driven learning. The specific methods are as follows:

[0098] Determine the confidence level in causal relationships derived from data-driven methods;

[0099] Assess the degree to which expert knowledge supports or opposes each causal relationship;

[0100] The causal relationship obtained by data-driven learning is corrected through the weight adjustment mechanism. When there is a conflict between expert knowledge and data-driven results, the final causal relationship is determined based on the degree of certainty of expert knowledge and the strength of data support.

[0101] The geological causal structure diagram obtained above is:

[0102] ;

[0103] in is a weighted directed graph. is a node set (including geological feature nodes and signal feature nodes), is a set of directed edges representing causal relationships, is the corresponding set of causal strength weights (i.e., the set containing all edge weights).

[0104] This geological causal structure diagram clearly characterizes the causal dependency between geological features and radar signal characteristics, providing a knowledge basis for subsequent signal processing and interpretation.

[0105] Causal-driven adaptive filtering module: builds a causal-driven adaptive filter;

[0106] In at least one embodiment of the present invention, a method for constructing an adaptive filter for geological radar signals through a causal weighted adaptive threshold algorithm is provided; the original geological radar signal and the obtained geological causal structure map are used as input, and the output is an adaptive filter that can selectively retain geological feature information based on geological causal relationships.

[0107] Multi-scale signal decomposition: Wavelet transform is used to decompose the original geological radar signal into coefficients of multiple frequency scales:

[0108] ;

[0109] in is the original geological radar signal, Respectively Layer approximation coefficient, Layer detail coefficient, Layer detail coefficient, is the number of wavelet transform layers.

[0110] Signal feature and causal node mapping: Establish the mapping relationship between wavelet coefficients and signal feature nodes in the causal structure diagram:

[0111] ;

[0112] in is the mapping function, is a subset of nodes in the causal structure diagram that represents signal characteristics.

[0113] Causal protection coefficient calculation: Calculate the causal protection coefficient of the signal feature node based on the causal structure diagram:

[0114] ;

[0115] in Signal feature node The causal protection coefficient, which represents the importance of signal characteristics to the inference of geological characteristics; is a subset of geological feature nodes; is a subset of signal feature nodes; For slave nodes To Node the causal strength of Indicates that for all geological feature nodes sum; is a set of geological feature nodes, and is the node index, For slave nodes To Node causal strength.

[0116] Construction of causal adaptive threshold function: Constructing an adaptive threshold function integrating causal protection in the wavelet domain:

[0117] ;

[0118] in For the Layer wavelet detail coefficients are at positions The threshold value at is the basic threshold, is the causal protection strength parameter, is the causal protection coefficient of the corresponding signal characteristic node, is a local geological feature significance function, quantifying the location The intensity of possible geological features at the location.

[0119] Adaptive soft threshold filtering implementation: Use the improved soft threshold function to process the wavelet coefficients:

[0120] ;

[0121] in is the processed wavelet coefficient, is a symbolic function, is the j-th layer detail coefficient after wavelet transform, is the jth layer wavelet detail coefficient at position The threshold value at .

[0122] Introduction of spatial correlation constraints: Introducing geological spatial correlation constraints in the filtering process to ensure the spatial continuity of geological features:

[0123] ;

[0124] in For location The neighborhood set of For location and The correlation coefficient between is the spatial constraint strength parameter, is the jth layer wavelet detail coefficient at position The processed wavelet coefficients at .

[0125] Signal reconstruction: Reconstruct the filtered signal using the processed wavelet coefficients:

[0126] ;

[0127] in is the filtered geological radar signal, represents the inverse wavelet transform operation, 、 They are respectively the Jth layer detail coefficient after processing, the J1th layer detail coefficient after processing, and the 1st layer detail coefficient after processing.

[0128] The obtained causal-driven adaptive filter incorporates geological causal relationship knowledge into the filtering process, can distinguish between noise and weak geological signals, selectively protect important geological features, and avoid the loss of geological information caused by excessive filtering. It is suitable for geological radar signal processing in complex geological environments with low signal-to-noise ratio.

[0129] Signal-geology bidirectional mapping module: builds a signal-geology bidirectional mapping model;

[0130] In at least one embodiment of the present invention, a signal-geology bidirectional mapping model is constructed, wherein the model receives as input a training dataset (including data pairs of signal features and geological features), a resulting geological causal structure map, and a resulting causal-driven adaptive filter, and generates as output a mapping model capable of bidirectional conversion between signal features and geological features;

[0131] The signal-geology bidirectional mapping model includes a forward mapping model and a reverse mapping model:

[0132] Build the forward mapping model (signal to geology):

[0133] Model Architecture Definition: Building a Causally Aware Deep Neural Network ,The structure includes the following layers: input layer (processing the filtered signal features), feature extraction layer, causal relationship layer, and geological feature prediction layer;

[0134] Causal structure mapping: Map the relationships in the geological causal structure diagram to network connections. The initial network weights are set based on the causal strength:

[0135] ;

[0136] in Connecting nodes in a neural network and The initial weight of For slave nodes To Node The causal strength of is the mapping function;

[0137] Definition of causal regularization term: Introducing causal consistency regularization term to ensure that the model learning results are consistent with the causal structure:

[0138] ;

[0139] in is the causal consistency regularization term, and is the regularization strength parameter, is the edge set in the causal graph, Represents all existing edges in the causal graph Sum, represents the sum of all non-existent edges in the causal graph.

[0140] Inverse mapping model (geology to signal) construction:

[0141] Model Architecture Definition: Building a Neural Network Based on Physical Constraints ,Its structure includes the following layers: input layer (receiving geological features), physical parameter inference layer, waveform simulation layer, and signal generation layer;

[0142] Integration of physical parameters and constraints: Embedding a physical model of electromagnetic wave propagation in the network to ensure that the generated signals conform to physical laws:

[0143] ;

[0144] in is the physical consistency loss metric, is the signal characteristic, For geological characteristics, Generate functions for signals based on physical models;

[0145] Geological continuity constraints: Introduce spatial continuity constraints of geological features:

[0146] ;

[0147] in is the spatial continuity constraint of geological features, For location The neighborhood point set of is the weight coefficient based on spatial distance, For location geological characteristics, For location geological characteristics.

[0148] Definition of bidirectional consistency measure: Construct a bidirectional consistency measure function between the signal domain and the geological domain:

[0149] ;

[0150] in For geological features; is the signal feature; It is a forward mapping model; It is a reverse mapping model; It is a two-way consistency measure that ensures the self-consistency of the two-way mapping, that is, the signals and geological features should be as close to the original data as possible after bidirectional transformation.

[0151] Joint optimization goal setting: Integrate various indicators and set the joint optimization goal of the bidirectional mapping model:

[0152] ;

[0153] in represents the total loss, and are the prediction losses for forward and reverse mapping, respectively, 、 、 、 is the coefficient for weighing each item.

[0154] Model training and validation:

[0155] Use the training dataset to train the model, using an alternating optimization strategy:

[0156] fixed Parameters, optimization Parameters;

[0157] fixed Parameters, optimization Parameters;

[0158] Jointly optimize the parameters of the two models to minimize the consistency loss;

[0159] Evaluate model performance using a validation dataset to ensure the model has sufficient generalization capabilities.

[0160] The constructed signal-geology bidirectional mapping model can realize the mutual conversion between the signal domain and the geological domain, in which the forward mapping ( ) converts the filtered radar signal characteristics into corresponding geological characteristics (including geological structure, physical parameters, etc.), and reverse mapping ( ) converts geological features into theoretical radar signals (including waveform, amplitude, phase, etc.). This model ensures the rationality and reliability of the conversion by introducing causal and physical constraints, providing basic support for subsequent dual-loop iterative optimization.

[0161] Double-loop iterative optimization module: realizes double-loop iterative optimization system;

[0162] In at least one embodiment of the present invention, a method for implementing a double-loop iterative optimization system is provided. The system uses a causal-driven dual-domain alternating optimization algorithm, receives the original geological radar signal, the geological causal structure diagram, the causal-driven adaptive filter and the signal-geology bidirectional mapping model as input, and outputs an optimized clean geological radar signal.

[0163] Initialization: Initial processing of the original signal using a causal-driven adaptive filter:

[0164] ;

[0165] in is the initial processing signal obtained in the 0th round of iteration, is the original geological radar signal, It is a causally driven adaptive filter.

[0166] Forward loop (signal → geology): for iterative rounds , use the forward mapping model to infer geological features:

[0167] ;

[0168] in For the The geological characteristics of the rounds of iterative inference, It is the forward mapping model in the signal-geology bidirectional mapping model. For the The processed signal obtained by round iteration.

[0169] Reverse loop (geology → signal): Use the reverse mapping model to generate the corresponding theoretical signal:

[0170] ;

[0171] in Based on geological characteristics The theoretical signal generated is It is the reverse mapping model in the signal-geology bidirectional mapping model.

[0172] Signal fusion and update: Adaptively fuse the filtered signal with the theoretical signal to obtain an updated signal:

[0173] ;

[0174] in is a position-dependent fusion weight function that is dynamically adjusted according to signal reliability and geological feature certainty. is the processed signal obtained in the round of iteration.

[0175] Causal feedback adjustment: Adjust the parameters of the adaptive filter based on the geological characteristics and signal quality of the current iteration:

[0176] ;

[0177] in For the The filter threshold used in the round iteration, is the change in protection factor calculated based on the current geological interpretation results, To adjust the rate parameter, For the The filter threshold used in the round iteration.

[0178] Re-filtering: Use the updated filter parameters to re-filter the original signal:

[0179] ;

[0180] in is the re-filtering result, is the original geological radar signal, For the The filter threshold used in the round iteration, It is a causally driven adaptive filter.

[0181] Further integrate the re-filtering result with the fusion result:

[0182] ;

[0183] in For the final clean geological radar signal, For the The processed signal obtained by round iteration, is the re-filtering result, To balance the parameters, control the relative importance of the two signals.

[0184] Convergence determination and early stopping: Calculate the changes in signals and geological characteristics between two adjacent iterations, and stop the iteration when the convergence condition is met or the maximum number of iterations is reached.

[0185] Counterfactual verification: The reliability of the processing results is evaluated by making small disturbances to geological features and observing the corresponding signal changes. When the disturbance response ratio exceeds a preset threshold (the preset threshold is the confidence level of the geological feature prediction in the signal-geology bidirectional mapping model), the corresponding area is marked as uncertain, indicating that further verification by experts is required.

[0186] The clean geological radar signal obtained through this iterative optimization has a higher signal-to-noise ratio and geological interpretability, effectively removing noise interference while preserving key geological features. This dual-loop iterative optimization system integrates signal processing and geological interpretation into a unified closed-loop optimization process. Through mutual verification and correction in the signal and geological domains, it continuously improves processing quality and is particularly suitable for processing high-noise radar data in complex geological environments.

[0187] Causal explanation module: builds a causal explanation system;

[0188] In at least one embodiment of the present invention, a method is provided for constructing an interpretable system for processing results using a causal path tracing algorithm. The system receives as input the original signal, the optimized clean signal, the geological causal structure diagram, and the iterative optimization process record, and outputs a causal interpretation report containing textual explanations, visual charts, and interactive exploration components.

[0189] Identification of key change areas: Compare the original signal and the processed signal to identify areas where significant changes have occurred:

[0190] ;

[0191] in, For location The signal variation at For the optimized clean signal at position The value at is the original signal at position The value at position ,when ( When is the threshold, the position is marked as a key change area.

[0192] Change attribution analysis: Analyze the cause of change in each key change area to determine whether it is noise removal or signal enhancement, and classify it based on the local signal-to-noise ratio and signal change amplitude.

[0193] Causal path tracing: For each key change area, corresponding nodes are identified in the signal feature space and geological feature space, and the important paths connecting them are found in the causal graph. Several paths with the highest scores are selected as the explanatory path set;

[0194] Calculate the importance score for each path:

[0195] ;

[0196] in For path The importance score of For the path, is the edge on the path, is the causal intensity value corresponding to the geological causal structure diagram;

[0197] Physical law mapping: Mapping the explanatory path set to geophysical laws, transforming abstract causal relationships into specific geophysical explanations.

[0198] Uncertainty Quantification: Quantitatively assess the uncertainty of interpretations:

[0199] ;

[0200] in Lower values ​​indicate higher certainty in the explanation. represents the summation symbol, For all possible causal paths, To explain the path set.

[0201] Multi-level explanation generation:

[0202] Three levels of explanation depth are generated based on user needs: foundational explanations (concise and concise descriptions of processing decisions), technical explanations (including processing parameters and algorithm details), and specialized explanations (including detailed geological causal relationships and physical principles).

[0203] Visualization: Generates intuitive visual explanation components, including pre- and post-processing signal comparison charts, causal path highlighting, decision uncertainty heatmaps, and A / B comparison views.

[0204] Interactive Exploration Interface: Provides a functional interface for experts to interactively explore explanations, allowing experts to click on areas of interest to obtain detailed explanations, adjust the level of explanation detail, explore hypothetical questions, and provide feedback.

[0205] The generated causal interpretation report includes textual explanations, visual charts, and interactive exploration components, transforming the "black box" process of signal processing into an interpretable, transparent process. By tracing causal paths, the system reveals the geophysical principles behind processing decisions, enabling experts to understand, evaluate, and trust the processing results, thereby improving the reliability and acceptance of geological radar data interpretation.

[0206] Feedback learning module: system integration and feedback learning mechanism implementation;

[0207] In at least one embodiment of the present invention, a method for implementing a system integration and feedback learning mechanism is provided. The method adopts a modular integration architecture and an incremental causal learning algorithm, takes the functional modules constructed as above as input, and outputs an integrated system with continuous learning capabilities and a continuously updated causal knowledge base.

[0208] Modular system integration: Integrate various functional modules (geological causal structure diagram, causal-driven adaptive filter, signal-geology bidirectional mapping model, dual-loop iterative optimization system, and causal interpretation system) into a unified processing platform. A loosely coupled architecture is adopted to ensure that the modules are independently upgradeable. This includes defining unified data interface standards, building a central scheduler, implementing a message passing mechanism between modules, and establishing a unified parameter management system.

[0209] Expert feedback collection: Build an expert feedback collection component to obtain evaluations and correction suggestions for processing results, including annotation interface support, structured feedback forms, interactive parameter adjustment functions, and expert explanation and evaluation collection mechanisms.

[0210] Feedback data structuring and classification: Expert feedback (including algorithm parameter feedback, interpretation feedback, and causal relationship feedback) is structured and classified:

[0211] ;

[0212] in For expert feedback, For feedback on causality, is feedback about the algorithm parameters, Provide feedback on interpretation results.

[0213] Causal knowledge update: Update the causal structure diagram based on expert feedback:

[0214] Feedback on causality , update the weights of causal edges:

[0215] ;

[0216] in is the smoothing factor (value range 0-1), For experts The recommended weight value of and are the causal edge weights before and after the update, respectively.

[0217] Add, remove, or modify causal edges based on expert knowledge:

[0218] ;

[0219] in and are the edge sets before and after the update, is the edge set to be added, is the set of edges that need to be removed.

[0220] Adaptive adjustment of parameters and interpretation criteria: Based on algorithm parameter feedback and interpretation feedback, optimize the algorithm parameters and interpretation criteria of each module. Use Bayesian optimization and other methods to integrate expert experience and historical data to achieve continuous optimization of system parameters and continuous improvement of interpretation criteria.

[0221] Incremental learning and knowledge transfer: Use incremental learning to continuously optimize the system, perform incremental training on newly collected data, and achieve knowledge transfer between different geological environments, so that the system can adapt to a wider range of application scenarios.

[0222] System performance monitoring and optimization: Establish a complete performance evaluation system, including indicators such as processing accuracy, efficiency, and consistency, and guide the iterative upgrade of the system through regular evaluation, resource optimization, and performance analysis.

[0223] The integrated system achieved through the above is a complete geological radar signal processing platform, equipped with causal-physical dual-loop processing capabilities and expert feedback learning capabilities. The system implements a closed-loop mechanism of "processing feedback learning optimization", which continuously improves processing capabilities and can adapt to signal processing requirements in different geological environments.

[0224] A computer-readable storage medium is used to store computer-readable instructions, which can run the above-mentioned three-dimensional geological radar AI processing and interpretation platform when the computer-readable instructions are read by a computer.

[0225] Here, the present invention provides an implementation example: This embodiment is applied during the construction of a high-speed railway tunnel. The engineering geological survey team needs to accurately detect the geological conditions ahead of the tunnel, especially to identify possible unfavorable geological bodies such as karst caves, faults, and karst fissures, in order to prevent safety accidents such as water inrush and collapse during the construction process.

[0226] The survey area is located in a complex karst development area with a complex stratum structure. It is also subject to significant interference from surrounding construction machinery, resulting in severe geological radar signal noise.

[0227] This application scenario has the typical characteristics of "high noise, multiple interferences, and complex geological structures", which poses severe challenges to geological radar signal processing. Traditional processing methods are difficult to obtain sufficiently clear geological structure images in this environment, resulting in insufficient reliability of survey results and posing potential risks to tunnel construction safety.

[0228] In this tunnel engineering project, we first constructed a geological causal structure map to reveal the causal relationship between radar signals and geological features in the karst area;

[0229] Data Collection and Preprocessing: We collected 142 sets of historical geological radar data from the region and similar geological conditions, including 36 validation borehole data with known karst features. These data were standardized and used for causal structure learning.

[0230] Physical constraint condition extraction: Based on the electromagnetic wave propagation characteristics in karst areas, the following were extracted: the relationship between electromagnetic wave velocity and dielectric constant of the medium, the relationship between interface reflection coefficient and dielectric constant difference, the relationship between water content in the karst area and signal amplitude attenuation, the relationship between fracture density and scattering attenuation, and the relationship between cave space and hyperbolic reflection characteristics;

[0231] Causal structure learning: Applying the PC algorithm based on physical constraints (PCPhyK) to conditional independence significance level = 0.05 for causal skeleton learning and integrating expert knowledge to determine the edge direction. The types and numbers of nodes in the constructed geological causal structure graph are shown in Table 1:

[0232] Table 1: Types and numbers of nodes in the constructed geological causal structure diagram

[0233]

[0234] The constructed causal structure diagram clearly demonstrates the causal dependencies between geological features and radar signal characteristics in karst areas. For example, groundwater level changes affect electromagnetic wave attenuation by altering the water content of the medium; karst caves affect the hyperbolic characteristics of the signal through boundary reflections; and fracture density influences the signal's spectral characteristics through scattering. These causal relationships provide physically meaningful knowledge guidance for subsequent signal processing.

[0235] Based on the constructed geological causal structure map, we developed a causal-driven adaptive filter for georadar signals in karst areas.

[0236] Wavelet decomposition and causal node mapping: The original radar signal is decomposed into five layers using the db5 wavelet basis, obtaining five detail coefficient layers and one approximation coefficient. The mapping relationship between the wavelet coefficients and the signal nodes of the causal graph is shown in Table 2:

[0237] Table 2: Mapping table of wavelet coefficients and causal graph signal feature nodes

[0238]

[0239] Calculation of causal protection coefficient: Based on the connection strength between each signal feature node and the geological feature node in the causal graph, the causal protection coefficients of different frequency bands are calculated as shown in Table 3:

[0240] Table 3: Causal protection coefficients for signals in different frequency bands

[0241]

[0242] Adaptive threshold function implementation: The parameters of the adaptive threshold function designed based on the causal protection coefficient are shown in Table 4:

[0243] Table 4: Adaptive threshold function parameter settings

[0244]

[0245] where σ is the noise level estimate, obtained by estimating in the non-target area.

[0246] 2.3 Signal-geology bidirectional mapping model and double-loop iterative optimization example

[0247] In the geological radar data processing of the area ahead of the tunnel, we constructed a signal-geology bidirectional mapping model for the karst environment and implemented a double-loop iterative optimization.

[0248] Bidirectional mapping model construction and training: Using the geological radar data and drilling verification data obtained in the previous tunnel project as the training set, the forward mapping (signal → geology) and reverse mapping (geology → signal) models were constructed, as shown in Table 5:

[0249] Table 5: Bidirectional mapping model architecture and parameters

[0250]

[0251] The training process adopts an alternating optimization strategy to minimize the following comprehensive loss function, as shown in Table 6:

[0252] Table 6: Loss function weight configuration for bidirectional mapping model training

[0253]

[0254] The model training employed an early stopping strategy, terminating training when the overall performance on the validation set reached optimality. The final model achieved an 83.6% accuracy in forward mapping geological feature prediction on the test set, and an average error of 0.14 in reverse mapping signal reconstruction.

[0255] Double-loop iterative optimization implementation: A double-loop iterative optimization system was applied to the high-noise radar profile in front of the tunnel. The key parameters of the iterative process are shown in Table 7:

[0256] Table 7: Double loop iterative optimization process parameters and convergence records

[0257]

[0258] After the sixth iteration, the changes in both the signal and geological features fell below the preset convergence threshold, and the system automatically stopped iterations. The optimized radar profile signal quality improved significantly, with the signal-to-noise ratio increasing from an initial 8.2dB to +4.3dB, significantly enhancing the discernibility of key geological features.

[0259] Counterfactual Verification: To validate the reliability of the processed results, we conducted counterfactual intervention tests on key geological features. For example, we perturbed the predicted cave area by ±20%. Using reverse mapping, we generated theoretical signals and calculated the perturbation response ratio (R). The results showed that in the predicted major cave area, the R value averaged 2.3 (well above the threshold of 1.0), confirming the high reliability of the processed results for this geological feature. However, in some weak reflection areas, the R value was only 0.7, marking them as uncertain areas requiring drilling verification.

[0260] To support tunnel engineering geology experts in understanding and trusting the processing results, we built a causal interpretation system for the processed radar data.

[0261] Key change areas and change attribution: The system automatically identifies and marks the key change areas in the profile and analyzes the causes of the changes as shown in Table 8:

[0262] Table 8: Attribution statistics of key change areas

[0263]

[0264] Example of causal path analysis: The system generates causal path analysis for important geological feature areas (such as the predicted main cave). Table 9 shows a partial causal path analysis of the cave feature at the 170m position on profile 25.

[0265] Table 9: Key causal path analysis of cave characteristics (order of importance)

[0266]

[0267] Multi-level interpretation generation: For each key change area, the system generates different levels of interpretation information to meet the needs of different users. An example of multi-level interpretation is shown in Table 10:

[0268] Table 10: Example of multi-level interpretation (for cave characteristics at 170m on survey line 25)

[0269]

[0270] Visualization and Interactive Interface: The system provides geologists with a variety of visual interpretation components, including before-and-after comparison charts, causal path visualizations, and uncertainty heat maps, along with support for interactive exploration. These features enable engineering geologists to fully understand the rationale behind treatment decisions and enhance their confidence in the results.

[0271] In the actual application of this tunnel project, we verified the technical effectiveness of this method in many aspects, focusing on the following two key technical effects:

[0272] To verify the geological feature recognition capability of this method in a high-noise environment, we conducted drilling verification on radar profiles before and after processing. The comparison of the accuracy of geological feature recognition under different noise environments is shown in Table 11:

[0273] Table 11: Comparison of geological feature recognition accuracy under different noise environments

[0274]

[0275] During actual tunnel excavation, this method successfully predicted two previously undiscovered karst caves (4.2 x 3.1 m and 2.8 x 2.3 m, respectively). These caves, masked by noise using conventional processing methods, became clearly visible after this method's treatment and were subsequently verified through drilling. This result avoided potential construction safety risks and demonstrates the practical value of this method in complex geological environments.

[0276] To verify the interpretability of this method, we conducted a geological expert evaluation experiment. The evaluation comparison of geological experts on the results of different processing methods is shown in Table 12:

[0277] Table 12: Comparison of geological experts’ evaluation of the results of different treatment methods

[0278]

[0279] The results show that compared with traditional filtering methods and pure deep learning methods, this method not only achieved higher scores in result reliability, but also had significant advantages in interpretation satisfaction and the credibility of decision-making basis. In particular, the willingness of experts to make engineering decisions based on the results of this method reached 89.3%, far higher than other methods.

[0280] Expert feedback indicates that understanding the geophysical rationale behind processing decisions is a key factor in building trust. Causal path explanations enable experts to connect signal processing results to their familiar geological knowledge, significantly improving their acceptance of automated processing results.

[0281] In actual engineering applications, the interpretation system of this method provides a reliable basis for tunnel design and construction plan adjustments, helping the engineering team to make rapid and reasonable decisions after the discovery of caves and faults, avoiding potential engineering risks while optimizing construction progress and costs.

[0282] In at least one embodiment of the present invention, a 3D geological radar AI processing and interpretation platform is disclosed. Based on the technical solution of embodiment 1, while retaining the basic processing flow, a hierarchical collaborative scheduling method with asynchronous decision-making is introduced, including:

[0283] Construct regional geological causal structure maps;

[0284] Input data types: 3D geological radar original data set, geological prior knowledge base data, and regional segmentation parameters;

[0285] Implementation method:

[0286] The 3D geological radar data is divided into regions according to the complexity of geological features;

[0287] Construct a geological causal structure diagram for each region and mark the causal relationship strength between nodes;

[0288] Identify and label causal connections between regions;

[0289] Specific output results: A collection of regionalized geological causal structure maps, including digital representations of the causal relationships within each region and the causal connections between regions.

[0290] Generate region-level causal-driven adaptive filters;

[0291] Input data types: regionalized geological radar raw data, regionalized geological causal structure map collection;

[0292] Implementation method:

[0293] Perform wavelet decomposition on radar data in each region to obtain multi-scale coefficients;

[0294] Determine the causal protection coefficient matrix based on the causal structure diagram of each region;

[0295] Construct a region-adaptive threshold function:

[0296] ;

[0297] in is the basic threshold, is the causal protection coefficient, is the spatial constraint coefficient, and To adjust the parameters;

[0298] Specific output results: regional-level causal-driven adaptive filter parameter set, including the filtering parameters of each region and boundary filtering coordination parameters.

[0299] Construct a regionally adaptive signal-geology bidirectional mapping model;

[0300] Input data types: regional filtered radar dataset, regionalized geological causal structure map collection, regional characteristic description data.

[0301] Implementation method:

[0302] A parameter-adaptive forward mapping model is constructed for each region to map the filtered signal into a geological feature representation;

[0303] Construct a parameter-adaptive reverse mapping model for each region to map geological feature representations into ideal signals;

[0304] Adjust model parameters based on regional characteristics to ensure mapping consistency between regions;

[0305] Specific output results: regional adaptive signal geological bidirectional mapping model set, including forward mapping model parameters and reverse mapping model parameters of each region.

[0306] Perform asynchronous hierarchical double-loop iterative optimization;

[0307] Input data types: regionalized geological radar raw data, regional-level causal-driven adaptive filter parameter set, and regional-adaptive signal-geology bidirectional mapping model set;

[0308] Implementation method:

[0309] Data area division and task value assessment

[0310] Divide the data into multiple processing areas based on the geological characteristics of the 3D radar data ,in 、 、 These are the 1st, 2nd, and nth treatment areas respectively;

[0311] Calculate the task value index for each area:

[0312] ;

[0313] in For the region The task value index, For the region The causal importance index of is the feature significance index, To handle the uncertainty index, 、 、 is the weight coefficient;

[0314] Priority stratification and resource allocation: Divide areas into three levels of priority based on task value:

[0315] ;

[0316] ;

[0317] ;

[0318] in, is a collection of high priority areas, is a set of medium priority areas, is a collection of low-priority areas, and Assign thresholds for priority;

[0319] Allocate computing resources by priority:

[0320] ;

[0321] in 、 、 Respectively represent the amount of computing resources allocated to the high-priority region set, the medium-priority region set, and the low-priority region set;

[0322] Asynchronous iteration and boundary state management: Determine the frequency of iteration updates based on priority:

[0323] ;

[0324] That 、 、 Respectively represent the iterative update frequencies assigned to the high priority region set, the medium priority region set, and the low priority region set;

[0325] Implement status coordination on the borders of adjacent regions:

[0326] ;

[0327] in and Adjacent areas and The processing status, is the boundary state after coordination, is the boundary coordination factor;

[0328] Regional adaptive convergence judgment: Calculate the regional characteristic adaptive convergence threshold:

[0329] ;

[0330] in For the region The convergence threshold of is the benchmark convergence threshold, and To adjust the parameters, is causal importance, is the characteristic significance;

[0331] Termination condition of regional iteration: when the state change between two adjacent iterations is less than the adaptability threshold or the maximum number of iterations is reached, it is terminated;

[0332] Task dependency-aware scheduling: Builds a task dependency graph between regions and determines the execution order based on topological sorting, optimizing parallel execution by considering resource availability and dependencies.

[0333] Specific output results: regionalized geological radar signal dataset after asynchronous optimization processing, including optimized signal data and boundary coordination data of each region.

[0334] Generate multi-level causal explanation reports;

[0335] Input data types: Regionalized geological radar raw data, regionalized geological radar signal dataset after asynchronous optimization processing, regionalized geological causal structure map collection, regional priority allocation data

[0336] Implementation method:

[0337] Generate independent interpretation documents for each treatment area, including descriptions of key geological features and treatment parameters;

[0338] Generate explanations of priority decisions, explaining the basis for allocating priorities to each region;

[0339] Generate multi-scale interpretation views, including region-level processing views, feature-level interpretation views, and signal-level interpretation views;

[0340] Generate processing efficiency and quality analysis reports, comparing the effects of asynchronous and synchronous processing;

[0341] Specific output results: multi-level causal explanation report data package, including regional level explanation documents, priority decision explanation documents, multi-scale explanation view set and processing efficiency analysis report.

[0342] Implement a dynamic scheduler for computing resources;

[0343] Input data types: task priority data, system resource status data, processing node load data, and regional convergence status data;

[0344] Implementation method:

[0345] Collect resource usage indicators of each computing node, including CPU, memory, GPU and other resource usage;

[0346] Calculate load balancing indicators to ensure that the load between nodes is relatively balanced;

[0347] Dynamically allocate resources based on region priority and convergence status:

[0348] ;

[0349] in For the region exist The amount of resources allocated at a given moment, As the basic resource allocation, For the region exist The convergence of time, Adjust parameters for resource allocation;

[0350] Implement adaptive batch processing and dynamic task migration to optimize resource utilization efficiency;

[0351] Collect and process performance indicator data to provide a basis for subsequent resource allocation;

[0352] Specific output results: computing resource dynamic scheduling control data, including resource allocation strategy parameters, load balancing control parameters, batch optimization parameters and task scheduling sequence.

[0353] The above describes an embodiment of the present invention, but this embodiment is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Ordinary technicians in this field can also make more forms of equivalent embodiments based on the inspiration of this embodiment, all of which are protected by this embodiment.

Claims

1. 3D geological radar AI processing and interpretation platform, characterized by: include: A geological causal structure diagram construction module is used to construct a geological causal structure diagram. Based on a constraint-based causal discovery algorithm, a geological causal structure diagram is obtained that characterizes the causal dependency relationship between geological features and radar signals. The construction of the geological causal structure diagram includes: Standardize, detrend and detect outliers on historical geological radar data to obtain preprocessed datasets; Extract physical constraints from electromagnetic wave propagation theory and geological knowledge to form a constraint set; The causal graph skeleton is learned using a constraint-based PC algorithm that determines the conditional independence between nodes through the conditional independence test statistic: ； in is the conditional independence test statistic, For node sets For any two nodes in is the conditional independence test statistic, is the set of physical constraints, is the basic condition independence test value, Respectively and nodes, is the condition set, It is The influence factor of a physical constraint on the independence judgment of the condition, Indicates from arrive The continuous multiplication operation, is the total number of physical constraints; Based on the V-shaped structure, acyclic constraints and physical prior knowledge, the direction of the edges in the graph is determined to obtain a directed acyclic graph; A causal-driven adaptive filtering module constructs a causal-driven adaptive filter and uses a causal-weighted adaptive threshold algorithm to obtain an adaptive filter that can selectively retain geological feature information. The signal-geology bidirectional mapping module constructs a signal-geology bidirectional mapping model, establishes a bidirectional mapping relationship between the signal domain and the geological domain, and realizes the mutual conversion between signal features and geological features. The construction of the signal-geology bidirectional mapping model includes: Construct a forward mapping model to achieve the conversion from signal features to geological features; Construct a reverse mapping model to achieve the conversion from geological features to signal features; Construct a bidirectional consistency measurement function to ensure the self-consistency of the bidirectional mapping; Set a joint optimization objective that integrates forward mapping loss, reverse mapping loss, bidirectional consistency loss, physical consistency loss, and causal consistency regularization term; The bidirectional mapping model is trained through an alternating optimization strategy to achieve mutual conversion between the signal domain and the geological domain; The dual-loop iterative optimization module, based on the dual-loop iterative optimization, uses a causal-driven dual-domain alternating optimization algorithm to perform iterative optimization between the signal domain and the geological domain to obtain the optimized clean geological radar signal; The causal interpretation module receives the geological causal structure diagram and the iterative optimization process record as input, and outputs a causal interpretation report containing textual explanations, visual charts, and interactive exploration components; The feedback learning module takes the functional modules constructed above as input and outputs an integrated system with continuous learning capabilities and a continuously updated causal knowledge base.

2. The 3D geological radar AI processing and interpretation platform according to claim 1, characterized in that: The construction of the causal driven adaptive filter comprises: Wavelet transform is used to decompose the original geological radar signal into coefficients of multiple frequency scales; Establish the mapping relationship between wavelet coefficients and signal feature nodes in the causal structure diagram; Calculate the causal protection coefficient of the signal feature node based on the causal structure diagram: ； in Signal feature node The causal protection coefficient, which represents the importance of signal characteristics to the inference of geological characteristics; is a subset of geological feature nodes; is a subset of signal feature nodes; Slave nodes for computing To Node the causal strength of Indicates that for all geological feature nodes sum; is a set of geological feature nodes, and is the node index, For slave nodes To Node the causal strength of Constructing an adaptive threshold function for fusion causal protection in the wavelet domain: ； in For the Layer wavelet detail coefficients are at positions The threshold value at For the The base threshold of the layer, is the causal protection strength parameter, is the causal protection coefficient of the corresponding signal characteristic node, is a local geological feature significance function, quantifying the location the strength of geological features that may be present at the location; An improved soft threshold function is used to process the wavelet coefficients and reconstruct the filtered signal.

3. The 3D geological radar AI processing and interpretation platform according to claim 1, characterized in that: The double-loop iterative optimization includes: Initial processing of the raw signal using a causal-driven adaptive filter; inferring geological features using forward mapping models; Use the reverse mapping model to generate the corresponding theoretical signal; Adaptively fuse the filtered signal with the theoretical signal; Adjust the parameters of the adaptive filter according to the geological characteristics and signal quality of the current iteration; Re-filter the original signal using the updated filter parameters; Calculate the changes in signals and geological characteristics between two adjacent iterations, and stop the iteration when the convergence condition is met.

4. The 3D geological radar AI processing and interpretation platform according to claim 1, characterized in that: The causal explanation module includes: Compare the original signal and the processed signal to identify areas where significant changes have occurred; Analyze the cause of each key change area to determine whether it is noise removal or signal enhancement; For each key area of ​​change, trace the causal paths that influenced treatment decisions; Mapping the set of interpretation pathways to geophysical laws; Quantitative assessment of uncertainty in interpretation of results; Generate multi-level explanatory information based on user needs; Generate intuitive visual explanation components; Generate causal explanation reports that include textual explanations, visual charts, and interactive exploration capabilities.

5. The 3D geological radar AI processing and interpretation platform according to claim 1, characterized in that: The feedback learning module includes: Integrate various functional modules into a unified processing platform; Build an expert feedback collection component to obtain evaluations of processing results and correction suggestions; Structuring and categorizing expert feedback; Update the causal structure diagram based on expert feedback; Optimize algorithm parameters of each module based on expert parameter feedback; Improve interpretation criteria based on interpretation feedback; Use incremental learning to continuously optimize the system; Implement continuous monitoring and optimization mechanism of system performance.

6. The 3D geological radar AI processing and interpretation platform according to claim 4, characterized in that: The causal paths that trace back to influence processing decisions include: Identify corresponding nodes in the signal feature space; Identify relevant nodes in geological feature space; Find all paths from signal feature nodes to geological feature nodes in the causal graph; Calculate the importance score for each path: ； in represents the importance score of each path, For the path, is the edge on the path, is the corresponding causal strength value; Several paths with the highest scores are selected as the explanation path set.

7. The 3D geological radar AI processing and interpretation platform according to claim 5, characterized in that: The updating of the causal structure diagram based on expert feedback includes: For causal feedback, update the weight of the causal edge: ； in and are the causal edge weights before and after the update, is the smoothing factor, For experts The recommended weight value of Add, remove, or modify causal edges to align with expert knowledge: ； in represents the updated causal edge set, represents the causal edge set before the update, is the edge set to be added, is the set of edges that need to be removed.

8. A computer-readable storage medium, characterized in that It is used to store computer-readable instructions, which, when read by a computer, can run the three-dimensional geological radar AI processing and interpretation platform as described in any one of claims 1 to 7.

Images