Surrounding rock loose circle identification method and system, electronic equipment and storage medium

Abstract

The invention discloses a surrounding rock loose circle identification method and system, electronic equipment and a storage medium. The method comprises the following steps: constructing a surrounding rock loose circle identification network model through a preset sample data set; wherein the preset sample data set comprises surrounding rock loose circle sample data; inputting to-be-recognized image data into the surrounding rock loose circle recognition network model to perform surrounding rock loose circle feature recognition to obtain a surrounding rock loose circle image; wherein the to-be-recognized image data is obtained through shooting of a first drilling imager; and performing analysis and calculation according to the surrounding rock loose circle image and the moving speed of the first drilling imager to obtain surrounding rock loose circle information data. According to the embodiment of the invention, the method can achieve the recognition of the surrounding rock loose circle, and effectively improves the recognition accuracy of the surrounding rock loose circle. The method can be widely applied to the technical field of tunnel and underground space development.

Description

Technical Field

[0001] The present application relates to the technical field of tunnel and underground space development, and in particular to a method, system, electronic device and storage medium for identifying a loose zone of surrounding rock. Background Art

[0002] In tunnel engineering, the expansion force generated by the loose rock zone during its expansion is the primary factor causing tunnel damage. Identifying this zone is crucial for evaluating the overall stability of the tunnel and determining tunnel support methods and parameters. Related technologies use physical testing methods to determine the zone, but accurate identification is difficult. For example, acoustic testing is susceptible to interference and has low accuracy, while borehole imaging is complex and can be difficult to clearly identify when the surrounding rock quality is good.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The main purpose of the embodiments of the present application is to propose a method, system, electronic device and storage medium for identifying a loose zone of surrounding rock, which can realize the identification of the loose zone of surrounding rock and effectively improve the accuracy of the identification of the loose zone of surrounding rock.

[0005] To achieve the above-mentioned purpose, an embodiment of the present application provides a method for identifying a loose zone of surrounding rock, the method comprising the following steps:

[0006] Constructing a surrounding rock loosening zone identification network model through a preset sample data set; wherein the preset sample data set includes surrounding rock loosening zone sample data;

[0007] Inputting the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image; wherein the image data to be identified is obtained by photographing the first borehole imager;

[0008] The surrounding rock loosening zone information data is obtained by analyzing and calculating the surrounding rock loosening zone image and the moving speed of the first borehole imager.

[0009] In some embodiments, before executing the step of constructing a surrounding rock loosening zone identification network model using a preset sample data set, the method further includes:

[0010] acquiring tunnel boring video data by a second borehole imager;

[0011] Performing data preprocessing on the tunnel drilling video data to obtain the surrounding rock loosening zone sample data; wherein the data preprocessing includes video image extraction and video image labeling;

[0012] The surrounding rock loosening zone sample data is divided according to a preset ratio to construct the preset sample data set.

[0013] In some embodiments, constructing a surrounding rock loosening zone identification network model using a preset sample data set includes:

[0014] Build a fully convolutional neural network model;

[0015] Inputting the preset sample data set into the fully convolutional neural network model to predict a pixel prediction map;

[0016] Calculating error data between the pixel prediction image and the preset label image using a preset error function;

[0017] Performing reverse learning calculations based on the error data through a back propagation algorithm to obtain an optimal weight set for the model;

[0018] The parameters of the full convolutional neural network model are updated using the optimal weight set of the model to obtain the surrounding rock loosening zone identification network model.

[0019] In some embodiments, inputting the preset sample data set into the fully convolutional neural network model to predict a pixel prediction map includes:

[0020] Performing a preset operation on the preset sample data set through the fully convolutional neural network model to generate a feature spectrum; wherein the preset operation includes a convolution operation and a pooling operation;

[0021] The characteristic spectrum is restored by a deconvolution operation, and each pixel is predicted to obtain the pixel prediction map.

[0022] In some embodiments, inputting the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain the surrounding rock loosening zone image includes:

[0023] Inputting the image data to be identified into the surrounding rock loosening zone identification network model, so as to perform pixel prediction on the image data to be identified by the surrounding rock loosening zone identification network model to obtain surrounding rock loosening zone pixels;

[0024] The surrounding rock loosening zone image is obtained according to the pixel identification of the surrounding rock loosening zone.

[0025] In some embodiments, the analyzing and calculating the surrounding rock loosening zone image and the moving speed of the first borehole imager to obtain the surrounding rock loosening zone information data includes:

[0026] Determining a preset frame image according to the surrounding rock loosening zone image; wherein the preset frame image includes a start frame image and an end frame image;

[0027] Determining a preset passing time based on the start frame image and the end frame image by using preset video frame rate data; wherein the preset passing time includes the time when the first borehole imager passes through the surrounding rock loosening zone to be identified;

[0028] The thickness data of the surrounding rock loosening zone of the to-be-identified surrounding rock loosening zone is obtained by calculation according to the preset passing time and the moving speed.

[0029] In some embodiments, determining a preset frame image based on the surrounding rock loosening circle image includes:

[0030] Parsing the image data to be identified frame by frame, and taking the first video frame image corresponding to the first identified surrounding rock loosening zone image as the starting frame image;

[0031] The second video frame image corresponding to the last identified surrounding rock loosening circle image is used as the end frame image.

[0032] To achieve the above-mentioned purpose, another aspect of the present application provides a system for identifying a loosening zone of surrounding rock, the system comprising:

[0033] The first module is used to construct a surrounding rock loosening zone identification network model through a preset sample data set; wherein the preset sample data set includes surrounding rock loosening zone sample data;

[0034] The second module is used to input the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image; wherein the image data to be identified is obtained by photographing the first borehole imager;

[0035] The third module is used to analyze and calculate the surrounding rock loosening zone image and the moving speed of the first drilling imager to obtain surrounding rock loosening zone information data.

[0036] To achieve the above-mentioned object, another aspect of the present application provides an electronic device, comprising:

[0037] at least one processor;

[0038] at least one memory for storing at least one program;

[0039] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.

[0040] To achieve the above-mentioned purpose, another aspect of an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program implements the above-mentioned method when executed by a processor.

[0041] Embodiments of the present application include at least the following advantageous effects: The present application provides a method, system, electronic device, and storage medium for identifying loose rock zones. This solution constructs a network model for identifying loose rock zones using preset sample data. The preset sample data includes sample data of loose rock zones. Next, the embodiment of the present invention inputs image data to be identified into the network model for identifying loose rock zones, performing feature recognition of the loose rock zones and obtaining an image of the loose rock zones. Accordingly, the image data to be identified is captured by a first borehole imager. Finally, the embodiment of the present invention analyzes and calculates information data about the loose rock zones based on the image of the loose rock zones and the movement speed of the first borehole imager, thereby achieving identification of the loose rock zones. It is readily understood that the embodiment of the present invention effectively improves the accuracy of identifying loose rock zones by constructing a network model for identifying loose rock zones, performing feature recognition of the loose rock zones on the image data to be identified captured by the first borehole imager using the network model, and then calculating the information data based on the identified image of the loose rock zones and the movement speed of the first borehole imager. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 This is a flowchart of the steps of the method for identifying the loose zone of surrounding rock provided by an embodiment of the present invention;

[0043] Figure 2 1 is a schematic diagram of the process architecture of the surrounding rock loosening zone identification method provided by an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of tunnel drilling provided by an embodiment of the present invention;

[0045] Figure 4 2 is a schematic diagram comparing the false detection rates of models of different data sets provided by an embodiment of the present invention;

[0046] Figure 5 2 is a schematic diagram comparing the model false detection rates of different model algorithms provided by the embodiments of the invention;

[0047] Figure 6 Schematic diagram of the structure of the surrounding rock loosening zone identification system provided in an embodiment of the present application;

[0048] Figure 7 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are merely examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0050] It will be understood that the terms "first", "second", etc. used in this application may be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0051] The terms "at least one", "plurality", "each", "any", etc. used in this application include "at least one", "two" or more, "plurality" or "each", "any" or "any one", "each" or "any one" in the context of the present invention, and "at least one" or "at least one" includes one, two or more, "plurality" or "any one" includes two or more, "each" or "each one" in the context of the present invention, and "any" or "any one

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0053] Before explaining the embodiments of the present application in detail, some of the nouns and terms involved in the embodiments of the present application are first explained. The nouns and terms involved in the embodiments of the present application are subject to the following explanations.

[0054] The loosening zone of surrounding rock refers to the area where the surrounding rock stress equilibrium is disrupted due to roadway excavation, resulting in loosening and deformation. When a roadway is excavated, significant stress concentration forms around the roadway. Simultaneously, the surrounding rock stress state changes from three-dimensional compression to two-dimensional or even unidirectional compression, resulting in a decrease in surrounding rock strength. Accordingly, when the stress value exceeds the surrounding rock strength, a crushing zone of a certain thickness forms around the roadway, known as the loosening zone of surrounding rock.

[0055] In tunnel engineering, the expansion force generated by the loose rock zone during its expansion is the primary factor causing tunnel damage. Identifying this zone is crucial for evaluating the overall stability of the tunnel and determining tunnel support methods and parameters. Related technologies use physical testing methods to determine the zone, but accurate identification is difficult. For example, acoustic testing is susceptible to interference and has low accuracy, while borehole imaging is complex and can be difficult to clearly identify when the surrounding rock quality is good.

[0056] In view of this, an embodiment of the present application provides a surrounding rock loosening zone identification method, system, electronic device and storage medium, and the scheme constructs a surrounding rock loosening zone identification network model through preset sample data. Wherein, the preset sample data includes surrounding rock loosening zone sample data. Then, the embodiment of the present invention inputs the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image. Correspondingly, the image data to be identified is obtained by photographing the first borehole imager. Finally, the embodiment of the present invention performs analysis and calculation based on the surrounding rock loosening zone image and the moving speed of the first borehole imager to obtain surrounding rock loosening zone information data, thereby realizing the identification of the surrounding rock loosening zone and effectively improving the accuracy of surrounding rock loosening zone identification.

[0057] The surrounding rock loosening circle identification method provided in the embodiment of the present application relates to the technical field of tunnel and underground space development. The surrounding rock loosening circle identification method provided in the embodiment of the present application can be applied to a terminal, can be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements the surrounding rock loosening circle identification method, etc., but is not limited to the above forms.

[0058] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0059] Figure 1 This is an optional flow chart of the method for identifying the loose zone of surrounding rock provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S110 to S130.

[0060] Step S110: constructing a surrounding rock loosening zone identification network model using a preset sample data set, wherein the preset sample data set includes surrounding rock loosening zone sample data.

[0061] Step S120: Input the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image. The image data to be identified is obtained by photographing the first borehole imaging device.

[0062] Step S130 , analyzing and calculating the surrounding rock loosening zone image and the moving speed of the first borehole imager to obtain surrounding rock loosening zone information data.

[0063] During the working process of this specific embodiment, the embodiment of the present invention first constructs a surrounding rock loosening zone identification network model through preset sample data. Specifically, the preset sample data set in the embodiment of the present invention includes surrounding rock loosening zone sample data. Among them, the surrounding rock loosening sample data refers to the surrounding rock loosening zone data of the tunnel borehole obtained in advance, which is used to train the surrounding rock loosening zone identification network model. Accordingly, the embodiment of the present invention inputs the preset sample data set into the constructed initial neural network model for model training, and constructs a surrounding rock loosening zone identification network model. Next, the embodiment of the present invention inputs the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification, and obtains a surrounding rock loosening zone image. Specifically, the image data to be identified in the embodiment of the present invention is obtained by photographing the first borehole imager. Among them, the first borehole imager in the embodiment of the present invention is an in-hole detection instrument that integrates borehole video recording and panoramic wall image stitching functions, also known as a borehole television imager. It guides the camera into the borehole to record the surrounding rock of the borehole wall to form a color image. The embodiment of the present invention inputs the image to be identified, which is captured by the first borehole imager, into a trained surrounding rock loosening zone recognition network model to perform surrounding rock loosening zone feature recognition, so as to identify the surrounding rock loosening zone in the image to be identified and obtain a surrounding rock loosening zone image. Finally, the embodiment of the present invention performs analysis and calculation based on the surrounding rock loosening zone image and the moving speed of the first borehole imager to obtain surrounding rock loosening zone information data. Specifically, the surrounding rock loosening zone information data in the embodiment of the present invention refers to the relevant data of the surrounding rock loosening zone, such as the size and thickness of the surrounding rock loosening zone of each borehole. Correspondingly, the moving speed of the first borehole imager refers to the moving speed of the first borehole imager during the process of capturing the image data to be identified. The embodiment of the present invention calculates the ring width of the surrounding rock loosening zone by combining the identified surrounding rock loosening zone image and the moving speed of the first borehole imager during the image acquisition process, thereby determining the corresponding surrounding rock loosening zone information data.

[0064] In some embodiments of the present invention, before constructing a surrounding rock loosening zone identification network model using a preset sample data set, the surrounding rock loosening zone identification method provided by the embodiment of the present invention further includes but is not limited to the following steps:

[0065] The tunnel boring video data is acquired by the second borehole imager.

[0066] Data preprocessing is performed on the tunnel drilling video data to obtain sample data of the surrounding rock loosening zone. The data preprocessing includes video image extraction and video image labeling.

[0067] The sample data of the surrounding rock loosening zone are divided according to the preset ratio to construct a preset sample data set.

[0068] In this specific embodiment, the present invention first acquires tunnel drilling video data using a second borehole imager. Specifically, the present invention guides the second borehole imager into the borehole and records the surrounding rock of the borehole wall to obtain tunnel drilling video data. Next, the present invention performs data preprocessing on the borehole video data to obtain surrounding rock loosening zone sample data. Specifically, data preprocessing in the present embodiment includes video image extraction and video image labeling. For example, the present embodiment extracts and converts the acquired tunnel drilling video data at a preset frame rate, such as 20 frames per second. Based on the movement speed of the second borehole imager's borehole TV probe, frame images are captured within a preset range, such as 1-2 meters, from the borehole opening to construct an initial tunnel drilling image dataset, i.e., surrounding rock loosening zone sample data. Simultaneously, the present embodiment constructs a label image, i.e., a preset label image. In the present embodiment, each input original image, i.e., each original sample image in the surrounding rock loosening zone sample data, is classified by color, where different color regions correspond to different categories in the image. In the embodiment of the present invention, the label map is the standard answer for supervised learning, which is an annotated image corresponding to the input image. Each pixel is assigned a category label, indicating the category to which the pixel belongs. Accordingly, during the training process, the model will generate a prediction map, predict the category of each pixel, and compare it with the label map. By calculating the error between the two, the weight of the model is adjusted to gradually improve the accuracy of the prediction. Furthermore, the embodiment of the present invention divides the surrounding rock loosening zone sample data according to a preset ratio to construct a preset sample data set. Specifically, in order to avoid the influence of the difference in sample size on the recognition results, and to ensure that the number of samples participating in the training of photos in each borehole is roughly the same, the embodiment of the present invention divides the surrounding rock loosening zone sample data into a training set and a test set according to a preset division ratio, such as a ratio of 4:1 between the training set and the test set, and the data between the test set and the training set do not overlap. The model is trained by the training set, and the model performance is evaluated by the test set.

[0069] In some embodiments of the present invention, a network model for identifying surrounding rock loosening zones is constructed using a preset sample data set, including but not limited to the following steps:

[0070] Build a fully convolutional neural network model.

[0071] The preset sample data set is input into the fully convolutional neural network model to predict the pixel prediction map.

[0072] The error data between the pixel prediction image and the preset label image is calculated using a preset error function.

[0073] Based on the error data, reverse learning calculation is performed through the back propagation algorithm to obtain the optimal weight set of the model.

[0074] The parameters of the full convolutional neural network model are updated through the optimal weight set of the model to obtain the surrounding rock loosening zone identification network model.

[0075] In this specific embodiment, the embodiment of the present invention first constructs a fully convolutional neural network model and inputs a preset sample data set into the fully convolutional neural network model to predict a pixel prediction map. Specifically, the fully convolutional neural network model (FCN) in the embodiment of the present invention is a deep learning model based on the traditional convolutional neural network (CNN). By replacing the fully connected layer with a convolutional layer (or upsampling layer, global average pooling layer, etc.), the network can accept inputs of any size and output feature maps or segmentation results of corresponding sizes. Accordingly, the embodiment of the present invention inputs the original image and its corresponding label map into the fully convolutional neural network model to predict each pixel in the original image to obtain a pixel prediction map, such as each pixel in the original image is predicted to be a pixel in the surrounding rock loosening zone or a pixel in the non-surrounding rock loosening zone. Then, the embodiment of the present invention calculates the error data between the pixel prediction map and the preset label image through a preset error function. Specifically, the embodiment of the present invention calculates the error data between the pixel prediction map and the label map, that is, the preset label image, through the overall error function to complete one iteration of forward reasoning. Among them, the overall error function in the embodiment of the present invention is a mathematical function used to measure the difference between the model prediction result and the actual result, which can guide model learning and gradually improve the prediction accuracy of the model by continuously minimizing the error.

[0076] Then, the embodiment of the present invention performs reverse learning calculations based on the error data through the back propagation algorithm to obtain the optimal weight set of the model. Specifically, the back propagation algorithm in the embodiment of the present invention is a learning algorithm that calculates the gradient of the loss function with respect to the network parameters in order to update the parameters and thus minimize the loss function. The embodiment of the present invention minimizes the overall error function by the stochastic gradient descent method, and at the same time uses the back propagation algorithm to perform reverse learning calculations on the error gradient between the prediction map and the label map to complete one iteration of reverse learning. Accordingly, the embodiment of the present invention converges the learning curve of the full convolutional neural network model suitable for automatic identification of surrounding rock loosening zones through multiple iterative operations, and obtains the optimal weight set of the model. Among them, stochastic gradient descent is used in deep learning and machine learning to minimize the loss function of the model in order to update the parameters of the model. The stochastic gradient descent algorithm is a variant of gradient descent, which is suitable for large-scale data sets and can accelerate the training process and reduce computational overhead. Exemplarily, in the stochastic gradient workflow of the embodiment of the present invention, a sample or a small batch of samples is randomly selected, and the gradient of the loss function of the sample is calculated, and the model parameters are updated by the following formula (1):

[0077]

[0078] Where θ is the model parameter, η is the learning rate, is the gradient of the loss function. Accordingly, the convergence condition of the stochastic gradient descent in the embodiment of the present invention is θ * is the global optimal solution.

[0079] Finally, the embodiment of the present invention updates the parameters of the fully convolutional neural network model using the model's optimal weight set, resulting in a network model for identifying loose rock zones. Specifically, the goal of the machine learning model in the embodiment of the present invention is to find the optimal parameters by minimizing the loss function. Accordingly, gradient descent calculates the gradient of the loss function with respect to the model parameters and updates the parameters along the negative direction of the gradient, gradually approaching the minimum value. The model parameters are then updated using the determined optimal weight set to construct a network model for identifying loose rock zones.

[0080] In some embodiments of the present invention, a preset sample data set is input into a fully convolutional neural network model to predict a pixel prediction map, including but not limited to the following steps:

[0081] The fully convolutional neural network model performs preset operations on the preset sample data set to generate a feature spectrum. The preset operations include convolution and pooling operations.

[0082] The feature spectrum is restored through deconvolution operation, and each pixel is predicted to obtain a pixel prediction map.

[0083] In this specific embodiment, the embodiment of the present invention first performs a preset operation on the preset sample data set through a fully convolutional neural network model to generate a feature spectrum. Specifically, the preset operation processing in the embodiment of the present invention includes a convolution operation and a pooling operation. Accordingly, the embodiment of the present invention performs multiple convolution and pooling operations on the original image in the preset sample data set, thereby generating a corresponding feature spectrum while retaining the spatial information and semantic features of the image. Among them, the convolution operation in the embodiment of the present invention is used to extract local features in the image. For example, the embodiment of the present invention slides a convolution kernel or filter on the input image and performs a dot product operation to generate a new matrix, which is called a feature map. This process can extract information such as edges, textures, and shapes in the image. The convolution operation formula is shown in the following formula (2):

[0084] Z i,j,k =∑ m,n X i+m,j+n W m,n,k +b k (2)

[0085] Where X is the input feature map, X i,j Represents the pixel value of the input image at position (i, j), w represents the convolution kernel, W m,n,k represents the weight of the kth convolution kernel at position (m,n), b k Represents the bias term of the kth convolution kernel, Z i,j,k Represents the value of the output feature map at position (i, j) after the convolution operation.

[0086] In addition, in the embodiment of the present invention, pooling is a downsampling operation in a convolutional neural network, which is used to reduce the size of the feature map, thereby reducing the amount of calculation and parameters, while retaining important feature information. By reducing the spatial resolution of the feature map, it focuses on the most important features, alleviates the overfitting problem, and reduces the complexity of the model. Among them, the embodiment of the present invention adopts the maximum pooling operation, as shown in the following formula (3):

[0087]

[0088] Among them, X is the input feature map, Z i,j,k It is the output feature map after pooling, and the size of the pooling window is m·n.

[0089] Accordingly, in the present embodiment, a feature map is the intermediate output generated after the convolution and pooling operations, representing the features in the input image that were recognized by the convolution kernel. In the present embodiment, each convolution kernel generates a feature map that contains a response to a specific pattern in the input image, such as edges or shapes.

[0090] Next, the embodiment of the present invention restores the feature spectrum through a deconvolution operation, and predicts each pixel to obtain a pixel prediction map. Specifically, the embodiment of the present invention restores the generated feature spectrum through a deconvolution operation to restore it to the same size as the original image, and predicts each pixel to generate an output map with pixel-level prediction, i.e., a pixel prediction map. Among them, the deconvolution operation in the embodiment of the present invention is a method for upsampling operation in a convolutional neural network, which can restore the spatial resolution of the feature map from a lower level to a higher level. In the embodiment of the present invention, the image processing task deconvolution is used to restore the spatial size of the input image, so that the network can generate an output of the same size as the original input image. The deconvolution operation formula is shown in the following formula (4):

[0091]

[0092] Where X is the input feature map, s is the deconvolution stride, which indicates the step size of each deconvolution movement, V is the convolution kernel used for deconvolution, and Z is the output feature map after deconvolution.

[0093] In some embodiments of the present invention, inputting the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain the surrounding rock loosening zone image, including but not limited to the following steps:

[0094] The image data to be identified is input into the surrounding rock loosening zone identification network model, so that the surrounding rock loosening zone identification network model performs pixel prediction on the image data to be identified and obtains the surrounding rock loosening zone pixels.

[0095] The surrounding rock loosening zone image is obtained based on the pixel identification of the surrounding rock loosening zone.

[0096] In this specific embodiment, the embodiment of the present invention first inputs the image data to be identified into the surrounding rock loosening zone identification network model to perform pixel prediction to obtain the surrounding rock loosening zone pixels, and then determines the corresponding surrounding rock loosening zone image through the predicted surrounding rock loosening zone pixels. Specifically, the embodiment of the present invention first inputs the borehole image data collected by the first borehole imager, that is, the image data to be identified, into the trained full convolutional neural network model, that is, the surrounding rock loosening zone identification network model, and calls the weight set generated in the training model, that is, the model optimal weight set, and only performs forward reasoning to predict and judge each pixel of the input image to obtain the surrounding rock loosening zone pixels. Accordingly, in the pixel prediction process of the embodiment of the present invention, pixels belonging to the surrounding rock loosening zone will be predicted as foreground pixels with a grayscale value of 1, while pixels that do not belong to the surrounding rock loosening zone will be predicted as background pixels of 0. Through the difference in the image pixel prediction, the surrounding rock loosening zone is automatically identified and the surrounding rock loosening zone image is obtained.

[0097] In some embodiments of the present invention, analyzing and calculating the surrounding rock loosening zone image and the moving speed of the first borehole imager to obtain surrounding rock loosening zone information data includes but is not limited to the following steps:

[0098] The preset frame image is determined according to the surrounding rock loosening zone image, wherein the preset frame image includes a start frame image and an end frame image.

[0099] The preset passing time is determined based on the start frame image and the end frame image by using the preset video frame rate data, wherein the preset passing time includes the time when the first borehole imager passes through the surrounding rock loosening zone to be identified.

[0100] The thickness data of the surrounding rock loosening zone to be identified is calculated according to the preset passing time and moving speed.

[0101] In this specific embodiment, the present invention first determines preset frame images based on the images of the surrounding rock loosening zone. Specifically, the preset frame images in the present embodiment include a start frame image and an end frame image. In the present embodiment, the start frame image refers to the first frame image of the surrounding rock loosening zone captured by the first borehole imager when passing through the surrounding rock loosening zone, and the end frame image refers to the last frame image of the surrounding rock loosening zone captured by the first borehole imager when passing through the surrounding rock loosening zone. The present embodiment analyzes the identified surrounding rock loosening zone images to determine the corresponding start frame image and end frame image. Next, the present embodiment determines a preset transit time based on the start frame image and end frame image using preset video frame rate data. Specifically, the preset transit time in the present embodiment refers to the time it takes for the first borehole imager to pass through the surrounding rock loosening zone to be identified, that is, the time interval between the start frame image and the end frame image captured by the first borehole imager. Accordingly, the preset video frame rate data in the present embodiment refers to the video frame rate of the first borehole imager. The present embodiment calculates the preset transit time based on the number of frames between the start frame image and the end frame image and the preset video frame rate data. Finally, the embodiment of the present invention calculates the thickness data of the surrounding rock loosening circle of the surrounding rock loosening circle to be identified based on the preset passing time and moving speed. Specifically, the moving speed in the embodiment of the present invention refers to the moving speed of the borehole television probe, that is, the moving speed of the first borehole imager. The embodiment of the present invention calculates the thickness data of the gastritis loosening circle by multiplying the calculated preset passing time of the first borehole imager and the corresponding moving speed. Alternatively, the embodiment of the present invention first calculates the probe moving distance in the borehole corresponding to each frame of the image by the preset video frame rate data and the moving speed. Among them, the product of the moving speed of the probe and the video frame rate is expressed as the physical displacement of each frame of the image. Then, the embodiment of the present invention calculates the ring width of the surrounding rock loosening circle, that is, the thickness data of the surrounding rock loosening circle, by multiplying the physical displacement of each frame of the image with the total number of frames (consisting of the starting frame image and the ending frame image) of the first borehole imager passing through the surrounding rock loosening circle to be identified.

[0102] In some embodiments of the present invention, determining a preset frame image based on the surrounding rock loosening zone image includes but is not limited to the following steps:

[0103] The image data to be identified is parsed frame by frame, and the first video frame image corresponding to the first identified surrounding rock loosening circle image is used as the starting frame image.

[0104] The second video frame image corresponding to the last identified surrounding rock loosening circle image is used as the end frame image.

[0105] In this specific embodiment, the embodiment of the present invention first analyzes the image data to be identified frame by frame, so as to use the first video frame image corresponding to the first identified surrounding rock loosening circle image as the starting frame image, and the second video frame image corresponding to the last identified surrounding rock loosening circle image as the ending frame image. Specifically, the embodiment of the present invention inputs the data captured by the first drilling imager into the surrounding rock loosening circle identification network model for analysis. Then, the surrounding rock loosening circle identification network model parses the input image frame by frame. Accordingly, when characteristic pixels such as surrounding rock cracks are detected for the first time, that is, the first surrounding rock loosening circle image is identified, the frame, that is, the first video frame image, is marked as the starting frame image. At the same time, when it is determined that the characteristic pixels of the surrounding rock loosening circle can no longer be identified in the image frame after a certain frame, that is, when the last surrounding rock loosening circle image is identified, the frame (the second video frame image) is marked as the ending frame image.

[0106] The following describes the solution of the embodiment of the present invention in detail with reference to a specific scenario of identifying a loose zone of surrounding rock:

[0107] For example, referring to Figure 2 , Figure 2 This is a flow chart of the loose rock zone identification method provided by the embodiment of the present invention. Specifically, taking a certain project as an example, the embodiment of the present invention selected three sections for drilling according to different burial depths, with burial depths of 537m, 330m and 94m respectively. For each section, 8 holes were selected at equal intervals. The drilling distribution of the tunnel is as follows: Figure 3As shown. Accordingly, the embodiment of the present invention uses a multifunctional borehole imaging analyzer to collect video data from 3 sections and 24 boreholes. The collected videos are saved in H264 format with a bit rate of 3369Kbps and a resolution of 400 pixels (horizontal) × 288 pixels (vertical). Then, the embodiment of the present invention extracts and converts the acquired video data at a frame rate of 20 frames per second. The correspondence between the number of frames and the depth is determined based on the movement speed of the probe. Each frame corresponds to a movement distance of the camera of 10 cm / 20 = approximately 5 mm in the borehole. Among them, since the approximate range of the surrounding rock loosening zone is between 1-2 meters from the borehole mouth, the embodiment of the present invention selects images with a frame number between 200-400. Accordingly, after video analysis and image processing, the embodiment of the present invention extracts 200 photos from each borehole, resulting in a total of approximately 4800 (200 × 8 (number of boreholes) × 3 (number of sections)) borehole video images. To minimize the impact of sample size differences on recognition results and ensure that the number of photos participating in training is roughly the same for each borehole, the present embodiment uses a standard training set:test set ratio of 4:1. Therefore, the present embodiment randomly selects 160 photos from the video images of each borehole as training samples for that borehole, ultimately resulting in a total of 3,840 photos as the training sample for automatic recognition. The test set contains 960 images, and the training and validation data do not overlap. Furthermore, the present embodiment creates a label map, one for each input image, and classifies images by color, with different color regions corresponding to different categories. For example, the present embodiment marks loose zones as white, i.e., the color corresponding to the surrounding rock loose zones in the image is (R, G, B) = (256, 256, 256). The background image, without the surrounding rock loose zones, is set to black, i.e., the background color corresponding to the background is (R, G, B) = (0, 0, 0).

[0108] Next, the embodiment of the present invention constructs a fully convolutional neural network structure suitable for identifying loose zones in surrounding rocks through a related deep learning system. First, the embodiment of the present invention uses the original image and its corresponding label map as input to construct a sample data set for training, and inputs it into the fully convolutional neural network (FCN) model for subsequent training. Next, the embodiment of the present invention performs multiple convolution and pooling operations on the original image in the sample data set to generate a corresponding feature spectrum map, and retains the spatial information and semantic features of the image. Then, the embodiment of the present invention restores the generated feature spectrum map through a deconvolution operation to the same size as the original image, and predicts each pixel to generate an output map with pixel-level prediction. The error between the output prediction map and the label map is then calculated through the overall error function to complete one iteration of forward reasoning. Finally, the weights are updated, and the overall error function is minimized using the stochastic gradient descent method. At the same time, the error gradient between the prediction map and the label map is reversely learned and calculated through the back-propagation algorithm to complete one iteration of reverse learning. Among them, the embodiment of the present invention has a learning curve of the full convolutional neural network model suitable for automatic identification of surrounding rock loosening zones that tends to converge after 20,000 iterative operations. At this time, the optimal weight set of the model is obtained, and thus the network model for automatic identification of surrounding rock loosening zones is obtained through training.

[0109] Then, this embodiment of the present invention inputs the image data to be identified, captured by the borehole imager, into a trained FCN model. The weight set generated by the trained model is then applied, and forward inference is used to predict each pixel in the input image. Pixels belonging to the surrounding rock loosening zone are predicted as foreground pixels with a grayscale value of 1, while pixels not belonging to the surrounding rock loosening zone are predicted as background pixels with a grayscale value of 0. This difference in image pixel predictions enables automatic identification of surrounding rock loosening zones. Furthermore, this embodiment of the present invention inputs the video data captured by the borehole imager into a fully convolutional neural network model for analysis. The model analyzes the video frame by frame to determine the starting and ending frames. The number of frames between the starting and ending frames represents the time required for the borehole imager to complete the loosening zone. The thickness of the surrounding rock loosening zone is calculated based on the borehole imager's movement speed. For the three sections in this embodiment at burial depths of 95 meters, 150 meters, and 250 meters, the number of frames for the surrounding rock loosening zone is 320, 480, and 1000, respectively. At the same time, the embodiment of the present invention obtains video data at a frame rate of 20 frames per second, and determines the correspondence between the number of frames and the displacement of the probe by the moving speed of the probe. Each frame corresponds to the movement distance of the camera in the borehole of 10 cm / 20 = about 5 mm, and the obtained image information is converted and calculated according to the distance of each frame. Accordingly, the embodiment of the present invention calculates the thickness of the loose zone of the surrounding rock loosening circle of each borehole, and obtains that in section 1 with a burial depth of 95 meters, the average radius of the surrounding rock loosening circle is 0.8 meters; in section 2 with a burial depth of 150 meters, the average radius of the surrounding rock loosening circle is 1.2 meters; in section 3 with a burial depth of 250 meters, the average radius of the surrounding rock loosening circle is 2.5 meters. Among them, under the same rock conditions, the size of the surrounding rock loosening circle increases with the increase of the burial depth of the borehole section.

[0110] Accordingly, in the embodiment of the present invention, the image recognition analysis of the training set is performed, and the image defect false detection rate refers to the percentage of falsely detected pixels to the total number of pixels in the image, which is calculated by the following formula (5):

[0111]

[0112] Where w is the false detection rate, N is the number of falsely detected pixels, and N total is the total number of pixels in the image.

[0113] The embodiment of the present invention detects the accuracy of the images in the training set by recognizing and classifying the constructed FCN model. 100 images were randomly selected from the sample data set of the drilling for testing. The test results are: the number of falsely detected pixels is 20,000, and the false detection rate is 0.59%, indicating that the model has good recognition ability for the trained images and the model has good robustness. At the same time, the embodiment of the present invention performs recognition analysis on the test set image data that is not used in model training. By determining the false detection rate of the model, the generalization ability of the model can be verified, that is, whether the training model can achieve good recognition effect for images that have not participated in the training. The results show that the false detection rate of the FCN model constructed by the embodiment of the present invention is 1.2%, indicating that the model has good generalization and can be applied to specific engineering practices. Among them, Figure 4 The comparison between the training false positive rate and the test false positive rate of the model shows that the model has good robustness and generalization.

[0114] In addition, in order to verify the superiority of the fully convolutional neural network model proposed in the embodiment of the present invention, the embodiment of the present invention compared and analyzed three methods commonly used in image recognition: region growing method, great law method, and watershed method. The four different prediction algorithms were compared from two main aspects: false detection rate and running time. The specific comparison results are shown as follows: Figure 5 : The full convolutional neural network model has the lowest false positive rate of 1.2%; the second is the region growing method, which is 2.3%; the third is the big law method, which is 3.4%; and the worst is the watershed method, which is 5.1%. In terms of running time, the four methods are all operated in the same computer environment. From the results of the running time, it can be seen that the fastest algorithm is the big law method, followed by the full convolutional neural network model, and the worst is the region growing method. Therefore, considering the comprehensive false positive rate and running time, the surrounding rock loosening zone identification method provided by the embodiment of the present invention has the characteristics of low false positive rate and high operating efficiency, can realize the identification of surrounding rock loosening zone, and effectively improve the accuracy of surrounding rock loosening zone identification.

[0115] See also Figure 6 The embodiment of the present application further provides a system for identifying a loose zone of surrounding rock, which can implement the above-mentioned loose zone identification method of surrounding rock. The system includes:

[0116] The first module 210 is configured to construct a surrounding rock loosening zone identification network model using a preset sample data set, wherein the preset sample data set includes surrounding rock loosening zone sample data.

[0117] The second module 220 is configured to input the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image. The image data to be identified is obtained by photographing the first borehole imaging device.

[0118] The third module 230 is used to analyze and calculate the surrounding rock loosening zone image and the moving speed of the first borehole imager to obtain surrounding rock loosening zone information data.

[0119] It can be understood that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0120] The present application also provides an electronic device comprising a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method for identifying a loose zone of surrounding rock. The electronic device can be any smart terminal, including a tablet computer and an in-vehicle computer.

[0121] It can be understood that the contents of the above method embodiments are applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0122] See also Figure 7 , Figure 7 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0123] The processor 310 may be implemented as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is configured to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.

[0124] The memory 320 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 320 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 320 and is called by the processor 310 to execute the surrounding rock loosening zone identification method of the embodiments of this application.

[0125] Input / output interface 330, used to implement information input and output;

[0126] Communication interface 340, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WiFi, Bluetooth, etc.);

[0127] bus 350 , which transmits information between the various components of the device (e.g., processor 310 , memory 320 , input / output interface 330 , and communication interface 340 );

[0128] The processor 310 , the memory 320 , the input / output interface 330 and the communication interface 340 are connected to each other in communication within the device via the bus 350 .

[0129] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned surrounding rock loosening zone identification method is implemented.

[0130] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiment, the functions specifically implemented by the present storage medium embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0131] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0132] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0133] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0134] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0135] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0136] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0137] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0138] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the above units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0139] The units described above as separate components may or may not be physically separate, and 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 these units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0140] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0141] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store programs.

[0142] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A method for identifying a loose zone of surrounding rock, characterized in that: The method comprises the following steps: Constructing a surrounding rock loosening zone identification network model through a preset sample data set; wherein the preset sample data set includes surrounding rock loosening zone sample data; Inputting the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image; wherein the image data to be identified is obtained by photographing the first borehole imager; The surrounding rock loosening zone information data is obtained by analyzing and calculating according to the surrounding rock loosening zone image and the moving speed of the first borehole imager.

2. The method according to claim 1, characterized in that Before executing the step of constructing a surrounding rock loosening zone identification network model using a preset sample data set, the method further includes: Acquiring tunnel boring video data by a second borehole imager; Performing data preprocessing on the tunnel drilling video data to obtain the surrounding rock loosening zone sample data; wherein the data preprocessing includes video image extraction and video image marking; The surrounding rock loosening zone sample data is divided according to a preset ratio to construct the preset sample data set.

3. The method according to claim 1, characterized in that The method of constructing a surrounding rock loosening zone identification network model by using a preset sample data set includes: Build a fully convolutional neural network model; Inputting the preset sample data set into the fully convolutional neural network model to predict a pixel prediction map; Calculating error data between the pixel prediction image and a preset label image using a preset error function; Performing reverse learning calculations based on the error data through a back propagation algorithm to obtain an optimal weight set for the model; The parameters of the full convolutional neural network model are updated through the optimal weight set of the model to obtain the surrounding rock loosening zone identification network model.

4. The method according to claim 3, characterized in that: The step of inputting the preset sample data set into the fully convolutional neural network model to predict a pixel prediction map includes: Performing preset operation processing on the preset sample data set through the fully convolutional neural network model to generate a feature spectrum; wherein the preset operation processing includes convolution operation and pooling operation; The characteristic spectrum is restored through a deconvolution operation, and each pixel is predicted to obtain the pixel prediction map.

5. The method according to claim 1, characterized in that The step of inputting the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image includes: Inputting the image data to be identified into the surrounding rock loosening zone identification network model, so as to perform pixel prediction on the image data to be identified through the surrounding rock loosening zone identification network model to obtain surrounding rock loosening zone pixels; The surrounding rock loosening zone image is obtained according to the surrounding rock loosening zone pixel identification.

6. The method according to claim 1, characterized in that The analyzing and calculating according to the surrounding rock loosening zone image and the moving speed of the first borehole imager to obtain surrounding rock loosening zone information data includes: Determine a preset frame image according to the surrounding rock loosening circle image; wherein the preset frame image includes a start frame image and an end frame image; Determine a preset passing time according to the start frame image and the end frame image by using preset video frame rate data; wherein the preset passing time includes the time when the first borehole imager passes through the surrounding rock loosening zone to be identified; The thickness data of the surrounding rock loose zone of the to-be-identified surrounding rock loose zone is obtained by calculation according to the preset passing time and the moving speed.

7. The method according to claim 6, characterized in that The step of determining a preset frame image according to the surrounding rock loosening zone image comprises: Parsing the image data to be identified frame by frame, and taking the first video frame image corresponding to the first identified surrounding rock loosening circle image as the starting frame image; The second video frame image corresponding to the last identified surrounding rock loosening circle image is used as the end frame image.

8. A system for identifying loosened surrounding rock zones, characterized in that: The system comprises: The first module is used to construct a surrounding rock loosening zone identification network model through a preset sample data set; wherein the preset sample data set includes surrounding rock loosening zone sample data; The second module is used to input the image data to be identified into the surrounding rock loosening zone identification network model to perform surrounding rock loosening zone feature identification to obtain a surrounding rock loosening zone image; wherein the image data to be identified is obtained by photographing the first borehole imager; The third module is used to analyze and calculate according to the surrounding rock loosening zone image and the moving speed of the first drilling imager to obtain surrounding rock loosening zone information data.

9. An electronic device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images