Coal-based rare metal and buried structure cooperative detection method and device, electronic equipment and storage medium

Abstract

The embodiment of the application provides a coal-based rare metal and hidden structure cooperative detection method and device, electronic equipment and storage medium, relates to the technical field of data processing, and the method comprises the following steps: collecting original seismic tail wave scattering signals and rock mass pressure data of a target area; calculating a rock mass damage factor according to the rock mass pressure data, and performing anti-interference processing and correction on the original seismic tail wave scattering signals by using the rock mass damage factor to obtain effective scattering characteristic signals; extracting core scattering characteristics of the effective scattering characteristic signals, and combining the rock mass damage factor to construct a joint feature set; inputting the joint feature set into a pre-trained cooperative identification model to output distribution information of a rare metal enrichment area and a hidden structure. Through the method provided in the embodiment of the application, the identification accuracy of synchronous detection of the rare metal enrichment area and the hidden structure is significantly improved, and accurate cooperative detection of the two is realized.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, electronic device and storage medium for the collaborative detection of coal-bearing rare metals and concealed structures. Background Technology

[0002] Rare metals in coal-bearing strata are strategic mineral resources. Their enrichment and distribution are closely related to the formation of hidden geological structures (such as fault fracture zones and mineralized alteration zones). Therefore, achieving coordinated exploration of rare metals and investigation of hidden structures is of great significance for improving the efficiency of comprehensive resource utilization and ensuring safe production in deep mines.

[0003] Existing detection technologies mostly employ seismic coma scattering methods, but these typically separate device design from the detection method. The devices cannot dynamically adjust parameters based on real-time data, and the methods are not optimized for device characteristics, leading to a mismatch between device performance and method requirements. Furthermore, existing technologies are primarily designed for shallow, relatively intact strata. In deep, fragmented coal-bearing strata, they face challenges such as borehole collapse, significant excitation energy loss, and severe signal attenuation. They also simply correlate coma scattering characteristics with single targets like rare metals or concealed structures, failing to consider the interference of deep rock mass damage on the scattering signal. They lack quantification and correction of rock mass damage factors, making it difficult to effectively distinguish between scattering signals caused by minor rock mass damage and rare metal enrichment. This results in low accuracy and fails to meet the precision requirements for collaborative detection of two targets under complex deep geological conditions. Therefore, a solution to address these issues is urgently needed. Summary of the Invention

[0004] This application provides a method, apparatus, electronic device, and storage medium for the coordinated detection of coal-bearing rare metals and concealed structures, in order to address the deficiencies in the prior art.

[0005] This application provides a method for the coordinated detection of coal-bearing rare metals and concealed structures, applied to a coordinated detection device. The method includes: Acquire raw seismic coma scattering signals and rock pressure data of the target area; The rock mass damage factor is calculated based on the rock mass pressure data, and the rock mass damage factor is used to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The core scattering features of the effective scattering feature signal are extracted and combined with the rock mass damage factor to construct a joint feature set. The joint feature set is then input into a pre-trained collaborative identification model to output the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factor and map them to the degree of mineralization.

[0006] According to an embodiment of this application, a method for the coordinated detection of coal-bearing rare metals and concealed structures is provided, wherein the acquisition of raw seismic wake scattering signals and rock pressure data of the target area includes: Real-time acquisition of rock mass pressure data and borehole wall displacement data at the target depth in the target area; Each excitation point is excited sequentially according to a preset order, and the original seismic coma scattered signal is collected synchronously. During the collection process, the signal-to-noise ratio is monitored in real time. If the signal-to-noise ratio is lower than a preset threshold, a re-excitation operation is automatically triggered.

[0007] According to an embodiment of this application, a method for the coordinated detection of coal-bearing rare metals and concealed structures is provided. The method involves using the rock mass damage factor to perform anti-interference processing and correction on the original seismic wake scattering signal to obtain an effective scattering characteristic signal, including: The original seismic coma scattered signal is subjected to curvelet transform and empirical mode decomposition to obtain the first intermediate signal; The first intermediate signal is corrected according to the pre-established deep rock mass attenuation compensation algorithm to obtain the second intermediate signal; wherein, the deep rock mass attenuation compensation algorithm is used to dynamically calculate the attenuation compensation coefficient based on the rock mass pressure data and the signal propagation path; The second intermediate signal is corrected using the rock mass damage factor to obtain the effective scattering characteristic signal.

[0008] According to the embodiment of this application, a method for the coordinated detection of coal-bearing rare metals and concealed structures, after obtaining the effective scattering characteristic signal, the method further includes: The effective scattering feature signal is corrected in multiple dimensions to obtain the corrected effective scattering feature signal; The multi-dimensional correction includes terrain correction, rock layer shielding correction, and deep scattering path correction. The deep scattering path correction adopts an improved dynamic time warping algorithm and combines deep rock mass physical parameters to improve path matching accuracy. The improved dynamic time warping algorithm introduces deep rock mass physical parameters as path constraints into the traditional dynamic time warping algorithm to achieve scattering path matching.

[0009] According to an embodiment of this application, a method for the coordinated detection of coal-bearing rare metals and concealed structures is provided, wherein the extraction of the core scattering features of the effective scattering feature signal and the construction of a joint feature set in conjunction with the rock mass damage factor include: Extract core scattering features from the effective scattering feature signal; wherein, the core scattering features include energy decay rate, frequency drift rate, and scattering angle distribution; The core scattering features are combined with the rock mass damage factor to construct an initial joint feature set; By using principal component analysis and wavelet packet decomposition algorithms, redundant features in the initial joint feature set are removed, and features whose correlation with rare metal enrichment areas and hidden structures reaches a preset threshold are retained, thus obtaining the joint feature set.

[0010] According to an embodiment of this application, a method for the collaborative detection of coal-bearing rare metals and concealed structures is provided. The method involves inputting the joint feature set into a pre-trained collaborative identification model and outputting distribution information of rare metal enrichment areas and concealed structures, including: The joint feature set is input into the collaborative identification model, which simultaneously outputs the three-dimensional coordinates, grade, mineralization range of the rare metal enrichment area, as well as the type, scale, burial depth, development degree and rock mass damage level of the concealed structure. The grade is divided into low grade, medium grade and high grade according to the mass fraction of rare metals.

[0011] According to the embodiments of this application, a method for the coordinated detection of coal-bearing rare metals and concealed structures, after outputting the distribution information of rare metal enrichment areas and concealed structures, the method further includes: Obtain measured data; wherein, the measured data includes rare metal grade, structural parameters, and rock mass damage degree; The model parameters and identification thresholds of the collaborative identification model are automatically corrected based on the measured data to form a unique identification model adapted to the target area.

[0012] This application embodiment also provides a device for the coordinated detection of coal-bearing rare metals and concealed structures, the device comprising: The deep adaptive excitation-receiver module is used to acquire raw seismic coma scattering signals and rock pressure data in the target area; A multi-dimensional anti-interference signal acquisition and transmission module is used to calculate the rock mass damage factor based on the rock mass pressure data, and to use the rock mass damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The three-dimensional collaborative identification processing module is used to extract the core scattering features of the effective scattering feature signal, and construct a joint feature set by combining it with the rock mass damage factor. The joint feature set is then input into a pre-trained collaborative identification model, which outputs the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factors and map them to the degree of mineralization. The dynamic linkage control module is connected to the deep adaptive excitation-reception module, the multi-dimensional anti-interference signal acquisition and transmission module, and the three-dimensional collaborative identification and processing module, and includes an industrial-grade programmable logic controller, a visual control terminal, and a data interaction unit.

[0013] According to an embodiment of this application, a coal-bearing rare metal and concealed structure collaborative detection device is provided, wherein the deep adaptive excitation-receiving module includes: The adaptive excitation unit adopts an integrated design of a controllable vibration source and a drilling excitation head. The diameter of the excitation head is adjustable, and the end of the excitation head is equipped with a wear-resistant and anti-slip probe and adopts a hydraulic rigid coupling structure to dynamically adjust the coupling force according to the pressure of the deep rock mass. The high-precision receiving unit consists of an array of high-sensitivity detectors, a collapse-proof protective sleeve, and a signal amplification module. The sampling frequency of the detectors can be dynamically adjusted. The deep positioning and rock pressure monitoring unit integrates a positioning module and a miniature geostress sensor to collect the three-dimensional coordinates of the excitation and receiving points as well as deep rock pressure data in real time.

[0014] According to an embodiment of this application, a coal-bearing rare metal and concealed structure collaborative detection device is provided, wherein the multi-dimensional anti-interference signal acquisition and transmission module includes: The signal acquisition unit uses a multi-channel synchronous acquisition chip to synchronously acquire wake wave signals, rock pressure signals, and borehole wall displacement signals. The composite anti-interference preprocessing unit has a built-in composite algorithm chip for curve transform, empirical mode decomposition and deep rock mass attenuation compensation, which is used to remove noise and correct deep rock mass attenuation in the acquired signal. The dual-mode redundant transmission unit adopts three transmission modes: wired transmission, wireless transmission, and fiber optic transmission, and has a built-in high-speed cache unit to cache data when the network connection is lost.

[0015] According to an embodiment of this application, a coal-bearing rare metal and concealed structure collaborative detection device is provided, wherein the three-dimensional collaborative identification and processing module includes: The dynamic feature extraction unit has built-in principal component analysis and wavelet packet decomposition algorithm modules to extract the core features of wake scattering, and calculates the rock mass damage factor by combining rock mass pressure data to construct a joint feature set; The three-dimensional collaborative identification unit integrates a Bayesian-optimized convolutional neural network and a long short-term memory network collaborative identification algorithm chip, which is used to combine the correlation model between scattering features, rock mass damage factors and mineralization degree to simultaneously complete the collaborative identification of rare metals and concealed structures. The dynamic model correction unit is used to receive measured data from the verification borehole and automatically correct the correlation model and identification threshold.

[0016] According to an embodiment of this application, a coal-bearing rare metal and concealed structure collaborative detection device is provided, wherein the dynamic linkage control module includes: The module linkage control unit adopts an industrial-grade programmable logic controller to automatically adjust the excitation energy, acquisition interval and anti-interference parameters based on the identification results, rock pressure data and borehole wall displacement data. A visual control terminal is used to display in real time the wake scattering characteristic map, the distribution map of rare metal enrichment area, the distribution map of concealed structure, the distribution map of rock mass damage, and the working status of each module of the device. The data interaction unit is used to network and interact with the mine dispatch center and drilling equipment to realize the real-time uploading of detection data and identification results, as well as the reception of control commands.

[0017] According to an embodiment of this application, a collaborative detection device for coal-bearing rare metals and concealed structures is provided. The device further includes an auxiliary adaptation and expansion module, which includes: A deep-fitting component includes a retractable drill reinforcement sleeve and a high-pressure sealing joint. The drill reinforcement sleeve can adaptively expand and contract according to the drill diameter, and the high-pressure sealing joint is used to withstand the high-pressure environment in the deep environment. Modular expansion interface, located on the device casing, allows for flexible addition or removal of functional modules; The fault self-diagnosis unit has a built-in fault detection chip, which is used to monitor the working status of each module in real time, and automatically locate the fault location and output fault handling suggestions when a fault occurs.

[0018] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the collaborative detection method for coal-bearing rare metals and concealed structures as described above.

[0019] This application also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for coordinated detection of coal-bearing rare metals and concealed structures as described above.

[0020] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the collaborative detection method for coal-bearing rare metals and concealed structures as described above.

[0021] This application provides a method, device, electronic device, and storage medium for the collaborative detection of coal-bearing rare metals and concealed structures. The method involves acquiring raw seismic coma scattering signals and rock mass pressure data of a target area; calculating a rock mass damage factor based on the rock mass pressure data; and using the rock mass damage factor to perform anti-interference processing and correction on the raw seismic coma scattering signals to obtain effective scattering characteristic signals; extracting the core scattering features of the effective scattering characteristic signals; constructing a joint feature set by combining the rock mass damage factor; and inputting the joint feature set into a pre-trained collaborative identification model to output the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering characteristics with the rock mass damage factor and map them to the degree of mineralization. Therefore, this application embodiment can effectively remove the interference of deep rock damage on the scattering signal by introducing rock damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal, and obtain an effective scattering feature signal that better reflects the essence of the target. On this basis, the extracted core scattering features and rock damage factor are used to construct a joint feature set, which is then input into a pre-trained collaborative identification model. This achieves deep coupling and collaborative identification of scattering features and rock damage factor, and solves the technical problem of inaccurate correlation of scattering features and inability to distinguish between rock damage and rare metal mineralization signals caused by ignoring the influence of rock damage in the existing technology. This significantly improves the identification accuracy of synchronous detection of rare metal enrichment areas and concealed structures, and achieves accurate collaborative detection of the two. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating the method for the coordinated detection of coal-bearing rare metals and concealed structures provided in the embodiments of this application.

[0024] Figure 2 This is a complete flowchart of the method for the coordinated detection of coal-bearing rare metals and concealed structures provided in the embodiments of this application.

[0025] Figure 3 This is a schematic diagram of the structure of the coal-bearing rare metals and concealed structures collaborative detection device provided in the embodiments of this application.

[0026] Figure 4 This is a schematic diagram of the linkage logic of each module of the coal-bearing rare metals and concealed structures collaborative detection device provided in the embodiments of this application.

[0027] Figure 5 This is a structural cross-sectional view of the deep adaptive excitation-reception module provided in the embodiments of this application.

[0028] Figure 6 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0029] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the protection scope of the embodiments of this application.

[0030] The following describes, with reference to the accompanying drawings, an embodiment of the present application of a method, apparatus, electronic device, and storage medium for the coordinated detection of coal-bearing rare metals and concealed structures.

[0031] It should be noted that rare metals in coal-bearing strata, as strategic mineral resources, are closely related to the formation of concealed geological structures (such as fault fracture zones and mineralized alteration zones). However, these concealed structures are also core disaster hazards in deep mining. Therefore, achieving coordinated development of "rare metal exploration and concealed structure investigation" is crucial to improving the comprehensive utilization efficiency of coal-bearing resources and ensuring safe production in deep mines. Related detection technologies and equipment still have many limitations, specifically as follows: 1. Insufficient coordination between devices and methods: In related technologies, seismic coma detection devices and detection methods are mostly designed independently. The devices cannot dynamically adjust their working parameters based on real-time data during the detection process, and the methods have not optimized their processes for the hardware characteristics of the devices. This results in "the device performance cannot be fully utilized and the method requirements cannot be accurately implemented," making it difficult to achieve efficient linkage for dual-target collaborative detection. Especially in deep, soft, and fractured strata, detection failure is easily caused by parameter mismatch.

[0032] 2. Poor adaptability to deep, fractured, and soft strata: Wake wave detection devices are mostly designed for shallow, relatively intact coal-bearing strata and employ a rigid excitation-receiver structure. In deep, fractured, and soft strata, boreholes are prone to collapse and rock mass disturbance is intense, resulting in severe excitation energy loss and weak and distorted received signals. The detection method does not consider the influence of deep rock mass pressure and temperature on wake wave scattering characteristics and does not have a dedicated signal attenuation compensation mechanism, resulting in an extremely low signal-to-noise ratio and a significant decrease in detection accuracy.

[0033] 3. Inaccurate correlation and weak identification capability of wake scattering features: Related technologies only simply correlate wake scattering features with concealed structures or rare metals as a single target, without establishing a three-dimensional correlation model of "wake scattering features - rock mass damage degree - rare metal mineralization degree". It is unable to distinguish the difference in scattering signals caused by rare metal enrichment and slight rock mass damage, which is prone to misjudgment; the identification algorithms are mostly static algorithms, which cannot dynamically adjust the identification threshold according to real-time detection data, and have weak ability to adapt to the geological differences of different mining areas.

[0034] 4. Low modularity and poor scalability of the device: The functional modules of the detection device have low integration and non-universal interfaces, making it impossible to flexibly add or remove functional modules according to detection needs (such as adding rare earth detection units or deep rock pressure monitoring units); moreover, it lacks a unified control center, with each module working independently, making it impossible to achieve real-time data sharing and coordinated control, and difficult to adapt to the collaborative detection needs of different types of rare metals and different concealed structures. Based on this, embodiments of this application provide a method for collaborative detection of coal-bearing rare metals and concealed structures to solve at least one of the above problems.

[0035] Figure 1 This is a flowchart illustrating the method for the coordinated detection of coal-bearing rare metals and concealed structures provided in this application embodiment. Figure 1 As shown, this method is applied to a cooperative detection device, including the following: Step 100: Collect raw seismic coma scattering signals and rock pressure data of the target area.

[0036] Specifically, the original seismic coma scattering signal refers to the coma signal that is scattered and continues to propagate when the seismic wave encounters discontinuities in the medium, such as lithological interfaces, differences in mineral composition, structural fracture zones, or rare metal enrichment masses, as it propagates in coal-bearing strata. Its energy attenuation rate, frequency drift rate, and scattering angle distribution can comprehensively reflect the changes in the physical properties and structural development of the strata. Rock mass pressure data is the in-situ stress value of deep rock mass collected in real time by a miniature in-situ stress sensor integrated into the collaborative detection device, which is used to characterize the stress state of the rock mass under complex stress environment.

[0037] Step 200: Calculate the rock mass damage factor based on the rock mass pressure data, and use the rock mass damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal.

[0038] Specifically, the rock mass damage factor is used to characterize the degree of damage to deep rock masses under in-situ stress. For example, it can be quantified by the ratio of rock mass pressure to rock mass compressive strength or by the rate of change of rock mass wave velocity. Its purpose is to eliminate the interference of rock mass damage on the scattered signal. Subsequently, the rock mass damage factor is used to perform anti-interference processing and correction on the original seismic coma scattering signal. The anti-interference processing refers to removing random noise, high-frequency interference and mud interference in the signal through algorithms such as curve transform and empirical mode decomposition. The correction refers to correcting the energy loss and waveform distortion caused by rock mass absorption, scattering and inelastic attenuation during the deep propagation of the signal through a pre-established deep rock mass attenuation compensation algorithm. Finally, an effective scattering characteristic signal is obtained. After correction by the rock mass damage factor, the interference of rock mass damage and propagation attenuation is eliminated, and it can more realistically reflect the inherent properties of rare metal mineralization and concealed structures.

[0039] Step 300: Extract the core scattering features of the effective scattering feature signal, and construct a joint feature set by combining it with the rock mass damage factor. Input the joint feature set into the pre-trained collaborative identification model, and output the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factor and map them to the degree of mineralization.

[0040] Specifically, core scattering characteristics include, but are not limited to, energy decay rate, frequency drift rate, and scattering angle distribution. The energy decay rate reflects how quickly the signal loses energy as it propagates in the medium, the frequency drift rate reflects the trend of signal frequency change with propagation distance, and the scattering angle distribution reflects the geometric shape and spatial distribution characteristics of the scatterer.

[0041] These core scattering features are combined with the aforementioned rock mass damage factors to construct a joint feature set containing multi-dimensional information. This joint feature set integrates scattering features reflecting mineralization and structure with damage factors reflecting rock mass damage, enabling a more comprehensive characterization of the underground medium's properties. The constructed joint feature set is then input into a pre-trained collaborative identification model. This collaborative identification model is a dynamic model pre-trained with a large amount of sample data, used to correlate scattering features with rock mass damage factors and map them to the degree of mineralization. Specifically, it can be constructed using a Bayesian-optimized convolutional neural network and a long short-term memory network. The convolutional neural network is used to extract spatial features from the joint feature set, the long short-term memory network is used to capture temporal evolution patterns, and Bayesian optimization is used to automatically optimize the model's hyperparameters, enabling the model to simultaneously process spatial and temporal features. The model establishes a nonlinear mapping relationship between scattering characteristics, rock mass damage factors, rare metal mineralization degree, and concealed structure development degree. The final model synchronously outputs the distribution information of rare metal enrichment areas and concealed structures, including the three-dimensional coordinates, grade, and mineralization range of rare metal enrichment areas, as well as the type, scale, burial depth, development degree, and rock mass damage grade of concealed structures. The grade can be divided into low grade, medium grade, and high grade according to the mass fraction of rare metals. The types of concealed structures can include deep micro-faults, collapse columns, goaf boundaries, or fault water-conducting channels.

[0042] The above describes the steps of the collaborative detection method for coal-bearing rare metals and concealed structures provided in the embodiments of this application. As can be seen from the above description, the collaborative detection method for coal-bearing rare metals and concealed structures provided in the embodiments of this application involves: collecting original seismic coma scattering signals and rock mass pressure data of the target area; calculating a rock mass damage factor based on the rock mass pressure data; using the rock mass damage factor to perform anti-interference processing and correction on the original seismic coma scattering signals to obtain effective scattering characteristic signals; extracting the core scattering features of the effective scattering characteristic signals; constructing a joint feature set by combining the rock mass damage factor; inputting the joint feature set into a pre-trained collaborative identification model; and outputting the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factors and map them to the degree of mineralization. Therefore, this application embodiment can effectively remove the interference of deep rock damage on the scattering signal by introducing rock damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal, and obtain an effective scattering feature signal that better reflects the essence of the target. On this basis, the extracted core scattering features and rock damage factor are used to construct a joint feature set, which is then input into a pre-trained collaborative identification model. This achieves deep coupling and collaborative identification of scattering features and rock damage factor, and solves the technical problem of inaccurate correlation of scattering features and inability to distinguish between rock damage and rare metal mineralization signals caused by ignoring the influence of rock damage in the existing technology. This significantly improves the identification accuracy of synchronous detection of rare metal enrichment areas and concealed structures, and achieves accurate collaborative detection of the two.

[0043] Based on the above embodiments, in this embodiment, step 100, acquiring the original seismic coma scattering signal and rock pressure data of the target area, includes: Step 110: Collect rock pressure data and borehole wall displacement data at the target depth in the target area in real time.

[0044] Step 120: Excite each excitation point in a preset order, and simultaneously collect the original seismic coma scattering signal. During the collection process, monitor the signal-to-noise ratio in real time. If the signal-to-noise ratio is lower than a preset threshold, automatically trigger the re-excitation operation.

[0045] It should be noted that before collecting data, the collaborative detection device must first be modularly assembled and the detection parameters initialized according to the detection target.

[0046] Specifically, based on the type of rare metal (such as gallium, niobium, rare earth elements, or lithium) and the type of concealed structure (such as deep micro-faults, collapse columns, goaf boundaries, or fault water-conducting channels) required to be detected in the target area, corresponding functional modules can be flexibly added or removed through standardized expansion interfaces on the collaborative detection device. For example, when it is necessary to detect rare earth elements, a rare earth detection unit can be added; when it is necessary to detect goaf boundaries, a goaf boundary detection unit can be added, thereby achieving rapid adaptation to different detection scenarios. After completing modular assembly, combined with deep geological data of the target area (including stratum depth, rock mass...), The detection parameters (such as cementation and geostress range) are initialized on the dynamic linkage control terminal. Specifically, this includes setting the excitation head diameter based on the borehole depth and diameter, setting the initial excitation energy based on the burial depth, setting the geophone sampling frequency and acquisition interval based on the stratum characteristics, setting the signal-to-noise ratio threshold based on the on-site noise level, and setting the initial threshold of the rock mass damage factor based on the rock mechanics parameters. At the same time, 3 to 5 calibration points (including shallow and deep calibration points) are set up around the target detection area, and auxiliary positioning and monitoring units are installed to complete the device calibration, ensuring the accuracy and reliability of subsequent data acquisition.

[0047] Based on this, step 110 involves real-time acquisition of rock mass pressure data and borehole wall displacement data at the target depth in the target area. The rock mass pressure data is the in-situ stress value of the deep rock mass acquired in real-time by a miniature in-situ stress sensor integrated into the deep adaptive excitation-receive module, used for subsequent calculation of the rock mass damage factor. The borehole wall displacement data is the borehole wall deformation monitored in real-time by a borehole wall displacement sensor built into the anti-collapse protective sleeve, used to assess borehole stability and provide early warning of collapse risks. Step 120 involves sequentially exciting each excitation point according to a preset order, simultaneously acquiring the original seismic coma scattering signal. The excitation points can be arranged using a rhomboid grid, rectangular grid, or linear array, with the spacing between adjacent excitation points set to 10 to 20 meters according to the required detection accuracy. Each excitation point is repeatedly excited 2 to 3 times to ensure accuracy. To ensure signal stability, synchronous acquisition refers to the simultaneous reception of wake scattering signals from different directions and distances using an array of high-sensitivity detectors (typically 8 channels or more). During acquisition, the signal-to-noise ratio (SNR) is monitored in real time by an abnormal signal identification module. This SNR is the ratio of effective signal energy to noise energy. When the SNR is detected to be lower than a preset threshold (e.g., 30 dB), it indicates that the quality of the currently acquired signal is insufficient to support subsequent processing and identification. The system automatically triggers a re-excitation operation, including adjusting the excitation energy, increasing the number of superpositions, or adjusting the detector coupling state, until a signal that meets the quality requirements is acquired. This ensures that the raw data acquired in the complex environment of deep, fractured strata has sufficient SNR and reliability, laying the foundation for subsequent anti-interference processing and feature extraction.

[0048] The method for collaborative detection of coal-bearing rare metals and concealed structures provided in this embodiment effectively solves the problem of uncontrollable data quality caused by weak signals and strong interference in deep, fractured strata by monitoring the signal-to-noise ratio in real time during the acquisition process and automatically triggering a re-excitation operation when it falls below a preset threshold. This ensures that the acquired original seismic coma scattering signal has sufficient signal-to-noise ratio and reliability, providing a high-quality data foundation for subsequent high-precision anti-interference processing and collaborative identification.

[0049] Based on the above embodiments, in this embodiment, step 200 utilizes the rock mass damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal, including: Step 210: Perform curvelet transform and empirical mode decomposition on the original seismic coma scattered signal to obtain the first intermediate signal.

[0050] Step 220: Correct the first intermediate signal according to the pre-established deep rock mass attenuation compensation algorithm to obtain the second intermediate signal; wherein, the deep rock mass attenuation compensation algorithm is used to dynamically calculate the attenuation compensation coefficient based on the rock mass pressure data and the signal propagation path.

[0051] Step 230: Correct the second intermediate signal using the rock mass damage factor to obtain the effective scattering characteristic signal.

[0052] Specifically, curvelet transform is a multi-scale geometric analysis method that can decompose a signal into curvelet coefficients of different directions and scales, effectively separating directional effective signals from random noise. Empirical mode decomposition (EMD) is an adaptive signal decomposition method that can decompose nonlinear and non-stationary tailwave signals into several intrinsic mode function components. By eliminating high-frequency noise components and low-frequency trend components, it effectively suppresses random noise, high-frequency interference, and mud interference. The combination of the two can significantly improve the signal-to-noise ratio. Based on this, step 220 corrects the first intermediate signal according to the pre-established deep rock mass attenuation compensation algorithm to obtain the second intermediate signal. This deep rock mass attenuation compensation algorithm is a mathematical model based on the viscoelastic constitutive relationship of deep rock mass. It is used to dynamically calculate the attenuation compensation coefficient based on the real-time collected rock mass pressure data and signal propagation path, thereby providing reverse compensation for the energy loss and waveform distortion caused by rock mass absorption, scattering, and inelastic attenuation during signal propagation in the deep region, restoring the original amplitude and phase characteristics of the signal. Finally, step 230 uses the rock mass damage factor calculated above to correct the second intermediate signal to obtain an effective scattering feature signal. The rock mass damage factor is a quantitative index used to characterize the degree of damage to deep rock masses under complex geostress. For example, it can be quantified by the ratio of rock mass pressure to rock mass compressive strength or by the rock mass wave velocity change rate. Using this factor to correct the signal can effectively remove the scattering signal distortion caused by rock mass damage and retain the inherent scattering features that are only related to rare metal enrichment and concealed structures. Thus, an effective scattering feature signal is obtained by eliminating the triple interference of noise, propagation attenuation and rock mass damage, providing high-quality input data for subsequent feature extraction and collaborative identification.

[0053] Furthermore, after obtaining the effective scattering characteristic signal in step 230, the method further includes: The effective scattering feature signal is corrected in multiple dimensions to obtain the corrected effective scattering feature signal; The multi-dimensional correction includes terrain correction, rock layer shielding correction, and deep scattering path correction. The deep scattering path correction adopts an improved dynamic time warping algorithm and combines deep rock mass physical parameters to improve path matching accuracy. The improved dynamic time warping algorithm introduces deep rock mass physical parameters as path constraints into the traditional dynamic time warping algorithm to achieve scattering path matching.

[0054] Specifically, topographic correction eliminates the impact of surface elevation variations on seismic wave propagation time and amplitude by normalizing the elevation data of the acquisition points to a unified reference surface, correcting time differences and energy variations caused by topographic changes. Rock strata shielding correction eliminates the shielding and absorption effect of overlying rock layers on seismic wave energy. Based on the thickness, density, and wave velocity of the rock strata, it calculates the energy attenuation coefficient and performs reverse compensation to restore the target signal weakened by the shielding layer. Deep scattering path correction addresses the problem of complex propagation paths and difficulty in accurately locating scattering points of seismic coma in deep, fractured strata. It employs an improved dynamic time warping algorithm combined with deep rock mass physical parameters to improve... To improve path matching accuracy, the dynamic time warping algorithm, originally used to measure the similarity between two time series and find the optimal matching path through nonlinear alignment, is improved in this embodiment by introducing deep rock mass physical parameters as path constraints. Specifically, a weighted term for these parameters is added when calculating the path matching cost function, allowing the algorithm to consider not only the temporal similarity of the waveforms but also the physical constraints of deep rock mass on the seismic wave propagation path. This improves path matching accuracy to within 0.2 meters, effectively solving the positioning error problem caused by wave velocity variations and multipath scattering in deep, complex strata. Through the multi-dimensional comprehensive processing of terrain correction, rock layer shielding correction, and deep scattering path correction, the final corrected effective scattering feature signal eliminates multiple interferences from the surface, strata, and deep propagation paths, providing a higher-precision and higher-fidelity input signal for subsequent feature extraction and collaborative identification.

[0055] The method for collaborative detection of coal-bearing rare metals and concealed structures provided in this embodiment achieves step-by-step stripping and precise compensation of multi-source interference in deep, fractured strata, significantly improving the signal-to-noise ratio and fidelity of effective scattering characteristic signals. Through deep scattering path correction, signal distortion and positioning errors are systematically eliminated from the three dimensions of surface, strata and deep propagation paths, providing high-precision and high-fidelity input signals for subsequent feature extraction and collaborative identification.

[0056] Based on the above embodiments, in this embodiment, step 300 extracts the core scattering features of the effective scattering feature signal and constructs a joint feature set in conjunction with the rock mass damage factor, including: Step 310: Extract core scattering features from the effective scattering feature signal; wherein the core scattering features include energy decay rate, frequency drift rate, and scattering angle distribution.

[0057] Step 320: Combine the core scattering features with the rock mass damage factor to construct an initial joint feature set.

[0058] Step 330: Using principal component analysis and wavelet packet decomposition algorithms, redundant features in the initial joint feature set are removed, and features with a correlation degree with rare metal enrichment areas and hidden structures that reaches a preset threshold are retained to obtain the joint feature set.

[0059] Specifically, step 310 extracts core scattering features from the effective scattering feature signals. Among them, the energy decay rate refers to the rate at which the amplitude or energy of the seismic coma decays per unit propagation distance or unit time, reflecting the absorption characteristics and scattering intensity of the medium. In rare metal-rich areas, due to the difference in physical parameters caused by changes in mineral composition, the energy decay rate usually shows an abnormally high value. In the fractured zone of concealed structures, due to the increase in porosity caused by rock fracture, the energy decay rate also shows a significant change. The frequency drift rate refers to the rate at which the dominant frequency of the seismic wave shifts with distance or time during propagation, reflecting the inelastic properties of the medium. Rare metal mineralization often has high density and elastic modulus, which can lead to abrupt changes in the frequency drift characteristics. The scattering angle distribution refers to the distribution characteristics of the coma scattered energy at different angles, reflecting the geometric shape and spatial distribution of the scatterer. Concealed structures such as faults and collapse columns usually show anisotropic scattering angle distribution characteristics, while rare metal-rich areas may show isotropic or scattering modes in specific directions.

[0060] Step 320 combines the extracted core scattering features, such as energy decay rate, frequency drift rate, and scattering angle distribution, with the rock mass damage factor calculated above to construct an initial joint feature set containing multi-dimensional information. This joint feature set integrates scattering features reflecting mineralization and structure with damage factors reflecting rock mass damage, and can more comprehensively characterize the comprehensive properties of the underground medium.

[0061] Step 330 uses principal component analysis (PCA) and wavelet packet decomposition (WPD) algorithms to remove redundant features from the initial joint feature set, retaining features whose correlation with rare metal enrichment areas and hidden structures reaches a preset threshold, thus obtaining the final joint feature set. Through the joint dimensionality reduction and filtering of PCA and WPD, the final joint feature set retains core information highly correlated with rare metal enrichment areas and hidden structures while eliminating redundant noise features, significantly reducing the feature dimensionality. This provides a high-quality, low-redundancy input feature set for the efficient training and accurate prediction of the subsequent collaborative identification model.

[0062] The coal-series rare metals and concealed structures collaborative detection method provided in this embodiment effectively improves the characterization ability and signal-to-noise ratio of input data while reducing feature dimensionality. It provides a high-quality, low-redundancy joint feature set for the collaborative identification model, thereby significantly improving the accuracy and computational efficiency of model identification.

[0063] Based on the above embodiments, in this embodiment, step 300 inputs the joint feature set into a pre-trained collaborative identification model and outputs the distribution information of rare metal enrichment areas and hidden structures, including: Step 340: Input the joint feature set into the collaborative identification model and simultaneously output the three-dimensional coordinates, grade, mineralization range of the rare metal enrichment area, as well as the type, scale, burial depth, development degree and rock mass damage level of the concealed structure. The grade is divided into low grade, medium grade and high grade according to the mass fraction of rare metals.

[0064] Specifically, the collaborative identification model is a dynamic identification model pre-trained with a large amount of sample data. It is used to correlate scattering characteristics with rock mass damage factors and map them to mineralization levels. Specifically, it is constructed using a Bayesian-optimized convolutional neural network and a long short-term memory network. The model simultaneously outputs the distribution information of rare metal enrichment areas and concealed structures. The three-dimensional coordinates of the rare metal enrichment areas refer to the coordinates of the excitation and receiving points obtained through the GPS and BeiDou dual-mode positioning modules in the deep positioning and rock mass pressure monitoring unit. Combined with the positioning information after scattering path correction, the specific spatial location of the rare metal enrichment area is deduced, with an accuracy of ±0.3 meters. The grade is classified according to the mass fraction of rare metals. For example, for gallium, a mass fraction between 0.01% and 0.05% is classified as low grade, between 0.05% and 0.1% as medium grade, and greater than or equal to 0.1% as high grade. The mineralization range refers to the distribution range of the rare metal enrichment area in three-dimensional space, including long... Information on depth, width, and thickness; types of concealed structures include deep micro-faults (length ≥ 30 meters), collapse columns (diameter ≥ 8 meters), goaf boundaries, and fault water-conducting channels, etc. Scale refers to the spatial dimensions of the structure, including fault length, fault displacement, and fracture zone width for faults, and diameter and height for collapse columns. Burial depth refers to the vertical depth of the top of the structure from the surface. Development degree refers to the degree of fragmentation, cementation, and activity of the structure, which can be divided into three levels: weak development, moderate development, and strong development. Rock mass damage level is the degree of rock mass damage classified according to the numerical range of the rock mass damage factor. For example, a rock mass damage factor less than 0.03 is considered slight damage, 0.03 to 0.08 is considered moderate damage, and greater than 0.08 is considered severe damage.

[0065] The coal-bearing rare metals and concealed structures collaborative detection method provided in this embodiment achieves accurate collaborative identification of dual targets and integrated output of multi-dimensional information, providing data support for resource mining and disaster prevention and control, and significantly improving the practicality and decision-making efficiency of comprehensive detection of deep coal-bearing strata.

[0066] Based on the above embodiments, in this embodiment, after outputting the distribution information of rare metal enrichment areas and concealed structures in step 300, the method further includes: Obtain measured data; wherein, the measured data includes rare metal grade, structural parameters, and rock mass damage degree; The model parameters and identification thresholds of the collaborative identification model are automatically corrected based on the measured data to form a unique identification model adapted to the target area.

[0067] Specifically, firstly, 5 to 8 verification boreholes are selected, and core samples are collected. Rare metal grades are detected using inductively coupled plasma mass spectrometry (ICP-MS) and X-ray fluorescence spectrometry (XRF). Simultaneously, borehole imaging technology is used to verify the type, scale, development level of concealed structures, and the degree of rock mass damage, obtaining measured data. Rare metal grade refers to the mass fraction of rare metal elements in the core sample. Structural parameters include geometric parameters such as fault length, fault displacement, fracture zone width, or diameter and height of collapse columns. The degree of rock mass damage can be quantified through core recovery rate, rock quality indicators, or fracture density in borehole imaging. The measured data are then input into a dynamic model correction unit. Based on the difference between the measured data and the model prediction results, this unit automatically corrects the model parameters and identification thresholds of the collaborative identification model (such as the boundary for grade classification and the classification threshold for structural development) using a backpropagation algorithm. This allows the model to better adapt to the specific geological characteristics of the deep, soft strata in the target area, forming a specific identification model adapted to the target area. This specific identification model can be continuously iterated and optimized with the accumulation of subsequent exploration data, possessing self-learning capabilities.

[0068] Based on this, the integrated output unit outputs wake scattering characteristic distribution maps, rare metal enrichment area distribution maps, concealed structure distribution maps, rock mass damage distribution maps, and collaborative detection reports of the detection area. These results can be directly printed, exported, or uploaded to the mine dispatch center, simultaneously displaying the working status and detection parameters of each module of the device, providing intuitive and visualized results for the construction of transparent geology in the mine. Simultaneously, the dynamic linkage control module transmits the identification results (high-risk structural areas, high-grade rare metal areas) to the mine drilling equipment and dispatch center, linking and adjusting the drilling trajectory and drilling parameters. For example, when a high-risk fault or collapse column is detected ahead, the borehole dip or azimuth angle is automatically adjusted to avoid the structural area. When a high-grade rare metal enrichment area is detected, the borehole is precisely guided through the area for core sampling verification, thereby achieving integrated linkage of "detection-identification-drilling-prevention".

[0069] After the exploration is completed, the drilling-while-drilling excitation-receiving module is retrieved using a drilling rig. The various modules of the device are cleaned, calibrated, and maintained to ensure the stable performance of the device in subsequent explorations. All exploration data (raw wake signals, rock pressure data, borehole displacement data, etc.), identification results (distribution information of rare metals and concealed structures), verification data (measured grade, structural parameters, etc.), and device operating parameters (excitation energy, sampling frequency, coupling strength, etc.) are archived and stored to form an exploration database. This database can provide a reference for subsequent exploration work and can also be used for continuous iterative optimization and knowledge transfer of the dedicated identification model, supporting subsequent collaborative exploration work in the mining area and surrounding areas.

[0070] The method for collaborative detection of coal-bearing rare metals and concealed structures provided in this embodiment automatically corrects the model parameters and identification thresholds of the collaborative identification model by acquiring measured data, forming a dedicated identification model adapted to the target area. This achieves self-learning and iterative optimization of the model. At the same time, the identification results are linked to control the drilling equipment to form an integrated closed loop of "detection-identification-drilling-prevention". After the detection is completed, the data is archived to form a detection database, which significantly improves the method's adaptability, engineering applicability and data reuse value.

[0071] Figure 2 This is a complete flowchart of the method for coordinated detection of coal-bearing rare metals and concealed structures provided in the embodiments of this application. The following is a summary of the process. Figure 2 This application provides a complete description of the method for the coordinated detection of coal-bearing rare metals and concealed structures provided in the embodiments of this application.

[0072] like Figure 2 As shown, in the first step of data acquisition, seismic stations deployed in the target area synchronously record the wake scattering signals generated when seismic waves encounter media discontinuities during propagation. These wake scattering signals contain information about the physical properties of the subsurface medium.

[0073] The second step is data preprocessing, which involves denoising the acquired raw wake scattering signals to remove random noise and environmental interference, filtering to retain the signal components of the target frequency band, and performing time calibration to eliminate time differences between different stations caused by clock synchronization errors or propagation path differences, thus providing high signal-to-noise ratio and high time consistency signal data for subsequent analysis.

[0074] Next, in the third step of feature extraction, the wake scattering energy is extracted from the preprocessed signal to reflect the spatial distribution of scattering intensity, the attenuation coefficient is extracted to characterize the absorption characteristics of the medium to the seismic wave energy, and the frequency features are extracted to reveal the variation law of signal frequency with propagation distance. These features together constitute the key indicators for describing the properties of the underground medium.

[0075] In the fourth step of collaborative analysis, the feature data extracted from multiple seismic stations are correlated and analyzed. Through spatial complementarity and mutual verification of data from multiple stations, a scattering feature matrix covering the detection area is constructed. This matrix integrates scattering information from different orientations and distances, and can more comprehensively characterize the heterogeneity of the subsurface medium.

[0076] Finally, in the fifth step of the output, a scattering feature distribution map of the detection area is generated based on the constructed scattering feature matrix. This map visually displays the spatial distribution of wake scattering energy, attenuation coefficient, and frequency characteristics, providing a foundation for geological interpretation and target identification.

[0077] The following describes the coal-series rare metal and concealed structure co-detection device provided in the embodiments of this application. The coal-series rare metal and concealed structure co-detection device described below and the coal-series rare metal and concealed structure co-detection method described above can be referred to in correspondence.

[0078] Figure 3 This is a schematic diagram of the structure of the coal-bearing rare metal and concealed structure collaborative detection device provided in the embodiments of this application, as shown below. Figure 3 As shown in the embodiment of this application, the coal-bearing rare metal and concealed structure co-detection device includes: The deep adaptive excitation-receiver module is used to acquire raw seismic coma scattering signals and rock pressure data in the target area; A multi-dimensional anti-interference signal acquisition and transmission module is used to calculate the rock mass damage factor based on the rock mass pressure data, and to use the rock mass damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The three-dimensional collaborative identification processing module is used to extract the core scattering features of the effective scattering feature signal, and construct a joint feature set by combining it with the rock mass damage factor. The joint feature set is then input into a pre-trained collaborative identification model, which outputs the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factors and map them to the degree of mineralization. The dynamic linkage control module is connected to the deep adaptive excitation-reception module, the multi-dimensional anti-interference signal acquisition and transmission module, and the three-dimensional collaborative identification and processing module, and includes an industrial-grade programmable logic controller, a visual control terminal, and a data interaction unit.

[0079] Figure 4 This is a schematic diagram of the linkage logic of each module of the coal-bearing rare metal and concealed structure collaborative detection device provided in the embodiments of this application, such as... Figure 4As shown, after the device is started, external signals first enter the input module. This module is responsible for receiving raw signals from sensors, external devices, or user commands, and performing preliminary analysis, format conversion, and feature extraction on the signals through the built-in data processing unit, preparing standardized input data for subsequent processing. The processed signal is then transmitted to the processing module, which acts as the core decision-making unit. Based on preset algorithm logic, this module performs in-depth analysis, calculation, and decision-making on the input data, executing action outputs to generate control commands or intermediate results. Simultaneously, it feeds the processing results back to the input module or previous stages, enabling result feedback adjustments and dynamically correcting subsequent processing parameters based on the current output state. The control commands output by the module are further transmitted to the output module, which is responsible for converting the processing results into specific execution actions (such as device driving, command sending, result display, etc.). This module also has a result feedback adjustment function, sending execution status or effect information back to the preceding module, forming a multi-level feedback chain. The diagram also includes another input module and a feedback module. The input module receives adjustment signals from the feedback chain, merges them with the main input signal, and re-enters the processing flow. The feedback module is specifically responsible for collecting the execution results and system operating status of the output module, performing a comprehensive evaluation, generating adjustment commands, and injecting them back into the input module, thus constructing a complete closed-loop control system. Based on this, this application deeply couples the dynamic linkage control module with the deep adaptive excitation-reception module, the multi-dimensional anti-interference signal acquisition and transmission module, and the three-dimensional collaborative identification and processing module, realizing fully automated collaborative control from data acquisition and signal processing to identification results.

[0080] The coal-bearing rare metal and concealed structure collaborative detection device provided in this application collects raw seismic coma scattering signals and rock mass pressure data of the target area; calculates a rock mass damage factor based on the rock mass pressure data, and uses the rock mass damage factor to perform anti-interference processing and correction on the raw seismic coma scattering signals to obtain effective scattering characteristic signals; extracts the core scattering features of the effective scattering characteristic signals, and constructs a joint feature set in combination with the rock mass damage factor; inputs the joint feature set into a pre-trained collaborative identification model, and outputs the distribution information of rare metal enrichment areas and concealed structures; wherein, the collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factors and map them to mineralization degree. Therefore, this application embodiment can effectively remove the interference of deep rock damage on the scattering signal by introducing rock damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal, and obtain an effective scattering feature signal that better reflects the essence of the target. On this basis, the extracted core scattering features and rock damage factor are used to construct a joint feature set, which is then input into a pre-trained collaborative identification model. This achieves deep coupling and collaborative identification of scattering features and rock damage factor, and solves the technical problem of inaccurate correlation of scattering features and inability to distinguish between rock damage and rare metal mineralization signals caused by ignoring the influence of rock damage in the existing technology. This significantly improves the identification accuracy of synchronous detection of rare metal enrichment areas and concealed structures, and achieves accurate collaborative detection of the two.

[0081] Based on the above embodiments, in this embodiment, the deep adaptive excitation-reception module includes: The adaptive excitation unit adopts an integrated design of a controllable vibration source and a drilling excitation head. The diameter of the excitation head is adjustable, and the end of the excitation head is equipped with a wear-resistant and anti-slip probe and adopts a hydraulic rigid coupling structure to dynamically adjust the coupling force according to the pressure of the deep rock mass. The high-precision receiving unit consists of an array of high-sensitivity detectors, a collapse-proof protective sleeve, and a signal amplification module. The sampling frequency of the detectors can be dynamically adjusted. The deep positioning and rock pressure monitoring unit integrates a positioning module and a miniature geostress sensor to collect the three-dimensional coordinates of the excitation and receiving points as well as deep rock pressure data in real time.

[0082] Figure 5 This is a structural cross-sectional view of the deep adaptive excitation-reception module provided in the embodiments of this application, as shown below. Figure 5As shown, the deep adaptive excitation-receiver module adopts an integrated design for drilling, allowing it to be deployed with the drilling rig to the target depth. The module integrates an automatic adaptive adjustment unit that dynamically adjusts the excitation and receiving parameters based on deep rock pressure data, borehole wall displacement data, and changes in formation lithology. Specifically, the automatic adaptive adjustment includes automatically adjusting the coupling strength of the excitation unit based on real-time monitored rock pressure to ensure optimal contact between the excitation head and the borehole wall, minimizing energy loss; automatically adjusting the excitation energy based on borehole depth and formation attenuation characteristics to compensate for deep signal attenuation; and automatically adjusting the geophone position based on borehole wall stability monitoring results to avoid collapse risks. Regarding the excitation function, the module uses a controllable source to excite seismic waves in a preset sequence. The diameter of the excitation head can be adjusted according to the borehole diameter. The end is equipped with a wear-resistant, anti-slip probe and a hydraulic rigid coupling structure, ensuring efficient and stable energy transfer even in deep, fractured, and soft strata. For receiving, the module integrates an array of high-sensitivity detectors, capable of simultaneously acquiring wake scattering signals from different directions. Built-in anti-collapse protective sleeves and borehole wall displacement sensors monitor the borehole status in real time, ensuring safe operation of the receiving unit in complex environments. Simultaneously, the module also integrates a deep positioning and rock pressure monitoring unit. It obtains precise three-dimensional coordinates of the excitation and receiving points through a GPS and BeiDou dual-mode positioning module, and collects rock pressure data in real time through a miniature geostress sensor, providing real-time feedback data for the automatic adaptive adjustment unit, forming a closed-loop control.

[0083] The coal-bearing rare metal and concealed structure collaborative detection device provided in this embodiment uses an automatic adaptive excitation-receiving module to achieve real-time optimization of excitation energy, coupling strength and receiving position through an automatic adaptive adjustment unit, which significantly improves the signal excitation efficiency and acquisition quality in deep fractured and soft strata.

[0084] Based on the above embodiments, in this embodiment, the multi-dimensional anti-interference signal acquisition and transmission module includes: The signal acquisition unit uses a multi-channel synchronous acquisition chip to synchronously acquire wake wave signals, rock pressure signals, and borehole wall displacement signals. The composite anti-interference preprocessing unit has a built-in composite algorithm chip for curve transform, empirical mode decomposition and deep rock mass attenuation compensation, which is used to remove noise and correct deep rock mass attenuation in the acquired signal. The dual-mode redundant transmission unit adopts three transmission modes: wired transmission, wireless transmission, and fiber optic transmission, and has a built-in high-speed cache unit to cache data when the network connection is lost.

[0085] Specifically, to address issues such as strong vibrations, rock mass attenuation, and mud interference in deep, fractured strata, the anti-interference and transmission functions are optimized, including: Signal acquisition unit: Employing a multi-channel synchronous acquisition chip, it can simultaneously acquire wake wave signals, rock mass pressure signals, and borehole wall displacement signals. The acquisition interval is dynamically adjustable (0.05~1s). It has a built-in abnormal signal identification module that monitors the signal-to-noise ratio in real time. When the signal-to-noise ratio falls below 30dB, it automatically triggers re-excitation and signal enhancement commands. Composite anti-interference preprocessing unit: It incorporates a chip with a composite algorithm of "curved wave transform + EMD + deep rock mass attenuation compensation." First, it uses curved wave transform and EMD to remove random noise and high-frequency interference. Then, it uses a deep rock mass attenuation compensation algorithm to correct signal distortion and energy loss caused by deep rock mass attenuation, improving the signal-to-noise ratio to over 40dB and effectively preserving subtle scattering signal changes caused by rare metal enrichment. Dual-mode redundant transmission unit: It adopts a three-mode transmission system of "wired + wireless + fiber optic backup". Wired transmission uses mining flame-retardant cable (transmission rate ≥20Mbps), wireless transmission uses mining intrinsically safe 5G module (transmission distance ≥100m), and fiber optic transmission serves as a backup for deep long-distance transmission (transmission rate ≥100Mbps). It has a built-in 32GB high-speed cache unit, which automatically caches data when the connection is lost and uploads it synchronously after the connection is restored, ensuring that data is not lost in the deep and complex environment.

[0086] The coal-bearing rare metal and concealed structure collaborative detection device provided in this embodiment effectively solves the problems of easy borehole collapse, large excitation energy loss and weak signal reception in deep soft strata, and significantly improves the success rate and signal-to-noise ratio in signal acquisition.

[0087] Based on the above embodiments, in this embodiment, the three-dimensional collaborative identification processing module includes: The dynamic feature extraction unit has built-in principal component analysis and wavelet packet decomposition algorithm modules to extract the core features of wake scattering, and calculates the rock mass damage factor by combining rock mass pressure data to construct a joint feature set; The three-dimensional collaborative identification unit integrates a Bayesian-optimized convolutional neural network and a long short-term memory network collaborative identification algorithm chip, which is used to combine the correlation model between scattering features, rock mass damage factors and mineralization degree to simultaneously complete the collaborative identification of rare metals and concealed structures. The dynamic model correction unit is used to receive measured data from the verification borehole and automatically correct the correlation model and identification threshold.

[0088] Specifically, the 3D collaborative identification processing module includes: a dynamic feature extraction unit: This unit incorporates a PCA + wavelet packet decomposition algorithm module to automatically extract three core features of wake scattering: energy, frequency, and spatial characteristics. Simultaneously, it combines rock mass pressure data to calculate rock mass damage factors, constructing a joint feature set of "scattering features - rock mass damage factors." Redundant features are eliminated, retaining core features with a correlation of ≥0.9 with rare metals and concealed structures. The 3D collaborative identification unit integrates a BO-CNN-LSTM collaborative identification algorithm chip, introducing temporal dynamic analysis capabilities. Combined with a "scattering features - rock mass damage factors - mineralization degree" correlation model, it simultaneously completes the collaborative identification of rare metals and concealed structures. It can output the coordinates and grade of rare metals (low: 0.01%~0.05%, medium: 0.05%~0.1%, high: ≥0.1%), and the type, scale, development degree, and rock mass damage level of concealed structures. The identification accuracy is improved to ≥93% (rare metals) and ≥95% (concealed structures), respectively. Dynamic model correction unit: It can receive measured data from verification boreholes (rare metal grade, structural parameters, rock mass damage degree) and real-time data collected by the device, automatically correct the associated model and identification threshold, form a dedicated identification model adapted to the deep fractured and soft strata of the target area, and support remote updates of model parameters, and has self-learning capabilities.

[0089] The coal-bearing rare metal and concealed structure co-detection device provided in this embodiment significantly improves the synchronization, anti-interference ability and transmission stability of signal acquisition in deep, fractured and soft strata.

[0090] Based on the above embodiments, in this embodiment, the dynamic linkage control module includes: The module linkage control unit adopts an industrial-grade programmable logic controller to automatically adjust the excitation energy, acquisition interval and anti-interference parameters based on the identification results, rock pressure data and borehole wall displacement data. A visual control terminal is used to display in real time the wake scattering characteristic map, the distribution map of rare metal enrichment area, the distribution map of concealed structure, the distribution map of rock mass damage, and the working status of each module of the device. The data interaction unit is used to network and interact with the mine dispatch center and drilling equipment to realize the real-time uploading of detection data and identification results, as well as the reception of control commands.

[0091] Specifically, the dynamic linkage control module is the core of this device, realizing the dynamic linkage between various modules and the detection method. This includes: a module linkage control unit: employing an industrial-grade PLC controller, enabling real-time linkage between the excitation-receiving module, the acquisition and transmission module, and the identification and processing module. Based on the identification results, rock pressure data, and borehole displacement data, it automatically adjusts the excitation energy, acquisition interval, and anti-interference parameters without manual intervention, achieving fully automated operation. A visual control terminal: using an industrial-grade 15-inch touchscreen display, it can display real-time wake scattering characteristic maps, rare metal enrichment area distribution maps, concealed structure distribution maps, and rock damage distribution maps. It synchronously displays the working status and parameters of each module, supporting manual parameter adjustment, remote control, and fault alarms for more convenient operation. A data interaction unit: supporting network interaction with the mine dispatch center and drilling equipment, it can upload detection data and identification results to the dispatch center in real time, while simultaneously receiving control commands from the dispatch center to link the drilling equipment and adjust drilling parameters, achieving integrated linkage of "detection-drilling-prevention".

[0092] The coal-series rare metal and concealed structure collaborative detection device provided in this embodiment significantly improves the accuracy of collaborative identification of rare metal enrichment areas and concealed structures, and enables the model to have self-learning and adaptive capabilities.

[0093] Based on the above embodiments, in this embodiment, the device further includes an auxiliary adaptation and expansion module, which includes: A deep-fitting component includes a retractable drill reinforcement sleeve and a high-pressure sealing joint. The drill reinforcement sleeve can adaptively expand and contract according to the drill diameter, and the high-pressure sealing joint is used to withstand the high-pressure environment in the deep environment. Modular expansion interface, located on the device casing, allows for flexible addition or removal of functional modules; The fault self-diagnosis unit has a built-in fault detection chip, which is used to monitor the working status of each module in real time, and automatically locate the fault location and output fault handling suggestions when a fault occurs.

[0094] Specifically, regarding the device's scalability and deep adaptability, the following innovative functions are designed: Deep Adaptability Components: These include a retractable borehole reinforcement sleeve and a high-pressure sealing joint. The borehole reinforcement sleeve can adaptively expand and contract according to the borehole diameter, preventing damage to the device from the collapse of deep, soft rock masses. The high-pressure sealing joint can withstand deep high pressure (≤20MPa), preventing mud and groundwater from entering the device and extending its service life. Modular Expansion Interface: Standardized expansion interfaces are provided, allowing for flexible addition or removal of rare earth detection units, groundwater ion detection units, goaf boundary detection units, etc., adapting to the detection needs of different types of rare metals and different concealed structures, offering extremely high scalability. Fault Self-Diagnosis Unit: A built-in fault detection chip monitors the working status of each module in real time. When a fault occurs, it automatically locates the fault position, issues an alarm signal, and outputs fault handling suggestions, reducing the difficulty of on-site maintenance.

[0095] The coal-bearing rare metal and concealed structure co-detection device provided in this embodiment significantly improves the device's adaptability, scalability, reliability, and ease of on-site maintenance in deep and complex environments.

[0096] The following embodiment provides a complete description of the method and apparatus for the coordinated detection of coal-bearing rare metals and concealed structures provided in this application.

[0097] 1. Project Background: The deep area (1000~1500m deep) of a certain coal-bearing mining area belongs to the deep, soft, and fractured coal-bearing strata. The rock mass has extremely poor cementation and high ground stress (15~25MPa), making it prone to borehole collapse. The area has the potential for enrichment of rare metals such as gallium, niobium, and light rare earth elements (cerium, lanthanum). At the same time, it has developed hidden structures such as deep micro-faults (length ≥30m), collapse columns (diameter ≥8m), and goaf boundaries. It is necessary to simultaneously complete rare metal exploration and hidden structure investigation, and coordinate drilling equipment to adjust operating parameters. Existing technologies and the aforementioned shallow exploration schemes are not suitable for the deep environment, and the detection accuracy and efficiency are extremely low. The area of ​​this exploration is determined to be 2km×2.5km. The exploration objectives are: to accurately delineate the rare metal enrichment area (grade ≥0.01%), identify deep hidden structures and classify risk levels, and coordinate drilling equipment to avoid high-risk areas.

[0098] 2. Device Assembly and Parameter Initialization: Based on the detection target, assemble the collaborative detection device through a standardized expansion interface, adding a rare earth detection unit; combined with regional deep geological data, initialize the parameters as follows: excitation head diameter Φ91mm, initial excitation energy 600J, detector sampling frequency 15kHz, acquisition interval 0.1s, signal-to-noise ratio threshold 30dB, and initial threshold for rock mass damage factor 0.03~0.08; set up 5 calibration points (including 2 deep calibration points) around the detection area, install auxiliary positioning and monitoring units, and complete device calibration.

[0099] 3. Deep Adaptive Excitation-Reception and Multi-Parameter Acquisition: The drilling rig is started and drilled to the target depth (1000~1500m). The excitation-reception module is sent into the borehole. The deep positioning and rock pressure monitoring unit collects rock pressure (15~23MPa) and borehole wall displacement data in real time and transmits them to the control module. The control module automatically adjusts the coupling force (15~35kN) and excitation energy (500~800J) according to the rock pressure data. The controllable seismic source is started, and the excitation points are excited in a diamond grid order (10m spacing). Each excitation point is repeated twice. The array geophone synchronously collects the wake wave scattering signal. During the acquisition process, the abnormal signal identification module detects that the signal-to-noise ratio of 8 excitation points is lower than 30dB and automatically triggers the re-excitation and signal enhancement command to ensure data validity. All data is uploaded to the acquisition and transmission module through three transmission modes and is cached locally.

[0100] 4. Deep-specific anti-interference processing and feature correction: The composite anti-interference preprocessing unit of the acquisition and transmission module processes the data, removes noise through curve transform and EMD, and corrects signal distortion through a deep rock mass attenuation compensation algorithm. After processing, the signal-to-noise ratio is improved to over 42dB. The rock mass damage factor is calculated by combining rock mass pressure data to correct the scattering characteristics. At the same time, topography, rock layer shielding, and deep scattering path correction are completed (path deviation correction accuracy within 0.15m) to obtain effective scattering characteristic signals.

[0101] 5. Dynamic Feature Extraction and 3D Collaborative Identification: The 3D collaborative identification processing module extracts core scattering features such as energy decay rate and frequency drift rate, and combines them with rock mass damage factors to construct a joint feature set. Redundant features are removed using PCA + wavelet packet decomposition algorithm. The joint feature set is then input into the BO-CNN-LSTM algorithm to complete dual-target identification: three gallium-enriched areas, two niobium-enriched areas, and two light rare earth enriched areas (two high-grade areas and three medium-grade areas) are delineated, and their coordinates and grade ranges are clearly defined (accuracy ±0.3m). Five micro-faults (length 30~100m), three collapse columns (diameter 8~30m), and two goaf boundaries are identified, and their types, scales, development degrees, and rock mass damage levels are clearly defined (one fault and one collapse column are strongly developed). During the identification process, the control module reported abnormal signals twice. After automatically adjusting the excitation parameters and re-acquiring, the identification accuracy met the standard.

[0102] 6. Results Correction, Output, and Linkage Control: Six verification boreholes (depth 1000~1500m, diameter 91mm) were selected to collect core samples for rare metal grade detection. Borehole imaging technology was used to verify the structure and rock mass damage level. The verification results showed that the rare metal detection error was ≤5.8%, and the accuracy rate of concealed structure identification was 96%, meeting the engineering requirements. The verification data was input into the model correction unit to correct the model parameters and form a dedicated identification model. The integrated detection results were output through the results output unit and uploaded to the mine dispatch center. The control module linked the borehole equipment to adjust the borehole trajectory offset by 12° / m to avoid high-risk structural areas and accurately locate high-grade rare metal areas.

[0103] 7. Implementation Effect Analysis: This project completed the dual-target detection of a 2km×2.5km area in just 12 days. Compared with the existing technology of "step-by-step detection" (25 days) and the aforementioned shallow case (15 days), the exploration cycle was shortened by 52% and the exploration cost was reduced by 45%. The drilling success rate reached 96%, far exceeding the existing technology and the aforementioned version. The detection accuracy and efficiency were greatly improved, successfully solving the detection problem of deep and soft strata, and realizing the integrated linkage of "detection-drilling-prevention". This fully verifies the feasibility, innovation and engineering practicality of this application.

[0104] Figure 6 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 6 As shown, the electronic device can be a robot or other electronic device. This electronic device may include: a processor 610, a communication interface 620, a memory 630, and a communication bus 640. The processor 610, communication interface 620, and memory 630 communicate with each other via the communication bus 640. The processor 610 can call logical instructions from the memory 630 to execute a method for the coordinated detection of coal-bearing rare metals and concealed structures, applied to a coordinated detection device. The method includes: Acquire raw seismic coma scattering signals and rock pressure data of the target area; The rock mass damage factor is calculated based on the rock mass pressure data, and the rock mass damage factor is used to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The core scattering features of the effective scattering feature signal are extracted and combined with the rock mass damage factor to construct a joint feature set. The joint feature set is then input into a pre-trained collaborative identification model to output the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factor and map them to the degree of mineralization.

[0105] Furthermore, the logical instructions in the aforementioned memory 630 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application embodiment, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in at least one embodiment of this application embodiment. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0106] On the other hand, this application also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the coal-bearing rare metals and concealed structures collaborative detection method provided by the above methods, and apply it to a collaborative detection device. The method includes: Acquire raw seismic coma scattering signals and rock pressure data of the target area; The rock mass damage factor is calculated based on the rock mass pressure data, and the rock mass damage factor is used to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The core scattering features of the effective scattering feature signal are extracted and combined with the rock mass damage factor to construct a joint feature set. The joint feature set is then input into a pre-trained collaborative identification model to output the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factor and map them to the degree of mineralization.

[0107] In another aspect, embodiments of this application also provide a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, this computer program implements the methods for the coordinated detection of coal-bearing rare metals and concealed structures provided by the above methods, and is applied to a coordinated detection device. The method includes: Acquire raw seismic coma scattering signals and rock pressure data of the target area; The rock mass damage factor is calculated based on the rock mass pressure data, and the rock mass damage factor is used to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The core scattering features of the effective scattering feature signal are extracted and combined with the rock mass damage factor to construct a joint feature set. The joint feature set is then input into a pre-trained collaborative identification model to output the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factor and map them to the degree of mineralization.

[0108] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0109] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the embodiments of this application, and are not intended to limit them; although the embodiments of this application have been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for synergistically detecting coal-based rare metals and concealed structures, characterized in that, Applied to a collaborative detection device, the method includes: Acquire raw seismic coma scattering signals and rock pressure data of the target area; The rock mass damage factor is calculated based on the rock mass pressure data, and the rock mass damage factor is used to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The core scattering features of the effective scattering feature signal are extracted and combined with the rock mass damage factor to construct a joint feature set. The joint feature set is then input into a pre-trained collaborative identification model to output the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factor and map them to the degree of mineralization.

2. The coal-based rare metal and concealed structure collaborative detection method according to claim 1, characterized in that, The raw seismic coma scattering signals and rock pressure data of the target area are collected, including: Real-time acquisition of rock mass pressure data and borehole wall displacement data at the target depth in the target area; Each excitation point is excited sequentially according to a preset order, and the original seismic coma scattered signal is collected synchronously. During the collection process, the signal-to-noise ratio is monitored in real time. If the signal-to-noise ratio is lower than a preset threshold, a re-excitation operation is automatically triggered.

3. The coal-based rare metal and concealed structure collaborative detection method according to claim 1, characterized in that, The process of using the rock mass damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal includes: The original seismic coma scattered signal is subjected to curvelet transform and empirical mode decomposition to obtain the first intermediate signal; The first intermediate signal is corrected according to the pre-established deep rock mass attenuation compensation algorithm to obtain the second intermediate signal; wherein, the deep rock mass attenuation compensation algorithm is used to dynamically calculate the attenuation compensation coefficient based on the rock mass pressure data and the signal propagation path; The second intermediate signal is corrected using the rock mass damage factor to obtain the effective scattering characteristic signal.

4. The coal-based rare metal and concealed structure collaborative detection method according to claim 3, characterized in that, After obtaining the effective scattering characteristic signal, the method further includes: The effective scattering feature signal is corrected in multiple dimensions to obtain the corrected effective scattering feature signal; The multi-dimensional correction includes terrain correction, rock layer shielding correction, and deep scattering path correction. The deep scattering path correction adopts an improved dynamic time warping algorithm and combines deep rock mass physical parameters to improve path matching accuracy. The improved dynamic time warping algorithm introduces deep rock mass physical parameters as path constraints into the traditional dynamic time warping algorithm to achieve scattering path matching.

5. The coal-based rare metal and concealed structure collaborative detection method according to claim 4, characterized in that, The extraction of the core scattering features of the effective scattering feature signal, combined with the rock mass damage factor to construct a joint feature set, includes: Extract core scattering features from the effective scattering feature signal; wherein, the core scattering features include energy decay rate, frequency drift rate, and scattering angle distribution; The core scattering features are combined with the rock mass damage factor to construct an initial joint feature set; By using principal component analysis and wavelet packet decomposition algorithms, redundant features in the initial joint feature set are removed, and features whose correlation with rare metal enrichment areas and hidden structures reaches a preset threshold are retained, thus obtaining the joint feature set.

6. The method for coordinated detection of coal-bearing rare metals and concealed structures according to claim 1, characterized in that, The step of inputting the joint feature set into a pre-trained collaborative identification model and outputting distribution information of rare metal enrichment areas and hidden structures includes: The joint feature set is input into the collaborative identification model, which simultaneously outputs the three-dimensional coordinates, grade, mineralization range of the rare metal enrichment area, as well as the type, scale, burial depth, development degree and rock mass damage level of the concealed structure. The grade is divided into low grade, medium grade and high grade according to the mass fraction of rare metals.

7. The method for coordinated detection of coal-bearing rare metals and concealed structures according to claim 6, characterized in that, After outputting the distribution information of rare metal enrichment regions and concealed structures, the method further includes: Obtain measured data; wherein, the measured data includes rare metal grade, structural parameters, and rock mass damage degree; The model parameters and identification thresholds of the collaborative identification model are automatically corrected based on the measured data to form a unique identification model adapted to the target area.

8. A device for the coordinated detection of coal-bearing rare metals and concealed structures, characterized in that, include: The deep adaptive excitation-receiver module is used to acquire raw seismic coma scattering signals and rock pressure data in the target area; A multi-dimensional anti-interference signal acquisition and transmission module is used to calculate the rock mass damage factor based on the rock mass pressure data, and to use the rock mass damage factor to perform anti-interference processing and correction on the original seismic coma scattering signal to obtain an effective scattering characteristic signal. The three-dimensional collaborative identification processing module is used to extract the core scattering features of the effective scattering feature signal, and construct a joint feature set by combining it with the rock mass damage factor. The joint feature set is then input into a pre-trained collaborative identification model, which outputs the distribution information of rare metal enrichment areas and concealed structures. The collaborative identification model is a dynamic identification model used to correlate scattering features with rock mass damage factors and map them to the degree of mineralization. The dynamic linkage control module is connected to the deep adaptive excitation-reception module, the multi-dimensional anti-interference signal acquisition and transmission module, and the three-dimensional collaborative identification and processing module, and includes an industrial-grade programmable logic controller, a visual control terminal, and a data interaction unit.

9. The coal-bearing rare metal and concealed structure co-detection device according to claim 8, characterized in that, The deep adaptive excitation-reception module includes: The adaptive excitation unit adopts an integrated design of a controllable vibration source and a drilling excitation head. The diameter of the excitation head is adjustable, and the end of the excitation head is equipped with a wear-resistant and anti-slip probe and adopts a hydraulic rigid coupling structure to dynamically adjust the coupling force according to the pressure of the deep rock mass. The high-precision receiving unit consists of an array of high-sensitivity detectors, a collapse-proof protective sleeve, and a signal amplification module. The sampling frequency of the detectors can be dynamically adjusted. The deep positioning and rock pressure monitoring unit integrates a positioning module and a miniature geostress sensor to collect the three-dimensional coordinates of the excitation and receiving points as well as deep rock pressure data in real time.

10. The coal-bearing rare metal and concealed structure co-detection device according to claim 8, characterized in that, The multi-dimensional anti-interference signal acquisition and transmission module includes: The signal acquisition unit uses a multi-channel synchronous acquisition chip to synchronously acquire wake wave signals, rock pressure signals, and borehole wall displacement signals. The composite anti-interference preprocessing unit has a built-in composite algorithm chip for curve transform, empirical mode decomposition and deep rock mass attenuation compensation, which is used to remove noise and correct deep rock mass attenuation in the acquired signal. The dual-mode redundant transmission unit adopts three transmission modes: wired transmission, wireless transmission, and fiber optic transmission, and has a built-in high-speed cache unit to cache data when the network connection is lost.

11. The coal-bearing rare metal and concealed structure co-detection device according to claim 8, characterized in that, The three-dimensional collaborative identification and processing module includes: The dynamic feature extraction unit has built-in principal component analysis and wavelet packet decomposition algorithm modules to extract the core features of wake scattering, and calculates the rock mass damage factor by combining rock mass pressure data to construct a joint feature set; The three-dimensional collaborative identification unit integrates a Bayesian-optimized convolutional neural network and a long short-term memory network collaborative identification algorithm chip, which is used to combine the correlation model between scattering features, rock mass damage factors and mineralization degree to simultaneously complete the collaborative identification of rare metals and concealed structures. The dynamic model correction unit is used to receive measured data from the verification borehole and automatically correct the correlation model and identification threshold.

12. The coal-bearing rare metal and concealed structure co-detection device according to claim 8, characterized in that, The dynamic linkage control module includes: The module linkage control unit adopts an industrial-grade programmable logic controller to automatically adjust the excitation energy, acquisition interval and anti-interference parameters based on the identification results, rock pressure data and borehole wall displacement data. A visual control terminal is used to display in real time the wake scattering characteristic map, the distribution map of rare metal enrichment area, the distribution map of concealed structure, the distribution map of rock mass damage, and the working status of each module of the device. The data interaction unit is used to network and interact with the mine dispatch center and drilling equipment to realize the real-time uploading of detection data and identification results, as well as the reception of control commands.

13. The coal-bearing rare metal and concealed structure co-detection device according to any one of claims 8-12, characterized in that, The device further includes an auxiliary adaptation and expansion module, which includes: A deep-fitting component includes a retractable drill reinforcement sleeve and a high-pressure sealing joint. The drill reinforcement sleeve can adaptively expand and contract according to the drill diameter, and the high-pressure sealing joint is used to withstand the high-pressure environment in the deep environment. Modular expansion interface, located on the device casing, allows for flexible addition or removal of functional modules; The fault self-diagnosis unit has a built-in fault detection chip, which is used to monitor the working status of each module in real time, and automatically locate the fault location and output fault handling suggestions when a fault occurs.

14. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for coordinated detection of coal-bearing rare metals and concealed structures as described in any one of claims 1 to 7.

15. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the method for coordinated detection of coal-bearing rare metals and concealed structures as described in any one of claims 1 to 7.

Images