Intelligent fault identification method and system based on improved U-Net network

Abstract

The invention relates to an intelligent fault identification method and system based on an improved U-Net network. The method comprises the following steps: obtaining original seismic profile data and processing the original seismic profile data to obtain fault data; establishing an improved residual module and a channel attention mechanism module based on the fault data; constructing a U-Net model based on an improved residual module and the channel attention mechanism module; optimizing the U-Net model by adopting a weight cross entropy function, and training the optimized U-Net model based on the fault data to obtain a final model; and inputting the original seismic profile data into the final model to obtain a fault identification result. According to the method, the distinction degree of features in each stage is improved, and the depth of the network is increased; the consistency of fault prediction is enhanced, and the accuracy of fault identification is improved; the problem that fault data and seismic data are extremely unbalanced is relieved by using the weight loss function, and the accuracy of network fault identification is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of oil and gas exploration, and in particular relates to an intelligent fault identification method and system based on an improved U-Net network. Background Art

[0002] Faults play a vital role in oil and gas accumulation and development. As a crucial step in seismic data processing and interpretation, the rapid and effective identification of faults based on seismic data combined with existing geological information has been a key area of ​​research both domestically and internationally. In early seismic data interpretation, professionals manually interpreted faults by establishing fault identification markers based on reflection isotropic characteristics. Later, seismic data was combined with well logging data for joint well-seismic identification, as well as seismic attributes such as coherence volumes and curvature to identify faults. Today, researchers use machine algorithms such as ant algorithms, extreme gradient descent, edge detection, and support vector machines to characterize faults, and the accuracy and efficiency of fault identification continue to improve.

[0003] Automatic fault identification methods have achieved some success, but many problems remain. First, automatic fault identification methods based on seismic attributes often weaken the lateral resolution of seismic profiles to highlight fault characteristics, resulting in low accuracy. Second, automatic fault identification methods require extensive computation and emphasize local features, making them unsuitable for constructing faults from a holistic perspective. Furthermore, automatic fault identification methods are highly sensitive to noise, prone to misidentification of faults, requiring extensive corrections by professionals later. Finally, various improvements to automatic fault identification methods also face many new challenges, such as incorporating variance attributes into dip estimation, which still cannot eliminate the influence of strata.

[0004] Deep learning, as a key core technology in machine learning, has been widely applied in natural speech processing, computer vision, and other interdisciplinary fields. When applied to fault identification, deep learning methods continuously optimize network parameters and weights to extract hidden low- and high-dimensional features of faults, including features such as low continuity and high discontinuity in reflection events. This optimization of the network ultimately leads to accurate fault identification. Furthermore, deep learning itself requires a large amount of data to train the model and can process large amounts of data quickly and efficiently. Seismic exploration precisely meets the prerequisites for deep learning, and the exploration and development of oil and gas fields urgently require the powerful feature extraction capabilities of deep learning. For these reasons, many researchers have incorporated various deep learning theories and methodological models into fault identification.

[0005] In existing technologies, U-Net has been used to achieve excellent fault recognition results in response to the problems of manually picking faults or low accuracy in seismic attribute detection in complex seismic profiles. Using an encoding-decoding structure combined with a convolutional neural network, the pixels of the seismic profile image are predicted one by one to determine whether each pixel is a fault, and finally, good recognition results are obtained on real data. U-Net can not only be used directly for automatic fault recognition, but also as a high-performance network framework; the specific idea is to design corresponding extraction modules based on the actual fault characteristics, and build a corresponding model based on the U-Net framework to improve the accuracy of fault recognition. A multi-scale attention convolutional neural network is introduced into the U-Net's autoencoder structure to merge and refine the different spatial features of the fault. The network was tested using synthetic and field data. The test results show that it is more effective than conventional convolutional neural network methods in identifying faults in complex seismic profiles.

[0006] Although certain results have been achieved in fault identification based on the U-Net network, some problems still exist, such as gradient explosion and disappearance during gradient backpropagation, which leads to decreased network performance and poor accuracy and precision. Summary of the Invention

[0007] The purpose of the present invention is to provide an intelligent fault identification method and system based on an improved U-Net network in order to solve the above problems.

[0008] The present invention achieves the above-mentioned purpose through the following technical solutions:

[0009] An intelligent fault identification method based on an improved U-Net network includes the following steps:

[0010] Obtaining raw seismic profile data and processing it to obtain fault data;

[0011] Establishing an improved residual module and channel attention mechanism module based on the fault data;

[0012] Constructing a U-Net model based on the improved residual module and the channel attention mechanism module;

[0013] The U-Net model is optimized using a weighted cross entropy function, and the optimized U-Net model is trained based on the tomographic data to obtain a final model;

[0014] The original seismic profile data is input into the final model to obtain a fault identification result.

[0015] As a further optimization solution of the present invention, the fault data includes a seismic grayscale image and mask label data. The specific process of obtaining the original seismic profile data and processing it to obtain the fault data is as follows:

[0016] Convert the original seismic profile data containing faults into a grayscale image to obtain a seismic grayscale image;

[0017] Annotating faults based on the seismic grayscale image to generate mask label data;

[0018] The original seismic profiles and the corresponding mask label data are composed of training sets and test sets.

[0019] As a further optimization scheme of the present invention, the improved residual module consists of a dual channel consisting of one 1×1 convolution, two 3×3 convolutions, batch normalization, and an activation layer; wherein the convolution has a padding of 1 and a stride of 1; first, the seismic grayscale image is divided into two channels after 1×1 convolution, and the first channel is sequentially subjected to 3×3 convolution, batch normalization, ELU activation function, and 3×3 convolution, and then added to the result of 1×1 convolution; finally, the ELU activation function is used to output the feature tensor;

[0020] The formula of the ELU activation function is:

[0021]

[0022] Where α is the adjustment coefficient; y' is the feature output; The input of the feature.

[0023] As a further optimization scheme of the present invention, the channel attention mechanism module consists of two 3×3 convolutions, batch normalization, ELU activation function and sigmoid activation function; wherein the padding of the convolution is 1 and the step size is 1; first, the shallow features output by the improved residual module and the deep features output by the deconvolution module are fused, the fused features are activated by the ELU activation function after 3×3 convolution, the activated result is further convolved by 3×3, and then activated by the sigmoid activation function, and the activated result is multiplied with the shallow features; finally, the multiplied result is added to the deep features and output;

[0024] The formula of the sigmoid activation function is:

[0025]

[0026] Where x is the input of the feature; σ(x) is the output of the feature.

[0027] As a further optimization solution of the present invention, the specific process of constructing the U-Net model based on the improved residual module and the channel attention mechanism module is as follows:

[0028] Construct the input layer, hidden layer and output layer; the input layer is the input of the earthquake grayscale image; the output layer is the predicted probability distribution of the fault; the hidden layer is divided into two parts: upsampling and downsampling;

[0029] The downsampling includes 4 improved residual modules and 4 maximum pooling layers;

[0030] The upsampling consists of 4 deconvolution layers and 4 residual convolution modules;

[0031] Upsampling and downsampling are connected by four channel attention mechanism modules.

[0032] As a further optimization scheme of the present invention, the U-Net model is optimized using a weighted cross entropy function, and the optimized U-Net model is trained based on the tomographic data. The specific process of obtaining the final model is as follows:

[0033] The training set is input into the U-Net model, gradient back propagation is adopted, and weighted cross entropy is used as the loss function to measure the distance between the true value and the predicted value;

[0034] The formula for the weighted cross entropy is:

[0035]

[0036] Where α is the ratio of fault data in the entire data set; γ is the attention parameter; is the modulation coefficient;

[0037] Through each iteration of the data set, an adaptive algorithm is used to adjust the weights between neurons to reduce the loss function, thereby optimizing the network. The accuracy, recall rate, and F1 value are used to qualitatively judge the performance of the U-Net model, and the various parameters of the network model are initially saved. The iteration is repeated until the requirements for fault identification are met and the final model is saved.

[0038] The formulas for the precision rate Prec, the recall rate Rec and the F1 value are:

[0039]

[0040]

[0041]

[0042] Where TP is the number of pixels where faults are correctly detected; FP is the number of pixels where non-faults are predicted to be faults; and FN is the number of pixels where actual faults are not detected to be faults.

[0043] As a further optimization solution of the present invention, the fault identification result is a probability distribution map of the fault.

[0044] An intelligent fault identification system based on an improved U-Net network includes:

[0045] A data acquisition module is used to acquire original seismic profile data and process it to obtain fault data;

[0046] An improved module building module, configured to establish an improved residual module and a channel attention mechanism module based on the fault data;

[0047] A model construction module, configured to construct a U-Net model based on the improved residual module and the channel attention mechanism module;

[0048] A model optimization module is used to optimize the U-Net model using a weighted cross entropy function, and train the optimized U-Net model based on the fault data to obtain a final model;

[0049] The fault identification module is used to input the original seismic profile data into the final model to obtain a fault identification result.

[0050] An electronic device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;

[0051] Memory for storing computer programs;

[0052] The processor is used to implement an intelligent fault identification method based on an improved U-Net network when executing a program stored in the memory.

[0053] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements an intelligent fault identification method based on an improved U-Net network.

[0054] The beneficial effects of the present invention are:

[0055] The present invention replaces the convolutional layer of the U-Net network with an improved residual module, which solves the problem of gradient explosion and vanishing during gradient backpropagation, resulting in decreased network performance. At the same time, it refines features and integrates channel information, improves the discrimination of features at each stage, and increases the depth of the network. In upsampling, a channel attention mechanism is added to enhance the consistency of fault prediction and improve the accuracy of fault identification. Finally, a weighted loss function is used to alleviate the extreme imbalance between fault data and seismic data, further improving the accuracy of network fault identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 is a flow chart of the method of the present invention;

[0057] Figure 2 is a cropped schematic diagram of a seismic section in an embodiment of the present invention;

[0058] Figure 3 is a schematic diagram of mask label generation in an embodiment of the present invention;

[0059] Figure 4 is a schematic diagram of an improved residual module in an embodiment of the present invention;

[0060] Figure 5 Schematic diagram of the channel attention mechanism in an embodiment of the present invention;

[0061] Figure 6 Schematic diagram of the improved U-Net structure in an embodiment of the present invention;

[0062] Figure 7 is a flow chart in an embodiment of the present invention;

[0063] Figure 8 This is an improved U-Net fault recognition effect diagram in an embodiment of the present invention;

[0064] Figure 9 is a system structure block diagram in an embodiment of the present invention;

[0065] Figure 10 It is a block diagram of the device structure in an embodiment of the present invention. DETAILED DESCRIPTION

[0066] The present application is described in further detail below in conjunction with the accompanying drawings. It is necessary to point out that the following specific implementation methods are only used to further illustrate the present application and cannot be understood as limiting the scope of protection of the present application. Technicians in this field can make some non-essential improvements and adjustments to the present application based on the above application content.

[0067] like Figure 1 As shown, an intelligent fault identification method based on an improved U-Net network includes the following steps:

[0068] Obtaining raw seismic profile data and processing it to obtain fault data;

[0069] Establishing an improved residual module and channel attention mechanism module based on the fault data;

[0070] Constructing a U-Net model based on the improved residual module and the channel attention mechanism module;

[0071] The U-Net model is optimized using a weighted cross entropy function, and the optimized U-Net model is trained based on the tomographic data to obtain a final model;

[0072] The original seismic profile data is input into the final model to obtain a fault identification result.

[0073] In this embodiment, in order to efficiently identify faults in seismic profiles, the present invention adopts a combination of improved residual module and channel attention mechanism, such as Figure 4 and Figure 5 As shown in the figure, an improved U-Net fault recognition network model is built. Figure 7 , shows a flow chart of the present invention, comprising the following steps:

[0074] Step 1, fault input data preparation:

[0075] (1) Conversion and clipping of seismic grayscale profiles, such as Figure 2 As shown in Figure 1, the fault identification accuracy is 16×16, with a total of 500 sections. Each interpreted section contains a different number of seismic channels, a different number of sampling points per channel, and a sampling interval of 2ms. The amplitude data of the seismic sections were intercepted using the maximum rectangle. To further reduce the amount of calculation, the extracted seismic data was normalized and a single-channel grayscale image was generated, as shown in Figure 1. Figure 5 ;

[0076] (2) Generation of mask data, such as Figure 3 As shown in the figure, after cropping, each seismic profile is divided into several sub-profiles, and the entire dataset is expanded from 500 to 6,000. The input of the deep learning model becomes a single-channel grayscale value. The labelme annotation tool in the Python compilation language is used to annotate faults and generate layer mask labels. The size of the fault mask label data matrix is ​​the same as the seismic data size, where faults are marked as 1 and non-faults are marked as 0, as shown in the figure. Figure 7 ;

[0077] (3) The original seismic sub-sections and the corresponding fault mask data constitute the data set of the network model.

[0078] Step 2: Build an improved U-Net fault identification network.

[0079] (1) Construct an improved residual module.

[0080] This module consists of a dual-channel structure consisting of one 1×1 convolution, two 3×3 convolutions, batch normalization, and an activation layer. The convolutions have a padding of 1 and a stride of 1. First, the seismic grayscale image is divided into two channels after a 1×1 convolution. The first channel then undergoes a 3×3 convolution, batch normalization, an ELU activation function, and a 3×3 convolution. The result is then added to the 1×1 convolution. Finally, the ELU activation function is used to output the feature tensor.

[0081] The formula for the ELU activation function is:

[0082]

[0083] Where α is the adjustment coefficient; y' is the feature output; The input of the feature.

[0084] (2) Construct a channel attention mechanism.

[0085] This module consists of two 3×3 convolutions, batch normalization, ELU activation, and sigmoid activation. The convolution has a padding of 1 and a stride of 1. First, the shallow features output by the improved residual module are fused with the deep features output by the deconvolution module. The fused features are then subjected to a 3×3 convolution and activated using the ELU activation function. The activated results are then subjected to another 3×3 convolution and activated using the sigmoid activation function. The activated results are then multiplied with the shallow features. Finally, the multiplied results are added to the deep features and output.

[0086] The formula for the sigmoid activation function is:

[0087]

[0088] Where x is the input of the feature; σ(x) is the output of the feature.

[0089] Step 3: Use the designed residual module and channel attention mechanism module to establish the U-Net model.

[0090] Construct three network layers: input layer, hidden layer, and output layer. The input layer is the input of the earthquake grayscale image. The output layer is the predicted probability distribution of the fault. The hidden layer is divided into two parts: upsampling and downsampling. The downsampling contains 4 improved residual modules and 4 maximum pooling layers. The upsampling consists of 4 deconvolution layers and 4 residual convolution modules. The upsampling and downsampling are connected by 4 channel attention modules, such as Figure 6 .

[0091] Step 4: Input the training set preprocessed in step 1 into the network model built in step 3 through queuing, use gradient backpropagation, and use weighted cross entropy as the loss function to measure the distance between the true value and the predicted value.

[0092] The formula for weighted cross entropy is:

[0093]

[0094] Where α is the ratio of fault data in the entire dataset, which can be set by label data statistics; γ is the attention parameter; is the modulation coefficient, which reduces the weight of the background and makes the model focus more on the faults that are difficult to classify.

[0095] Step 5: Set the initial parameters of the model.

[0096] After the dataset and network model are built, the model parameters are initialized. The specific settings are:

[0097]

[0098] Step 6: Optimize the network model and save the model parameters.

[0099] Through each iteration of the data set, an adaptive algorithm is used to adjust the weights between neurons to reduce the loss function and thus optimize the network. The accuracy, recall rate, and F1 value are used to qualitatively judge the performance of the network model. The parameters of the network model are initially saved. The model is iterated repeatedly. If it does not meet the requirements for fault identification, continue training or return to step 5 to adjust the parameters and train the network. If it meets the requirements, stop iteration and save the final network model.

[0100] The formulas for precision, recall, and F1 value are:

[0101]

[0102]

[0103]

[0104] Where TP is the number of pixels where faults are correctly detected; FP is the number of pixels where non-faults are predicted to be faults; and FN is the number of pixels where actual faults are not detected to be faults.

[0105] Step 7: Use the optimized network to identify faults in the original earthquake profile and output a probability distribution map of the faults.

[0106] Figure 8 These are examples of fault recognition using the present invention, where (a) is the manual interpretation result; (b) is the fault recognition effect of FCN; (c) is the fault recognition effect of U-Net; (d) is the fault recognition effect of U-Net++; (e) is the fault recognition effect of Deeplap V3; and (f) is the fault recognition effect of the method of the present invention.

[0107] like Figure 9 As shown, an embodiment of the present disclosure provides an intelligent fault identification system based on an improved U-Net network, comprising:

[0108] A data acquisition module is used to acquire original seismic profile data and process it to obtain fault data;

[0109] An improved module building module, configured to establish an improved residual module and a channel attention mechanism module based on the fault data;

[0110] A model construction module, configured to construct a U-Net model based on the improved residual module and the channel attention mechanism module;

[0111] A model optimization module is used to optimize the U-Net model using a weighted cross entropy function, and train the optimized U-Net model based on the fault data to obtain a final model;

[0112] The fault identification module is used to input the original seismic profile data into the final model to obtain a fault identification result.

[0113] The implementation process of the functions and effects of each module in the above system is specifically described in the implementation process of the corresponding steps in the above method, which will not be repeated here.

[0114] For the system embodiment, since it basically corresponds to the method embodiment, the relevant parts can be referred to the partial description of the method embodiment. The system embodiment described above is only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, that is, they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of the present invention. Those of ordinary skill in the art can understand and implement it without paying any creative work.

[0115] In the above embodiment, any number of all modules can be combined into one module for implementation, or any one of the modules can be split into multiple modules. Alternatively, at least some of the functions of one or more of these modules can be combined with at least some of the functions of other modules and implemented in one module. At least one of all modules can be at least partially implemented as a hardware circuit, such as a field programmable gate array (FPGA), a programmable logic array (PLA), a system on a chip, a system on a substrate, a system on a package, an application specific integrated circuit (ASIC), or can be implemented by hardware or firmware such as any other reasonable way of integrating or packaging the circuit, or implemented in any one of the three implementation modes of software, hardware and firmware or in a suitable combination of any of them. Alternatively, at least one of all modules can be at least partially implemented as a computer program module, which can perform the corresponding function when the computer program module is run.

[0116] See also Figure 10The electronic device provided by an embodiment of the present disclosure includes a processor 1110, a communication interface 1120, a memory 1130 and a communication bus 1140, wherein the processor 1110, the communication interface 1120, and the memory 1130 communicate with each other through the communication bus 1140;

[0117] Memory 1130, for storing computer programs;

[0118] The processor 1110 is configured to implement the intelligent fault identification method based on the improved U-Net network when executing the program stored in the memory 1130 .

[0119] The communication bus 1140 can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. The communication bus 1140 can be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, the figure shows only one thick line, but this does not mean that there is only one bus or only one type of bus.

[0120] The communication interface 1120 is used for communication between the electronic device and other devices.

[0121] The memory 1130 may include a random access memory (RAM) or a non-volatile memory, such as at least one disk storage. Alternatively, the memory 1130 may be at least one storage device located away from the processor 1110.

[0122] The above-mentioned processor 1110 can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0123] The embodiments of the present disclosure further provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent fault identification method based on the improved U-Net network.

[0124] The computer-readable storage medium may be included in the device / apparatus described in the above embodiments, or may exist independently and not be incorporated into the device / apparatus. The computer-readable storage medium carries one or more programs that, when executed, implement the intelligent fault identification method based on the improved U-Net network according to the embodiments of the present disclosure.

[0125] According to an embodiment of the present disclosure, a computer-readable storage medium may be a non-volatile computer-readable storage medium, such as, but not limited to, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0126] The above-described embodiments merely illustrate several implementations of the present invention. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, and all such variations and improvements fall within the scope of protection of the present invention.

Claims

1. An intelligent fault identification method based on an improved U-Net network, characterized in that: The following steps are involved: Obtaining raw seismic profile data and processing it to obtain fault data; Establishing an improved residual module and channel attention mechanism module based on the fault data; Constructing a U-Net model based on the improved residual module and the channel attention mechanism module; The U-Net model is optimized using a weighted cross entropy function, and the optimized U-Net model is trained based on the tomographic data to obtain a final model; The original seismic profile data is input into the final model to obtain a fault identification result.

2. The intelligent fault identification method based on the improved U-Net network according to claim 1 is characterized in that: The fault data includes seismic grayscale images and mask label data. The specific process of obtaining the original seismic profile data and processing it to obtain the fault data is as follows: Convert the original seismic profile data containing faults into a grayscale image to obtain a seismic grayscale image; Annotating faults based on the seismic grayscale image to generate mask label data; The original seismic profiles and the corresponding mask label data are composed of training sets and test sets.

3. The intelligent fault identification method based on the improved U-Net network according to claim 2 is characterized in that: The improved residual module consists of a dual channel consisting of one 1×1 convolution, two 3×3 convolutions, batch normalization, and an activation layer. The convolution has a padding of 1 and a stride of 1. First, the seismic grayscale image is divided into two channels after 1×1 convolution. The first channel is sequentially subjected to 3×3 convolution, batch normalization, ELU activation function, and 3×3 convolution, and then added to the result of 1×1 convolution. Finally, the ELU activation function is used to output the feature tensor. The formula of the ELU activation function is: Where α is the adjustment coefficient; y' is the feature output; The input of the feature.

4. The intelligent fault identification method based on the improved U-Net network according to claim 3 is characterized in that: The channel attention mechanism module consists of two 3×3 convolutions, batch normalization, ELU activation function and sigmoid activation function; wherein the convolution has a padding of 1 and a stride of 1; first, the shallow features output by the improved residual module are fused with the deep features output by the deconvolution module, the fused features are activated by the ELU activation function after 3×3 convolution, the activated result is further 3×3 convolution, and then activated by the sigmoid activation function, and the activated result is multiplied with the shallow features; finally, the multiplied result is added to the deep features and output; The formula of the sigmoid activation function is: Where x is the input of the feature; σ(x) is the output of the feature.

5. The intelligent fault identification method based on the improved U-Net network according to claim 1 is characterized in that: The specific process of constructing the U-Net model based on the improved residual module and the channel attention mechanism module is as follows: Construct the input layer, hidden layer and output layer; the input layer is the input of the earthquake grayscale image; the output layer is the predicted probability distribution of the fault; the hidden layer is divided into two parts: upsampling and downsampling; The downsampling includes 4 improved residual modules and 4 maximum pooling layers; The upsampling consists of 4 deconvolution layers and 4 residual convolution modules; Upsampling and downsampling are connected by four channel attention mechanism modules.

6. The intelligent fault identification method based on the improved U-Net network according to claim 2 is characterized in that: The U-Net model is optimized using a weighted cross entropy function, and the optimized U-Net model is trained based on the tomographic data to obtain the final model. The specific process is as follows: The training set is input into the U-Net model, gradient back propagation is adopted, and weighted cross entropy is used as the loss function to measure the distance between the true value and the predicted value; The formula for the weighted cross entropy is: Where α is the ratio of fault data in the entire data set; γ is the attention parameter; is the modulation coefficient; Through each iteration of the data set, an adaptive algorithm is used to adjust the weights between neurons to reduce the loss function, thereby optimizing the network. The accuracy, recall rate, and F1 value are used to qualitatively judge the performance of the U-Net model, and the various parameters of the network model are initially saved. The iteration is repeated until the requirements for fault identification are met and the final model is saved. The formulas for the precision rate Prec, the recall rate Rec and the F1 value are: Where TP is the number of pixels where the fault is correctly detected; FP is the number of pixels that are not faults but are predicted to be faults; FN is the number of pixels that are actually faults but are not detected as faults.

7. The intelligent fault identification method based on the improved U-Net network according to claim 1 is characterized in that: The fault identification result is a probability distribution map of the fault.

8. An intelligent fault identification system based on an improved U-Net network, characterized in that: include: A data acquisition module is used to acquire original seismic profile data and process it to obtain fault data; An improved module building module, configured to establish an improved residual module and a channel attention mechanism module based on the fault data; A model construction module, configured to construct a U-Net model based on the improved residual module and the channel attention mechanism module; A model optimization module is used to optimize the U-Net model using a weighted cross entropy function, and train the optimized U-Net model based on the fault data to obtain a final model; The fault identification module is used to input the original seismic profile data into the final model to obtain a fault identification result.

9. An electronic device, characterized in that: The processor, the communication interface, the memory and the communication bus are connected to each other via the communication bus. Memory for storing computer programs; The processor is configured to implement the intelligent fault identification method based on the improved U-Net network described in any one of claims 1 to 10 when executing the program stored in the memory.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the intelligent fault identification method based on the improved U-Net network according to any one of claims 1 to 10 is implemented.

Images