Device and method for detecting surrounding rock fissure based on causal feature learning

Abstract

The application relates to the technical field of tunnel surrounding rock crack detection, in particular to a surrounding rock crack detection device and method based on causal feature learning, wherein the device comprises: a wall-climbing unmanned aerial vehicle platform used for controlling at least one wall-climbing unmanned aerial vehicle to perform detection and shooting; an airborne laser scanning system used for obtaining three-dimensional laser point cloud data of surrounding rock cracks; an oblique photography system used for collecting image data of the surrounding rock cracks, establishing a real three-dimensional model; an airborne ground penetrating radar system used for detecting information such as the width, position, internal trend and extension of the surrounding rock cracks; a data fusion system used for fusing and matching the information, generating a sample data set and performing sample calibration; and an image recognition system used for performing crack recognition on the sample data set, generating a crack identification model and identifying the crack type. Thus, the problems of large work load, low recognition accuracy and efficiency and high danger in the related art are solved.

Description

Technical Field

[0001] This application relates to the field of tunnel surrounding rock fissure detection technology, and in particular to a surrounding rock fissure detection device and method based on causal feature learning. Background Technology

[0002] The development and distribution pattern of fissures in the surrounding rock of a tunnel directly affects the stability of the tunnel. Tunnel excavation disrupts the stress balance of the surrounding rock, resulting in complex fissures. The development and distribution pattern of these fissures has a direct impact on the assessment of the degree of damage to the surrounding rock and the selection of support methods. Conducting fissure detection and identification in the surrounding rock of a tunnel helps to accurately assess the stability and potential risks of the surrounding rock.

[0003] However, in related technologies, due to the existence of large areas, numerous and varying sizes of damaged mountains with dangerous surrounding rocks, the workload of detecting cracks in the surrounding rock of tunnels is large, which easily leads to low efficiency and weak identification accuracy in the on-site survey for assessing the stability of dangerous surrounding rocks, and is also highly dangerous, which urgently needs to be solved. Summary of the Invention

[0004] This application provides a rock fissure detection device based on causal feature learning to solve the problems of large workload, low identification accuracy and efficiency, and high risk in tunnel rock fissure detection in related technologies. The first aspect of this application provides a rock fissure detection device based on causal feature learning, comprising: a wall-climbing unmanned aerial vehicle (UAV) platform for controlling at least one wall-climbing UAV to perform flight tasks and attitude control actions in a target area according to a preset flight path, thereby detecting and photographing the target detection area; an airborne laser scanning system for obtaining three-dimensional laser point cloud data of the detected rock fissures when the at least one wall-climbing UAV performs the flight tasks and attitude control actions; an oblique photography system for acquiring image data of the detected rock fissures and establishing a real-scene three-dimensional model of the target detection area; an airborne ground-penetrating radar system for detecting at least one of the following information: width, location, internal orientation, and extension of the detected rock fissures; a data fusion system for fusing and matching the three-dimensional laser point cloud data, the real-scene three-dimensional model, the image data, and / or the at least one of the information to generate a sample dataset, generate a sample database for sample calibration, and constitute a training set in the sample database; and an image recognition system for identifying fissures in the training set and generating a fissure identification model to identify any type of rock fissure using the fissure identification model.

[0005] Optionally, in one embodiment of this application, the wall-climbing drone platform includes: a flight control module for controlling at least one wall-climbing drone to perform flight tasks and attitude control actions; and a data storage module for storing the three-dimensional laser point cloud data and the image data.

[0006] Optionally, in one embodiment of this application, the data fusion system includes: a data fusion module, used to fuse and match the three-dimensional laser point cloud data, the real-world three-dimensional model, and the image data to obtain fused data; a sample database, used to crop and vectorize the fused data to the required size, generate a sample dataset according to the coordinate sequence, generate a sample database, and dynamically update the sample database in real time; and a sample calibration module, used to calibrate the geological elements in the oblique photography images and ground-penetrating radar profiles in the sample database to form a training set in the sample database.

[0007] Optionally, in one embodiment of this application, the image recognition system includes: a crack identification module, used to train and test the neural network in conjunction with the training set to generate a crack identification model.

[0008] Optionally, in one embodiment of this application, the image recognition system is further used to separate causal and non-causal features of the surrounding rock fracture sample data based on the neural network, so as to reduce the confusion of the non-causal features for prediction by a preset means, learn the causal relationship between the surface features and internal features of the surrounding rock fracture, and generate a fracture identification model.

[0009] Optionally, in one embodiment of this application, the image recognition system is further configured to input the sample dataset into the neural network, learn and update the representation vector of each node, transform the graph representation using a multilayer perceptron, separate causal features and non-causal features, randomly recombine the causal features and the non-causal features to obtain a new graph-level representation after intervention, which is then input into the loss function to optimize the parameters of the crack identification model.

[0010] The second aspect of this application provides a method for detecting surrounding rock fissures based on causal feature learning, comprising the following steps: a wall-climbing unmanned aerial vehicle (UAV) platform performs a detection flight mission and attitude positioning operation required for the mission according to a preset planned route, so as to conduct multi-level and multi-directional detection of surrounding rock fissure feature data; collects image data of the detected surrounding rock fissures and establishes a real-scene 3D model of the target detection area; detects at least one of the following information: width, location, internal direction, and extension of the detected surrounding rock fissures; fuses and matches the 3D laser point cloud data, the real-scene 3D model, and the image data to generate a sample dataset; performs sample calibration and fissure identification on the sample dataset to generate a fissure identification model, so as to use the fissure identification model to identify any type of surrounding rock fissure.

[0011] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the rock fracture detection method based on causal feature learning as described in the above embodiments.

[0012] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for detecting rock fissures based on causal feature learning.

[0013] A fifth aspect of this application provides a computer program, which, when executed, is used to implement the above-described method for detecting rock fissures based on causal feature learning.

[0014] This application embodiment utilizes a wall-climbing drone platform to control the drone to conduct detection and imaging in a target area according to a preset flight path. Using an onboard laser scanning system, it obtains three-dimensional laser point cloud data of the detected rock fissures. An oblique photography system collects image data of the detected rock fissures, establishing a realistic three-dimensional model of the target detection area. An onboard ground-penetrating radar system detects the width, location, internal orientation, and extension of the detected rock fissures. A data fusion system then performs data fusion and matching to generate a sample dataset. An image recognition system is then used to generate a fissure identification model, enabling the identification of different types of rock fissures, effectively improving the accuracy and efficiency of identification. This solves the problems of high workload, low accuracy and efficiency, and high risk associated with tunnel rock fissure detection in related technologies.

[0015] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0016] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0017] Figure 1 This is a schematic diagram of a rock fissure detection device based on causal feature learning according to an embodiment of this application;

[0018] Figure 2 This is a schematic diagram of the framework of a rock fissure detection device based on causal feature learning according to an embodiment of this application;

[0019] Figure 3 This is a schematic diagram of the structure of a deep learning neural network based on causal feature learning according to an embodiment of this application;

[0020] Figure 4 This is a schematic diagram of the causal structure of causal feature learning according to an embodiment of this application;

[0021] Figure 5 This is a flowchart of a method for detecting fractures in surrounding rock based on causal feature learning, according to an embodiment of this application.

[0022] Figure 6 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Detailed Implementation

[0023] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0024] The following describes an embodiment of the rock fissure detection device and method based on causal feature learning according to the accompanying drawings. Addressing the problems of high workload, low accuracy and efficiency, and high risk associated with traditional tunnel rock fissure detection methods mentioned in the background section, this application provides a rock fissure detection device based on causal feature learning. In this device, a wall-climbing drone platform can be used to control the drone to perform detection and imaging in a target area according to a preset flight path. An airborne laser scanning system is used to obtain three-dimensional laser point cloud data of the detected rock fissures. An oblique photography system is used to collect image data of the detected rock fissures, establishing a realistic three-dimensional model of the target detection area. An airborne ground-penetrating radar system is used to detect the width, location, internal orientation, and extension of the detected rock fissures. These data are then fused and matched using a data fusion system to generate a sample dataset. An image recognition system is then used to generate a fissure identification model, enabling the identification of different types of rock fissures and effectively improving the accuracy and efficiency of identification. This solves the problems of high workload, low efficiency, and high risk associated with traditional tunnel rock fissure detection methods.

[0025] Specifically, Figure 1 This is a schematic diagram of the structure of a rock fissure detection device based on causal feature learning, provided in an embodiment of this application.

[0026] like Figure 1 As shown, the rock fissure detection device 10 based on causal feature learning includes: a wall-climbing UAV platform 100, an airborne laser scanning system 200, an oblique photography system 300, an airborne ground-penetrating radar system 400, a data fusion system 500, and an image recognition system 600.

[0027] Specifically, the wall-climbing drone platform 100 is used to control at least one wall-climbing drone to perform flight tasks and attitude control actions in the target area according to a preset route, so as to detect and photograph the target detection area.

[0028] It is understood that the preset flight path refers to the pre-planned and set flight path of the wall-climbing drone, including but not limited to the starting point of the wall-climbing drone, the spacing of the flight path, the direction of travel, the altitude, etc. The specific settings can be made by professionals in this field according to actual needs and the characteristics of the target area, and no specific restrictions are imposed here.

[0029] Specifically, the embodiments of this application utilize a wall-climbing drone platform, which can carry various airborne devices to perform loop-like flight and detection tasks on the surrounding rock of the tunnel. It can also perform tasks such as route planning, flight control, and position recording for the wall-climbing drone. Wireless transmission technology is used, including but not limited to wireless local area network and mobile communication technology, to realize data transmission of the wall-climbing drone flight control system and airborne device detection data transmission.

[0030] This application embodiment can realize the autonomous flight and attitude control of the wall-climbing UAV platform based on its own autonomous route planning and autonomous navigation functions, so as to detect and photograph the surrounding rock of the target tunnel, ensure flight stability, ensure coverage of the entire target area, effectively improve mission execution efficiency, and ensure safety.

[0031] Optionally, in one embodiment of this application, the unmanned aerial vehicle (UAV) platform 100 includes a flight control module and a data storage module. The flight control module is used to control at least one UAV to perform flight tasks and attitude control actions; the data storage module is used to store three-dimensional laser point cloud data and image data.

[0032] It is understandable that flight missions may include, but are not limited to, takeoff, landing, hovering, and heading adjustments, and attitude control actions may include, but are not limited to, tilting, turning, climbing, and descending.

[0033] Specifically, in combination Figure 2 As shown, the wall-climbing UAV platform includes a flight control module and a data storage module. The flight control module can share information from the attitude and positioning sensors, inertial measurement unit, and airborne laser scanning. It can also achieve autonomous flight and attitude control based on the wall-climbing UAV platform's own autonomous route planning and autonomous navigation functions. The data storage module of the wall-climbing UAV platform can store the laser point cloud data and image data collected and recorded in real time by the airborne laser scanning system and oblique photography system in the following steps. The storage capacity meets the data acquisition requirements of a complete detection operation and also meets the requirements for real-time lossless data transmission during flight.

[0034] The embodiments of this application can utilize a flight control module to control the wall-climbing drone to perform flight missions and achieve attitude control actions, thereby achieving precise control over the flight process of the wall-climbing drone. This helps to adapt to different environments and mission requirements, maintain a stable flight state, and utilize a data storage module to ensure that the collected data can be stored and backed up in a timely and secure manner after the wall-climbing drone has performed its mission, providing strong support for the subsequent effective use of the data.

[0035] The airborne laser scanning system 200 is used to obtain three-dimensional laser point cloud data for detecting fissures in surrounding rock when at least one rock-climbing UAV is performing flight missions and attitude control actions.

[0036] This application embodiment can utilize an airborne laser scanning system to scan and detect the surface of the surrounding rock in a tunnel, extract surface features of the surrounding rock fissures, and measure the position and attitude information of a wall-climbing drone, thereby obtaining three-dimensional laser point cloud data for detecting surrounding rock fissures, effectively improving the feasibility of detecting surrounding rock fissures.

[0037] Optionally, in one embodiment of this application, the airborne laser scanning system 200 includes: an airborne laser scanner, an attitude positioning sensor, and an inertial measurement unit. The attitude positioning sensor and inertial measurement unit are used to calculate the attitude state of the wall-climbing unmanned aerial vehicle platform in real time; the airborne laser scanner is used to acquire three-dimensional laser point cloud data for detecting fissures in the surrounding rock.

[0038] It is understandable that pose state refers to parameters describing the position and attitude state of an object, including information such as position on the coordinate axes, pitch angle, roll angle, yaw angle and velocity. Three-dimensional laser point cloud data refers to the set of point coordinate data of the surface of a target object in three-dimensional space obtained by laser scanning and other equipment.

[0039] In actual implementation, combined with Figure 2 As shown, the embodiments of this application can use attitude positioning sensors to measure the position and attitude information of the wall-climbing UAV, compensate for dynamic errors through an inertial measurement unit, and obtain information such as three-dimensional laser point cloud data for detecting surrounding rock fissures through airborne laser scanning. Then, based on the autonomous route planning and autonomous navigation functions of the wall-climbing UAV platform, autonomous control of the wall-climbing UAV platform can be realized, thereby improving the intelligence level of the wall-climbing UAV platform in detecting surrounding rock fissures.

[0040] As one possible approach, embodiments of this application can utilize airborne laser scanning to quickly acquire three-dimensional coordinate point cloud data of the detection area, and can quickly acquire a high-precision, high-resolution digital surface model of the surrounding rock fissures, thereby improving the accuracy of the detection data acquisition.

[0041] This application embodiment utilizes attitude-fixing and positioning sensors and inertial measurement units to calculate the attitude state of the wall-climbing UAV in real time, and uses airborne laser scanning to collect three-dimensional laser point cloud data of surrounding rock fissures, which helps to improve the accuracy and effectiveness of data acquisition, and effectively ensures the accuracy and stability of the wall-climbing UAV's flight.

[0042] The oblique photography system 300 is used to acquire image data of surrounding rock fissures and establish a realistic 3D model of the target detection area.

[0043] In some embodiments, the present application embodiments utilize an oblique photography system to capture images of the surface conditions and surrounding rock fissures of the detection area from multiple angles, such as vertical and oblique, and establish a three-dimensional model of the real-world scene of the detection area. This helps to improve the accuracy and completeness of the obtained surface condition information of the detection area and enhances the accuracy of surrounding rock fissure detection.

[0044] Optionally, in one embodiment of this application, the oblique photography system 300 includes: a photography module and an automatic modeling module. The photography module is used to capture images of the target detection area from multiple angles to obtain image data; the automatic modeling module is used to generate a realistic 3D model based on the image data.

[0045] Understandably, the photography module contains multiple cameras used to capture images of the detection area from multiple angles, including vertical, front-view, rear-view, left-view, and right-view, to obtain image information of the surrounding rock fissures.

[0046] In some embodiments, combined with Figure 2 As shown, the oblique photography system includes a photography module and an automatic modeling module. The photography module contains multiple cameras that can capture images of the detection area from vertical, oblique, and other angles. In addition, the photography module can calculate data such as the relative flight altitude, flight path interval, and photo interval of the wall-climbing UAV platform according to the required mapping scale and resolution values, and transmit them back to the flight control module. The oblique photography system can be equipped with an automatic modeling module, which can automatically identify the image data of the detection area captured by the photography module from vertical, oblique, and other angles, and generate a 3D reality model using methods such as image matching, aerial triangulation, and texture mapping, thereby improving the accuracy of the 3D reality model of the surrounding rock fissures.

[0047] Airborne ground-penetrating radar system 400 is used to detect at least one of the following information: the width, location, internal orientation, and extension of surrounding rock fissures.

[0048] Specifically, the embodiments of this application can utilize an airborne ground-penetrating radar system to conduct multi-level and multi-directional detection of the width and direction of surface fissures on the damaged mountain surface, as well as the depth and extension structure of underground fissures, by emitting electromagnetic waves and receiving reflected signals. This is beneficial for more accurate analysis of the structure and fracture status of rock strata, providing important reference for subsequent geological exploration and rock strata treatment, improving the comprehensiveness and efficiency of monitoring, and reducing time and manpower costs.

[0049] The data fusion system 500 is used to fuse and match 3D laser point cloud data, real-world 3D models, image data and / or at least one of these information to generate a sample dataset, generate a sample database for sample calibration, and form a training set in the sample database.

[0050] This application embodiment can utilize a data fusion system to preprocess, coordinate match, texture map, and feature recognize data such as 3D laser point cloud data, 3D real-world models, and ground-penetrating radar profile images obtained by airborne laser scanning systems, oblique photography systems, and airborne ground-penetrating radar systems. This generates a 3D model of surrounding rock fissures with high-precision coordinates, high-accuracy surface features, and shallow underground radar detection features. This effectively improves data consistency and accuracy, generates a more realistic and reliable sample dataset, and performs sample calibration to form a training set in the sample database, which is beneficial for providing a foundation for subsequent data analysis and decision-making.

[0051] Optionally, in one embodiment of this application, the data fusion system 500 includes: a data fusion module, a sample database, and a sample calibration module. The data fusion module is used to fuse and match 3D laser point cloud data, real-world 3D models, and image data to obtain fused data. The sample database is used to crop and vectorize the fused data to the required size, generate a sample dataset according to a coordinate sequence, and dynamically update the sample database in real time. The sample calibration module is used to calibrate the geological elements in the oblique photographic images and ground-penetrating radar profiles in the sample database to form a training set combined with the sample database.

[0052] It is understandable that the coordinate sequence can be geographic coordinates, time series, or other ordered indicators.

[0053] Specifically, in combination Figure 2As shown, the data fusion system includes a data fusion module, a sample database, and a sample calibration module. The data fusion module, through functions such as data verification, coordinate transformation, and image matching, performs coordinate matching and fusion calibration of laser point cloud data with surface image information and subsurface geophysical data to obtain an accurate data model of the probed area with three-dimensional coordinate information. The sample database has data preprocessing functions, performing operations such as denoising, correction, and 3D reconstruction on the probed data. It can partition and store raw data, processed data, calibrated geological elements, and probe results to improve processing efficiency. Simultaneously, within the sample database… After cropping and vectorization, the samples can be organized according to coordinate sequences to form an ordered dataset, thereby generating a sample database, which is dynamically updated in real time. The sample calibration module can be used by professional technicians to manually calibrate geological elements in oblique photogrammetric images and ground-penetrating radar profiles based on the sample database. These geological elements include, but are not limited to, fracture types (including but not limited to joint fractures, joint fillings, rock layer fractures, vein mineral fractures, etc.); fracture density; fracture direction and angle; fracture width; and fracture morphology (including but not limited to linear, arc-shaped, and bifurcated types, etc.), thus forming a training set.

[0054] This application embodiment, by dynamically updating the sample database in real time, helps ensure the timeliness and accuracy of the sample dataset, enabling application scenarios with real-time feedback and rapid response.

[0055] Image recognition system 600 is used to identify fractures in sample datasets and generate fracture identification models to identify any type of fracture in surrounding rock.

[0056] It is understandable that the types of fractures in the surrounding rock include, but are not limited to, cracks, joints, faults, etc.

[0057] Specifically, in this application embodiment, an image recognition system can be used to perform sample calibration and crack identification on the sample dataset, construct a deep learning neural network based on causal feature learning, and train and test the neural network in combination with the training set in the sample database to generate a crack identification model and identify any type of surrounding rock crack.

[0058] Furthermore, combined Figure 2 As shown, the fracture discrimination model is based on a deep learning neural network constructed by causal feature learning. It can separate the causal and non-causal features of oblique photography images and ground-penetrating radar images. By using different methods, it can reduce the confusion of non-causal features in prediction, learn the causal relationship between the surface features and underground features of surrounding rock fractures, and thus obtain a model with more generalization ability, thereby improving the high-precision identification capability of surrounding rock fractures.

[0059] Optionally, in one embodiment of this application, the image recognition system 600 includes a crack identification module. The crack identification module is used to train and test a neural network using a training set to generate a crack identification model.

[0060] Specifically, in this embodiment, the crack identification module can construct a deep learning neural network based on causal feature learning, and train and test the neural network using a training set in a sample database to obtain a crack identification model. Furthermore, the deep learning neural network can use a loss function to compare the predicted values ​​with the true values ​​to obtain a loss value, which is used to measure the recognition accuracy of the crack identification model. Simultaneously, the optimizer will update the model weights based on the loss value.

[0061] As one possible approach, deep learning neural networks built based on causal feature learning can separate causal and non-causal features from graph data, and reduce the confusion of non-causal features in predictions through different means, thereby obtaining a more generalizable model and improving the ability to identify rock fissures with high precision.

[0062] Optionally, in one embodiment of this application, the image recognition system 600 is further used to separate the causal features and non-causal features of the surrounding rock fracture sample data based on a neural network, so as to reduce the confusion of non-causal features for prediction by a preset means, learn the causal relationship between the surface features and internal features of the surrounding rock fracture, and generate a fracture identification model.

[0063] Understandably, the pre-defined approach can be understood as randomly recombining causal and non-causal features to obtain a new graph-level representation after intervention.

[0064] Optionally, in one embodiment of this application, the image recognition system 600 is further configured to input the sample dataset into the neural network, learn and update the representation vector of each node, use a multilayer perceptron to transform the graph representation, separate causal features and non-causal features, and randomly recombine the causal features and non-causal features to obtain a new graph-level representation after intervention, which is then input into the loss function to optimize the parameters of the crack identification model.

[0065] As one possible way to achieve this, such as Figure 3 As shown, a deep learning neural network is constructed based on causal feature learning. The specific steps for implementing the neural network are as follows:

[0066] 1. Graph Feature Learning: A neural network model is used to learn and update the representation vectors of each node in the graph. Each node is represented as a vector containing its position in the graph structure, connectivity, and other relevant attributes. By reading out the function to aggregate the node representation information, these local features can be aggregated into a graph-level representation vector, thereby achieving a holistic understanding and learning of the entire surrounding rock fracture network.

[0067] 2. Causal Feature Separation: To better understand the causal relationships in the surrounding rock fracture network, a multilayer perceptron can be used to transform the graph representation in different ways, thereby separating causal and non-causal features. By controlling the optimization objective, the final graph representation can more clearly reflect the causal relationships in the surrounding rock fractures, which is helpful for subsequent analysis and prediction.

[0068] 3. Causal Intervention: After separating causal and non-causal features, these features can be randomly recombined to obtain a new graph-level representation after intervention. This representation can reduce the interference of non-causal features on causal features and increase the interpretability of the surrounding rock fracture network features.

[0069] 4. Joint Training: After obtaining the feature representation of the surrounding rock fracture network, the results from the classifier are aggregated into the loss function, and the model parameters are optimized using the loss value. This ultimately yields a prediction model that is more relevant to causal features, enabling more accurate prediction of the location, morphology, and other related properties of surrounding rock fractures.

[0070] In actual implementation, oblique photogrammetry images and ground-penetrating radar profiles contain surface features such as the width and orientation of surface fissures in the surrounding rock, and underground features such as the depth and extension structure of underground fissures. The distribution and extension of the surrounding rock fissures have a certain causal relationship with both surface and underground features. Using a deep neural network based on causal feature learning, features are extracted and learned from oblique photogrammetry images and ground-penetrating radar profiles to learn the causal relationship between surface and underground features of the surrounding rock fissures, thereby achieving a three-dimensional identification of the surrounding rock fissure network.

[0071] Furthermore, such as Figure 4 As shown, the causal structure graph for causal feature learning includes: input graph data G, graph representation R, causal features C, non-causal features N, and prediction result Y.

[0072] In the structural cause-effect graph, each link represents a causal relationship, specifically as follows:

[0073] (1) G→R represents the original graph data being transformed into a vector form after being learned by the neural network;

[0074] (2) C←R→N represents the graph representation that simultaneously contains two types of features that affect the prediction results: causal and non-causal.

[0075] (3) C→Y←N represents that the general model takes into account both causal and non-causal features when making predictions.

[0076] In actual execution, the sample dataset is input into the deep learning neural network. The oblique photogrammetry image and the ground penetrating radar profile are vectorized by the input graph data G, and graph feature extraction and graph classification are performed respectively to extract causal features and non-causal features. After causal intervention, the causal and non-causal features are entered into the classifier, and the output is used as the final prediction result.

[0077] In addition, combined Figure 1 As shown, in this embodiment, the airborne laser scanning system, oblique photography system, airborne ground-penetrating radar system, data fusion system, and image recognition system can be directly mounted on the wall-climbing drone platform to form an integrated connection, thereby achieving closer collaboration and functional integration. Alternatively, the airborne laser scanning system, oblique photography system, and airborne ground-penetrating radar system can be directly mounted on the wall-climbing drone platform, with the data fusion system and image recognition system physically connected to the wall-climbing drone platform, such as through cables or other direct hardware interfaces, to ensure the real-time performance and stability of data transmission.

[0078] It is worth noting that, Figure 1 The connection method shown is only an exemplary illustration of this application. Other different connection methods may be used according to different actual needs, including: (1) directly mounting the airborne laser scanning system, oblique photography system and airborne ground penetrating radar system on the wall-climbing UAV platform, and communicating with the data fusion system and image recognition system through Wi-Fi, Bluetooth or dedicated communication protocols to transmit data, thereby increasing the flexibility of the system; (2) directly mounting the airborne laser scanning system, oblique photography system and airborne ground penetrating radar system on the wall-climbing UAV platform, and deploying the data fusion system and image recognition system on an independent server connected to the wall-climbing UAV platform, so as to balance the needs of flexibility and computing resources.

[0079] Different connection methods can flexibly meet the connection needs between data fusion systems and image recognition systems and wall-climbing drone platforms under different circumstances. The specific choice can be made by professionals in the field based on the actual application scenario and technical limitations, and no specific restrictions are made here.

[0080] For example, the working principle of the embodiments of this application will be described in detail below with a specific example.

[0081] First, the embodiments of this application can check whether the wall-climbing drone platform and related airborne equipment are working properly, set initialization parameters, plan flight routes, and start the exploration mission through remote manual control of the flight.

[0082] Secondly, the embodiments of this application can acquire detection data such as laser point cloud data, oblique photography images, and ground-penetrating radar images of the detection area based on the actual situation of the detection area using equipment such as airborne laser scanning systems, oblique photography systems, and airborne ground-penetrating radar systems.

[0083] Furthermore, embodiments of this application can utilize a data fusion system to fuse and match probe data and dynamically update the sample database.

[0084] Finally, in this embodiment of the application, an image recognition system can be used to calibrate the sample database, import it into the fracture discrimination model, and identify the type of fracture in the surrounding rock.

[0085] In summary, the fracture identification model based on causal feature learning and deep learning neural network algorithm in this application realizes standardized identification of the type of fracture in the surrounding rock, saves a lot of manpower, improves work efficiency, ensures the safety of workers, and is easy to operate with simple steps, making it easy for workers to learn and master. The on-site preparation and debugging time is short, saving time and facilitating promotion and popularization.

[0086] The rock fissure detection device based on causal feature learning proposed in this application can use a wall-climbing drone platform to control the drone to conduct detection and imaging in the target area according to a preset flight path. It uses an airborne laser scanning system to obtain three-dimensional laser point cloud data of the detected rock fissures, and an oblique photography system to collect image data of the detected rock fissures, establishing a realistic three-dimensional model of the target detection area. Furthermore, it uses an airborne ground-penetrating radar system to detect the width, location, internal orientation, and extension of the detected rock fissures. Then, a data fusion system is used for fusion and matching to generate a sample dataset. Finally, an image recognition system is used to generate a fissure identification model, enabling the identification of different types of rock fissures, effectively improving the accuracy and efficiency of identification. This solves the problems of low efficiency, weak identification accuracy, and high risk associated with manual on-site surveys for assessing the stability of dangerous surrounding rock in related technologies.

[0087] Next, referring to the accompanying drawings, a method for detecting fractures in surrounding rock based on causal feature learning, according to an embodiment of this application, is described.

[0088] Figure 5 This is a flowchart of a rock fissure detection method based on causal feature learning, according to an embodiment of this application.

[0089] like Figure 5As shown, this method for detecting fractures in surrounding rock based on causal feature learning includes the following steps:

[0090] In step S501, at least one wall-climbing UAV is controlled to perform flight tasks and attitude control actions in the target area according to a preset flight path in order to obtain three-dimensional laser point cloud data for detecting fissures in the surrounding rock.

[0091] In step S502, image data of the surrounding rock fissures are acquired and a real-world three-dimensional model of the target detection area is established.

[0092] In step S503, at least one of the following information is detected: the width, location, internal orientation, and extension of the surrounding rock fissure.

[0093] In step S504, the 3D laser point cloud data, the real-world 3D model, and the image data are fused and matched to generate a sample dataset.

[0094] In step S505, the sample dataset is calibrated and fractures are identified to generate a fracture identification model, which is then used to identify any type of fracture in the surrounding rock.

[0095] It should be noted that the foregoing explanation of the embodiment of the surrounding rock fracture detection device based on causal feature learning also applies to the surrounding rock fracture detection method based on causal feature learning in this embodiment, and will not be repeated here.

[0096] The rock fissure detection method based on causal feature learning proposed in this application can utilize a wall-climbing drone platform to control the drone to conduct detection and imaging in a target area according to a preset flight path. Using an onboard laser scanning system, three-dimensional laser point cloud data of the detected rock fissures is obtained. An oblique photography system is used to collect image data of the detected rock fissures, establishing a realistic three-dimensional model of the target detection area. An onboard ground-penetrating radar system is used to detect the width, location, internal orientation, and extension of the detected rock fissures. Then, a data fusion system is used for fusion and matching to generate a sample dataset. Finally, an image recognition system is used to generate a fissure identification model, enabling the identification of different types of rock fissures, effectively improving the accuracy and efficiency of identification. This solves the problems of low efficiency, weak identification accuracy, and high risk associated with manual on-site surveys for assessing the stability of dangerous surrounding rock in related technologies.

[0097] Figure 6 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0098] The memory 601, the processor 602, and the computer program stored on the memory 601 and capable of running on the processor 602.

[0099] When the processor 602 executes the program, it implements the rock fissure detection method based on causal feature learning provided in the above embodiments.

[0100] Furthermore, electronic devices also include:

[0101] Communication interface 603 is used for communication between memory 601 and processor 602.

[0102] The memory 601 is used to store computer programs that can run on the processor 602.

[0103] The memory 601 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0104] If the memory 601, processor 602, and communication interface 603 are implemented independently, then the communication interface 603, memory 601, and processor 602 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 6 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0105] Optionally, in a specific implementation, if the memory 601, processor 602, and communication interface 603 are integrated on a single chip, then the memory 601, processor 602, and communication interface 603 can communicate with each other through an internal interface.

[0106] The processor 602 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0107] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described method for detecting rock fissures based on causal feature learning.

[0108] This application also provides a computer program that can run computer instructions, which, when executed by a processor, implement the above-described method for detecting rock fissures based on causal feature learning.

[0109] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0110] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0111] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0112] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0113] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0114] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0115] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0116] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A rock fracture detection device based on causal feature learning, characterized in that, include: A wall-climbing drone platform is used to control at least one wall-climbing drone to perform flight missions and attitude control actions in a target area according to a preset route, so as to detect and photograph the target detection area. An airborne laser scanning system is used to obtain three-dimensional laser point cloud data for detecting rock fissures when at least one rock-climbing UAV performs the flight mission and the attitude control actions. An oblique photography system is used to acquire image data of the detected rock fissures and to establish a realistic three-dimensional model of the target detection area; An airborne ground-penetrating radar system is used to detect at least one of the following information: the width, location, internal orientation, and extension of the surrounding rock fissure. The data fusion system is used to fuse and match at least one of the following information: the width, location, internal orientation, and extension of the detected surrounding rock fissures, to generate a sample dataset, generate a sample database for sample calibration, and constitute a training set in the sample database. An image recognition system is used to identify fractures in the training set and generate a fracture identification model, so as to identify any type of fracture in the surrounding rock using the fracture identification model. The image recognition system includes: The crack identification module is used to train and test the neural network by combining the training set to generate a crack identification model. The image recognition system is further used to separate causal and non-causal features of the surrounding rock fracture sample data based on the neural network, so as to reduce the confusion of the non-causal features for prediction by a preset means, learn the causal relationship between the surface features and internal features of the surrounding rock fracture, and generate a fracture identification model. The image recognition system is further used to input the sample dataset into the neural network, learn and update the representation vector of each node, use a multilayer perceptron to transform the graph representation, separate causal features and non-causal features, and randomly recombine the causal features and non-causal features to obtain a new graph-level representation after intervention, which is then input into the loss function to optimize the parameters of the crack identification model.

2. The apparatus according to claim 1, characterized in that, The wall-climbing drone platform includes: The flight control module is used to control the at least one wall-climbing UAV to perform flight missions and attitude control actions; A data storage module is used to store the three-dimensional laser point cloud data and the image data.

3. The apparatus according to claim 1, characterized in that, The data fusion system includes: The data fusion module is used to fuse and match at least one of the following information: the width, location, internal orientation, and extension of the detected surrounding rock fissures, to obtain fused data. The sample database is used to crop and vectorize the fused data to the required size, generate a sample dataset according to the coordinate sequence, generate the sample database, and dynamically update the sample database in real time. The sample calibration module is used to calibrate the geological elements in the oblique photographic images and ground-penetrating radar profiles in the sample database to form a training set in the sample database.

4. A method for detecting fractures in surrounding rock based on causal feature learning, characterized in that, The surrounding rock fracture detection device based on causal feature learning as described in any one of claims 1-3 is used, wherein the method includes the following steps: According to the preset flight path, control at least one wall-climbing UAV to perform flight missions and attitude control actions in the target area to obtain three-dimensional laser point cloud data for detecting surrounding rock fissures; Collect image data of the detected rock fissures and establish a real-scene 3D model of the target detection area; Detect at least one of the following information: the width, location, internal orientation, and extension of the detected surrounding rock fissure; The three-dimensional laser point cloud data, the real-scene three-dimensional model, and the image data are fused and matched to generate a sample dataset. The sample dataset is calibrated and fractures are identified to generate a fracture identification model, which is then used to identify any type of fracture in the surrounding rock.

5. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, the processor executing the program to implement the rock fracture detection method based on causal feature learning as described in claim 4.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the rock fracture detection method based on causal feature learning as described in claim 4.

7. A computer program product, comprising a computer program, characterized in that, The computer program is executed to implement the rock fracture detection method based on causal feature learning as described in claim 4.

Images