Method and device for synchronous detection of river channel and fault

Abstract

The embodiment of the present invention provides a method and device for synchronous detection of river channels and faults, belonging to the technical field of seismic data interpretation, the method comprising: acquiring three-dimensional post-stack seismic data; processing the three-dimensional post-stack seismic data based on the minimum amplitude and the maximum amplitude in the three-dimensional post-stack seismic data; using the processed three-dimensional post-stack seismic data as the input of a river channel and fault detection model to obtain river channel and fault detection results. The method and device for synchronous detection of river channels and faults of the present invention have the advantages of short detection process time, high efficiency, accurate detection results, strong generalization ability, ability to adapt to the detection of different lithology river channel positions formed in different sedimentary environments, ability to eliminate the influence of faults on accurate detection of river channels, and realization of detection of complex river channels.

Description

Technical Field

[0001] The present invention relates to the field of seismic data interpretation, and in particular to a method for synchronously detecting a river channel and a fault, a device for synchronously detecting a river channel and a fault, an electronic device and a machine-readable storage medium. Background Art

[0002] The river channel is one of the important geological targets in oil and gas exploration. It is crucial to accurately detect the location of the river channel and reveal its development process. In terms of basin and prospect scale analysis, understanding the characteristics of the river channel is helpful for quantitative analysis of the landform and accurate understanding of the sedimentary process, and helps to understand the sedimentary environment and the main direction of sediments over time; in terms of reservoir scale analysis, it is crucial to determine the volume of the river channel sandstone and the properties of the fluid filled in it; in the oil and gas field development stage, accurately describing the macroscopic distribution characteristics of the river channel plane, sedimentation and its changing laws can provide important basis for the optimization of development target blocks, the preparation of oil and gas development plans, and the design of well locations and well trajectories.

[0003] In the existing technology, 3D post-stack seismic data is widely used for river channel interpretation. However, manual interpretation of river channels is time-consuming, the interpretation results are inaccurate and affected by human subjectivity; conventional convolutional neural networks are highly dependent on labels and label generation methods, resulting in poor generalization ability in other work areas, and only considering the influence of folds, ignoring the influence of faults, resulting in greatly reduced detection accuracy. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a method and device for synchronous detection of river channels and faults, so as to solve the problems of manual interpretation of river channels, which is time-consuming, inaccurate interpretation results and affected by human subjectivity; using conventional convolutional neural networks for river channel detection, which is highly dependent on labels and label generation methods, resulting in poor generalization ability in other work areas, and only considering the influence of folds and ignoring the influence of faults, resulting in low detection accuracy.

[0005] In order to achieve the above object, an embodiment of the present invention provides a method for synchronously detecting a river channel and a fault, comprising:

[0006] Acquire 3D post-stack seismic data;

[0007] Processing the three-dimensional post-stack seismic data based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data;

[0008] The processed three-dimensional post-stack seismic data are used as the input of the river channel and fault detection model to obtain the river channel and fault detection results.

[0009] Optionally, processing the three-dimensional post-stack seismic data based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data includes:

[0010] Determining a minimum amplitude value and a maximum amplitude value of the three-dimensional post-stack seismic data;

[0011] Determine a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value;

[0012] Transforming the amplitude value of each data in the three-dimensional post-stack seismic data based on the amplitude minimum value to obtain first transformed data;

[0013] Based on the normalized amplitude value and the first transformed data, processed three-dimensional post-stack seismic data are obtained.

[0014] Optionally, determining a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value includes:

[0015] The normalized amplitude value is calculated using the following formula:

[0016] ΔA=A max -A min ;

[0017] Among them, ΔA is the normalized amplitude value; A max is the maximum amplitude; A min is the minimum amplitude.

[0018] Optionally, transforming the amplitude value of each data in the three-dimensional post-stack seismic data based on the amplitude minimum value to obtain first transformed data includes:

[0019] The first transformation data is obtained using the following calculation formula:

[0020] A′ i =A i -A min ;

[0021] Among them, A′ i is the first transformation data; A i is the maximum amplitude; A min is the minimum amplitude.

[0022] Optionally, obtaining processed three-dimensional post-stack seismic data based on the normalized amplitude value and the first transformed data includes:

[0023] The processed three-dimensional post-stack seismic data is calculated using the following calculation formula:

[0024]

[0025] Among them, A″i is the processed 3D post-stack seismic data; A′ i is the first transformed data; ΔA is the normalized amplitude value.

[0026] Optionally, the method further includes:

[0027] Synthesize multiple sample data based on geological and geophysical theories and construct a training sample set;

[0028] The training sample set is used as the input of the deep learning neural network to train the river channel and fault detection model.

[0029] The present invention also provides a device for synchronously detecting a river channel and a fault, comprising:

[0030] A data acquisition module, used for acquiring three-dimensional post-stack seismic data;

[0031] A data processing module, used for processing the three-dimensional post-stack seismic data based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data;

[0032] The result output module is used to use the processed three-dimensional post-stack seismic data as the input of the river channel and fault detection model to obtain the river channel and fault detection results.

[0033] The data processing module specifically includes:

[0034] A data amplitude determination module, used to determine the minimum amplitude and the maximum amplitude of the three-dimensional post-stack seismic data;

[0035] A normalized amplitude value determination module, used to determine a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value;

[0036] An amplitude value conversion module, used for converting the amplitude value of each data in the three-dimensional post-stack seismic data based on the amplitude minimum value to obtain first converted data;

[0037] A data output module is used to obtain processed three-dimensional post-stack seismic data based on the normalized amplitude value and the first transformed data.

[0038] On the other hand, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for synchronous detection of river channels and faults when executing the computer program.

[0039] On the other hand, the present invention further provides a machine-readable storage medium, on which instructions are stored, and the instructions are used to enable a machine to execute the above-mentioned method for synchronous detection of river channels and faults.

[0040] The technical solution processes the three-dimensional post-stack seismic data by acquiring the minimum amplitude value and the maximum amplitude value of the three-dimensional post-stack seismic data to obtain processed three-dimensional post-stack seismic data, and inputs the processed three-dimensional post-stack seismic data into a pre-trained channel and fault detection model to obtain channel and fault detection results. The technical solution can adapt to the detection of the positions of channels with different lithologies formed in different sedimentary environments. In addition, it can eliminate the influence of faults on the accurate detection of channels, realize the detection of complex channels, and the detection process is short in time, high in efficiency, accurate in detection results, and strong in generalization ability.

[0041] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present invention, but do not constitute a limitation on the embodiments of the present invention. In the accompanying drawings:

[0043] Figure 1 It is a flow chart of the method for synchronous detection of river channel and fault provided by the present invention;

[0044] Figure 2 It is a flow chart of data processing in the method for synchronous detection of river channels and faults provided by the present invention;

[0045] Figure 3 It is a structural schematic diagram of a river channel and fault synchronous detection device provided by the present invention;

[0046] Figure 4 It is a structural schematic diagram of a data processing module in the river channel and fault synchronous detection device provided by the present invention;

[0047] Figure 5 is a schematic diagram of the first synthetic three-dimensional post-stack seismic data provided by the present invention;

[0048] Figure 6 The present invention provides Figure 5 Schematic diagram of 3D river channel data obtained by synthesizing 3D post-stack seismic data;

[0049] Figure 7 The present invention provides Figure 5 Schematic diagram of 3D fault data obtained by synthesizing 3D post-stack seismic data;

[0050] Figure 8 is a schematic diagram of a second type of synthetic three-dimensional post-stack seismic data provided by the present invention;

[0051] Fig. 9 The present invention provides Figure 8 Schematic diagram of 3D river channel data obtained by synthesizing 3D post-stack seismic data;

[0052] Fig.10 The present invention provides Figure 8 Schematic diagram of 3D fault data obtained by synthesizing 3D post-stack seismic data;

[0053] Fig.11 It is a schematic diagram of three-dimensional post-stack seismic data of a certain work area provided by the present invention;

[0054] Fig.12 The present invention provides Fig.11 Schematic diagram showing the superposition of 3D post-stack seismic data and corresponding river channel detection results;

[0055] Fig.13 The present invention provides Fig.11 Schematic diagram of the superposition of 3D post-stack seismic data and corresponding fault detection results.

[0056] Description of Reference Numerals

[0057] 10-data acquisition module; 20-data processing module;

[0058] 30-result output module; 21-data amplitude determination module;

[0059] 22-normalized amplitude value determination module; 23-amplitude value conversion module;

[0060] 24-Data output module. DETAILED DESCRIPTION

[0061] The specific implementation of the embodiment of the present invention is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0062] The terms “first”, “second”, “third”, etc. are only used for distinguishing descriptions and should not be understood as indicating or implying relative importance.

[0063] In addition, the terms "roughly" and "basically" are intended to explain that the relevant content does not require absolute precision, but can have certain deviations. For example, "roughly equal" does not only mean absolute equality. Since it is difficult to achieve absolute "equality" in actual production and operation, there are generally certain deviations. Therefore, in addition to absolute equality, "roughly equal" also includes the above-mentioned situation with certain deviations. Taking this as an example, in other cases, unless otherwise specified, the terms "roughly" and "basically" have similar meanings to the above.

[0064] Figure 1 It is a flow chart of the method for synchronous detection of river channel and fault provided by the present invention; Figure 3 It is a structural schematic diagram of a river channel and fault synchronous detection device provided by the present invention; Figure 4 It is a structural schematic diagram of a data processing module in the river channel and fault synchronous detection device provided by the present invention; Figure 5 is a schematic diagram of the first synthetic three-dimensional post-stack seismic data provided by the present invention; Figure 6 The present invention provides Figure 5 Schematic diagram of 3D river channel data obtained by synthesizing 3D post-stack seismic data; Figure 7 The present invention provides Figure 5 Schematic diagram of 3D fault data obtained by synthesizing 3D post-stack seismic data; Figure 8 is a schematic diagram of a second type of synthetic three-dimensional post-stack seismic data provided by the present invention; Fig. 9 The present invention provides Figure 8 Schematic diagram of 3D river channel data obtained by synthesizing 3D post-stack seismic data; Fig.10 The present invention provides Figure 8 Schematic diagram of 3D fault data obtained by synthesizing 3D post-stack seismic data; Fig.11 It is a schematic diagram of three-dimensional post-stack seismic data of a certain work area provided by the present invention; Fig.12 The present invention provides Fig.11 Schematic diagram showing the superposition of 3D post-stack seismic data and corresponding river channel detection results; Fig.13 The present invention provides Fig.11 Schematic diagram of the superposition of 3D post-stack seismic data and corresponding fault detection results.

[0065] Example 1

[0066] like Figure 1 As shown, an embodiment of the present invention provides a method for synchronous detection of a river channel and a fault, comprising:

[0067] Step 1: Acquire three-dimensional post-stack seismic data;

[0068] Specifically, the relevant steps of obtaining three-dimensional post-stack seismic data based on existing seismic data are well known to those skilled in the art and will not be described in detail here.

[0069] Step 2: processing the three-dimensional post-stack seismic data based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data;

[0070] Specifically, Figure 2 As shown, the specific steps of step 2 include:

[0071] Step 201, determining the minimum amplitude and the maximum amplitude of the three-dimensional post-stack seismic data;

[0072] Step 202: Determine a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value;

[0073] Step 203, transforming the amplitude value of each data in the three-dimensional post-stack seismic data based on the minimum amplitude value to obtain first transformed data, wherein the first transformed data is three-dimensional post-stack seismic data with a minimum amplitude value of 0;

[0074] Step 204: Obtain processed three-dimensional post-stack seismic data based on the normalized amplitude value and the first transformed data, and the amplitude of the processed three-dimensional post-stack seismic data ranges between 0 and 1.

[0075] Further, determining a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value includes:

[0076] The normalized amplitude value is calculated using the following formula:

[0077] ΔA=A max -A min ;

[0078] Among them, ΔA is the normalized amplitude value; A max is the maximum amplitude; A min is the minimum amplitude.

[0079] Further, the amplitude value of each data in the three-dimensional post-stack seismic data is transformed based on the amplitude minimum value to obtain first transformed data, including:

[0080] The first transformation data is obtained using the following calculation formula:

[0081] A′ i =A i -A min ;

[0082] Among them, A′ i is the first transformation data; A i is the maximum amplitude; A min is the minimum amplitude.

[0083] Further, based on the normalized amplitude value and the first transformed data, the processed three-dimensional post-stack seismic data is obtained, including:

[0084] The processed three-dimensional post-stack seismic data is calculated using the following calculation formula:

[0085]

[0086] Among them, A″ i is the processed 3D post-stack seismic data; A′i is the first transformed data; ΔA is the normalized amplitude value.

[0087] Therefore, the specific steps of step 2 can also be described as:

[0088] The minimum amplitude value and the maximum amplitude value of the acquired three-dimensional post-stack seismic data are subtracted from the amplitude of the three-dimensional post-stack seismic data and the minimum amplitude value to obtain the three-dimensional post-stack seismic data with the minimum amplitude value of 0 (i.e., the first transformed data); the maximum amplitude value is subtracted from the minimum amplitude value to obtain a normalized amplitude value; the three-dimensional post-stack seismic data with the minimum amplitude value of 0 (i.e., the first transformed data) is divided by the normalized amplitude value to obtain the processed three-dimensional post-stack seismic data, and the amplitude of the data is between 0 and 1, which is specifically expressed by the following calculation formula:

[0089]

[0090] Among them, A″ i A is the processed 3D post-stack seismic data; i is the maximum amplitude; A max is the maximum amplitude; A min is the minimum amplitude.

[0091] Step 3: Use the processed three-dimensional post-stack seismic data as input to the river channel and fault detection model to obtain the river channel and fault detection results.

[0092] Specifically, Fig.11 , Fig.12 and Fig.13 As shown in the figure, the data size is 700*1600*400, the spatial sampling interval is 12.5m, the time sampling interval is 2ms, the longitudinal coordinate is the longitudinal survey line number, the transverse coordinate is the contact survey line number, and the vertical coordinate is time. The above method is used for detection and the detection results are obtained, as shown in the figure. Fig.12 and Fig.13 As shown, all river channel locations (dark locations) and fault locations (dark locations) in the 3D post-stack seismic data with a complex tectonic background are accurately detected.

[0093] Furthermore, the method further comprises:

[0094] Synthesize multiple sample data based on geological and geophysical theories and construct a training sample set;

[0095] The training sample set is used as the input of the deep learning neural network to train the river channel and fault detection model.

[0096] Specifically, in order to ensure the generalization ability of the channel and fault detection model, the training sample set of the channel and fault detection model is also processed (normalized) so that each sample consists of synthetic three-dimensional post-stack seismic data with a value range between 0 and 1 and the corresponding accurately labeled three-dimensional channel data and accurately labeled three-dimensional fault data. Therefore, in order to ensure the consistency of the input of the pre-trained channel and fault detection model, the input three-dimensional post-stack seismic data is also normalized. Considering that manual annotation of three-dimensional seismic data is time-consuming, inaccurate and highly subjective, in order to overcome this problem, based on geological and geophysical theories, 9600 pairs of synthetic three-dimensional post-stack seismic data with diverse and complex structural patterns and corresponding accurately labeled three-dimensional channel data and accurately labeled three-dimensional fault data with a size of 96*96*96 are automatically generated to form a training sample set.

[0097] In the embodiment of the present invention, the channel and fault detection model has two task outputs, namely, the processed three-dimensional post-stack seismic data is input into the pre-trained channel and fault detection model for prediction, and the channel detection result can be output by one of the branches of the channel and fault detection model, and the fault detection result can be output by another branch of the network model. It should be noted that in the embodiment of the present invention, the number of convolution layers, the size of the convolution kernel, the step size, the activation function and other elements of the channel and fault detection model are carefully designed for feasibility and necessity according to the requirements of the present invention. The channel and fault detection model in the present invention sequentially constructs the encoder part, the decoder part and the output part according to the principle of deep learning:

[0098] The encoder part consists of five stages, which are used to extract features of different scales from the input 3D post-stack seismic data. Each stage contains one or two convolutional layers, and learns a residual function, which is to add the input of each stage to the output of the last convolutional layer of that stage. The introduction of the residual learning mechanism helps to overcome the performance degradation problem and improve the accuracy. In each stage, the convolution kernel size of the convolution layer is 3*3*3, and the pooling is achieved by the convolution operation with a convolution kernel size of 2*2*2 and a stride of 2. It can be noticed that the number of features doubles and the resolution is halved in each stage. After each convolution layer, batch normalization and ReLU (Rectified Linear Unit) activation functions are applied to improve the stability and nonlinear approximation ability of the network.

[0099] The decoder part consists of four stages, which are used to systematically aggregate multi-scale information. Similar to the encoder part, each stage also contains one or two convolutional layers, and the residual learning mechanism is also adopted. The size of the convolution kernel of each stage convolution layer is also 3*3*3, and upsampling is achieved through deconvolution operations with a convolution kernel size of 2*2*2 and a step size of 2. Through the jump layer connection, the features extracted from the encoder part are connected to the corresponding stage of the decoder part for spatial resolution compensation.

[0100] The output part consists of two branches that share the same decoded features, each branch consists of two convolutional layers with a kernel size of 3*3*3 and one convolutional layer with a kernel size of l*l*l, followed by batch normalization and sigmoid activation function. The first branch is used for the river channel detection task, and the second branch is used for the fault detection task.

[0101] During training, the normalized synthetic 3D post-stack seismic data were input into the channel and fault detection model in batches, and the batch size was set to 4. The adaptive moment estimation optimization algorithm was used to optimize the network, the learning rate was set to 0.0001, the total training cycle was set to 100, and one cycle traversed all samples in the training sample set.

[0102] More specifically, Figure 5 , Figure 6 and Figure 7 The longitudinal coordinate in is the longitudinal survey line number, the transverse coordinate is the connection survey line number, and the vertical coordinate is time. Figure 5 , Figure 6 and Figure 7 As shown, in the first type of synthetic three-dimensional post-stack seismic data generated by the present invention, which has undergone complex structural deformation (including a series of alternating fold deformation and fault deformation), the river channel position and the fault position are accurately marked; wherein, Figure 6 The middle river channel position is marked as 1, and other positions are marked as 0; Figure 7 The fault layer position is marked as 1, and other positions are marked as 0.

[0103] More specifically, Figure 8 , Fig. 9 and Fig.10 The longitudinal coordinate is the longitudinal survey line number, the transverse coordinate is the connection survey line number, and the vertical coordinate is time. Figure 8 , Fig. 9 and Fig.10 As shown, the fold and fault structural information in the second type of synthetic 3D post-stack seismic data generated by the present invention is rich and diverse, and can be highly close to the actual data, and the river channel location and fault location are accurately marked, among which, Fig. 9 The middle river channel position is marked as 1, and other positions are marked as 0; Fig.10 The fault layer position is marked as 1, and other positions are marked as 0.

[0104] Taking the first type of synthetic 3D post-stack seismic data and the second type of synthetic 3D post-stack seismic data as examples, accurate labeled samples are provided for the training of the river channel and fault detection model. The existence of a sufficient number of labeled data in the training sample set lays a solid big data foundation for the rapid and accurate implementation of the river channel and fault synchronous detection method based on deep learning of the present invention.

[0105] Example 2

[0106] like Figure 3 As shown, an embodiment of the present invention further provides a device for synchronous detection of a river channel and a fault, comprising:

[0107] A data acquisition module 10 is used to acquire three-dimensional post-stack seismic data;

[0108] A data processing module 20, for processing the three-dimensional post-stack seismic data based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data;

[0109] The result output module 30 is used to use the processed three-dimensional post-stack seismic data as the input of the river channel and fault detection model to obtain the river channel and fault detection results.

[0110] Furthermore, if Figure 4 As shown, the data processing module 20 specifically includes:

[0111] A data amplitude determination module 21 is used to determine the minimum amplitude and the maximum amplitude of the three-dimensional post-stack seismic data;

[0112] A normalized amplitude value determining module 22, configured to determine a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value;

[0113] An amplitude value conversion module 23, used for converting the amplitude value of each data in the three-dimensional post-stack seismic data based on the amplitude minimum value to obtain first conversion data;

[0114] The data output module 24 is used to obtain processed three-dimensional post-stack seismic data based on the normalized amplitude value and the first transformed data.

[0115] Example 3

[0116] An embodiment of the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned method for synchronous detection of river channels and faults when executing the computer program.

[0117] Example 4

[0118] An embodiment of the present invention provides a machine-readable storage medium, on which instructions are stored, and the instructions are used to enable a machine to execute the above-mentioned method for synchronous detection of river channels and faults.

[0119] Through the above technical solutions, the detection accuracy and efficiency of river channel and fault detection have been significantly improved:

[0120] 1. Since different forms of river channels are added to the model in the form of relative impedance in the process of constructing the dual-task accurate annotation dataset, the river channel and fault detection model can adapt to the detection of different lithological channel locations formed in different sedimentary environments.

[0121] 2. Since the number of different river channel forms simulated in the model is randomly set in the process of constructing a dual-task accurate annotation dataset, and the direction, width, length, thickness and relative impedance value of the river channel vary with the spatial position, the river channel and fault detection model can adapt to the detection of complex river channel locations such as those that are closely spaced or intertwined.

[0122] 3. Due to the various folds and fault deformations in the process of building the dual-task accurate annotation dataset, coupled with data augmentation and knowledge sharing between tasks, the river channel and fault detection model can simultaneously detect the river channel and faults in the complex structural background, thereby effectively eliminating the impact of faults on the accurate detection of the river channel;

[0123] 4. Since the present invention is based on a deep learning algorithm and performs calculations on a GPU, it has a high computing efficiency. For a 4.56GB 3D seismic data volume, it only takes 4.4 minutes using an NVIDIA V100 GPU to complete the detection of all river channels and faults in the entire 3D seismic data volume, while manual interpretation requires several days and the accuracy cannot be guaranteed.

[0124] The optional implementation modes of the embodiments of the present invention are described in detail above in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above implementation modes. Within the technical concept of the embodiments of the present invention, various simple modifications can be made to the technical scheme of the embodiments of the present invention, and these simple modifications all belong to the protection scope of the embodiments of the present invention.

[0125] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not further describe various possible combinations.

[0126] Those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a program, and the program is stored in a storage medium, including a number of instructions to enable a single-chip microcomputer, a chip or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0127] In addition, various implementation modes of the embodiments of the present invention may be arbitrarily combined, and as long as they do not violate the concept of the embodiments of the present invention, they should also be regarded as the contents disclosed by the embodiments of the present invention.

Claims

1. A method for synchronous detection of river channels and faults, characterized in that: include: Acquire 3D post-stack seismic data; Processing the three-dimensional post-stack seismic data based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data; The processed three-dimensional post-stack seismic data is used as the input of the channel and fault detection model to obtain the channel and fault detection results. The channel and fault detection model is obtained by training the training sample set as the input of the deep learning neural network. The training sample set is constructed by multiple sample data synthesized based on geological and geophysical theories. In the process of constructing the training sample set, channels of different forms are added to the model in the form of relative impedance. The direction, width, length, thickness and relative impedance value of the channel vary with the spatial position, and various folds and fault deformations, data augmentation and knowledge sharing between tasks are performed. The channel and fault detection model includes an encoder part, a decoder part and an output part. The encoder part consists of five The decoder part consists of four stages, each of which contains one or two convolutional layers, and introduces a residual learning mechanism. Batch normalization and ReLU activation function are applied after each convolutional layer, and the convolution kernel size of the convolutional layer is 3*3*3; the output part consists of two branches sharing the same decoding features, each branch consists of two convolutional layers with a convolution kernel size of 3*3*3 and one convolutional layer with a convolution kernel size of l*l*l, followed by batch normalization and sigmoid activation function. The first branch is used to output the river detection results, and the second branch is used to output the fault detection results. The three-dimensional post-stack seismic data is processed based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data, including: Determining a minimum amplitude value and a maximum amplitude value of the three-dimensional post-stack seismic data; Determine a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value; Transforming the amplitude value of each data in the three-dimensional post-stack seismic data based on the amplitude minimum value to obtain first transformed data; Based on the normalized amplitude value and the first transformed data, processed three-dimensional post-stack seismic data are obtained.

2. The method according to claim 1, characterized in that Determining a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value includes: The normalized amplitude value is calculated using the following formula: ΔA=A max -IN min ; Among them, ΔA is the normalized amplitude value; A max is the maximum amplitude; A min is the minimum amplitude.

3. The method according to claim 1, characterized in that The amplitude value of each data in the three-dimensional post-stack seismic data is transformed based on the amplitude minimum value to obtain first transformed data, including: The first transformation data is obtained using the following calculation formula: A i ′=A i -A min ; Among them, A i ' is the first transformation data; A i is the maximum amplitude; A min is the minimum amplitude.

4. The method according to claim 1, characterized in that: Based on the normalized amplitude value and the first transformed data, the processed three-dimensional post-stack seismic data is obtained, including: The processed three-dimensional post-stack seismic data is calculated using the following calculation formula: Among them, A i ″ is the processed 3D post-stack seismic data; A i ' is the first transformation data; ΔA is the normalized amplitude value.

5. A device for synchronous detection of river channels and faults, characterized in that: include: A data acquisition module, used for acquiring three-dimensional post-stack seismic data; A data processing module, used for processing the three-dimensional post-stack seismic data based on the minimum amplitude value and the maximum amplitude value in the three-dimensional post-stack seismic data; The result output module is used to use the processed three-dimensional post-stack seismic data as the input of the channel and fault detection model to obtain the channel and fault detection results; the channel and fault detection model is obtained by training the training sample set as the input of the deep learning neural network, and the training sample set is constructed by multiple sample data synthesized based on geological and geophysical theories. In the process of constructing the training sample set, channels of different forms are added to the model in the form of relative impedance. The direction, width, length, thickness and relative impedance value of the channel vary with the spatial position, and various folds and fault deformations, data augmentation and knowledge sharing between tasks are performed; the channel and fault detection model includes an encoder part, a decoder part and an output part, wherein the encoder The detector part consists of five stages, each of which contains one or two convolutional layers and introduces a residual learning mechanism. After each convolutional layer, batch normalization and ReLU activation function are applied, and the convolution kernel size of the convolution layer is 3*3*3; the decoder part consists of four stages, each of which contains one or two convolutional layers and introduces a residual learning mechanism. The convolution kernel size of the convolution layer is 3*3*3; the output part consists of two branches that share the same decoding features, each branch consists of two convolutional layers with a convolution kernel size of 3*3*3 and one convolutional layer with a convolution kernel size of l*l*l, followed by batch normalization and sigmoid activation function. The first branch is used to output the river detection results, and the second branch is used to output the fault detection results; Wherein, the data processing module specifically includes: A data amplitude determination module, used to determine the minimum amplitude and the maximum amplitude of the three-dimensional post-stack seismic data; A normalized amplitude value determination module, used to determine a normalized amplitude value based on the minimum amplitude value and the maximum amplitude value; An amplitude value conversion module, used for converting the amplitude value of each data in the three-dimensional post-stack seismic data based on the amplitude minimum value to obtain first converted data; A data output module is used to obtain processed three-dimensional post-stack seismic data based on the normalized amplitude value and the first transformed data.

6. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the method for synchronous detection of river channels and faults described in any one of claims 1-4 is implemented.

7. A machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the method for synchronous detection of a river channel and a fault according to any one of claims 1 to 4.

Images