Rock cuttings identification methods, systems, media, and devices combining deep learning and XRF

Abstract

This invention discloses a method, system, medium, and device for rock cuttings identification combining deep learning and XRF. XRF technology is used to determine the types and contents of major elements in rock cuttings. During feature extraction using a neural network, channel expansion and compression are performed on the data, and an improved SE module is embedded during channel modification to correct the feature map. This allows the proposed one-dimensional convolutional neural network to autonomously learn the importance of each channel in the feature map, resulting in a lithology identification model with better generalization. This invention enables rapid and accurate lithology identification, saves labor costs, and plays a crucial role in the entire logging process.

Description

Technical Field

[0001] This invention belongs to the field of lithology identification technology, specifically relating to a method, system, medium, and device for identifying rock cuttings that combines deep learning and XRF. Background Technology

[0002] As early as 1912, Laue confirmed the diffraction of X-rays in crystalline materials. In 1922, Harding first used X-rays for elemental analysis of minerals. X-ray spectroscopy analysis in my country originated in the late 1950s, a relatively late start, and the technology was initially immature. To keep pace with foreign research and promote the development of mineral element identification in my country, relevant instruments were introduced for research, leading to the establishment of XRF laboratories. Later, the combination of XRF with cuttings logging technology provided another feasible method for identifying cuttings during logging, further increasing the reliability of cuttings logging. The main methods for interpreting and evaluating the elemental content of cuttings include: spectral methods, curve methods, and numerical methods.

[0003] Among these methods, the atlas method can roughly determine the type of rock cuttings, but it cannot perform more detailed classification during the rock cuttings identification process. The curve method requires a thorough understanding of the current area, thus placing high demands on the selection and skills of the geologists. The numerical method analyzes the changes in the ratios between elements during crustal movement, using these ratios to measure the presence of certain lithologies, but this approach is rarely used in well logging production.

[0004] In recent years, deep learning technology has seen tremendous development and has been widely applied in various fields. Compared to many traditional algorithms, deep learning technology, by learning prior knowledge from massive amounts of training data, possesses stronger generalization capabilities and more complex parameterized representations, and does not require adjusting algorithm parameters to adapt to different application scenarios. For the classification problem of one-dimensional data formats such as rock fragment element content, most researchers have focused on traditional approaches, with few studies exploring deep learning. Considering that rock fragment element contents do not have temporal dependencies, conventional deep networks such as LSTM, which are suitable for time-series relationships, are not applicable. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to address the shortcomings of the prior art by providing a rock cuttings identification method, system, medium and device that combines deep learning and XRF. The model trained by the deep learning algorithm identifies rock cuttings based on their elemental content, which has better generalization and can achieve the effects of fast and accurate identification and cost saving.

[0006] The present invention adopts the following technical solution:

[0007] A method for identifying rock cuttings combining deep learning and XRF includes the following steps:

[0008] S1. Improve the SE module in SE-Net and embed the improved SE module into a one-dimensional convolutional neural network to construct and train a rock cuttings recognition model that combines deep learning and XRF, thereby obtaining a one-dimensional convolutional neural network lithology recognition model based on channel attention. ;

[0009] S2. Perform X-ray fluorescence spectroscopy analysis on the mixed rock fragment samples to obtain the content ratio of various elements in the rock fragment samples, and then input them into the one-dimensional convolutional neural network lithology identification model based on channel attention trained in step S1. In this process, lithological classification results are obtained.

[0010] Specifically, step S1 is as follows:

[0011] A backbone of a one-dimensional convolutional neural network is constructed using five one-dimensional convolutional modules. An improved SE module is embedded after each of the three middle one-dimensional convolutional modules to form a Conv-SE. Two fully connected layers and a Softmax layer are added after the fifth one-dimensional convolutional module to output probabilities for the predicted rock debris lithology. This completes the construction of a rock debris recognition model combining deep learning and XRF. .

[0012] Furthermore, each improved SE module includes an adaptive average pooling layer and two one-dimensional convolutional layers, with a linear rectified unit layer following the first one-dimensional convolutional layer and a sigmoid layer following the second one-dimensional convolutional layer.

[0013] Furthermore, the dimensionality reduction factor of the two one-dimensional convolutional layers is 16.

[0014] Furthermore, all five one-dimensional convolutional modules undergo channel modification, the channel modification process being as follows: .

[0015] Furthermore, each one-dimensional convolutional module contains a one-dimensional convolutional layer, a batch regularization layer, and a ReLU layer.

[0016] Specifically, in step S1, the following is adopted: Loss function for a cuttings recognition model combining deep learning and XRF Conduct training, The loss function is as follows:

[0017]

[0018] in, This is an adjustable focusing parameter. To balance the weighting factors, This represents the predicted probability value.

[0019] Secondly, embodiments of the present invention provide a rock cuttings identification system combining deep learning and XRF, comprising:

[0020] The network module improves upon the SE module in SE-Net and embeds the improved SE module into a one-dimensional convolutional neural network. This constructs and trains a rock cuttings identification model combining deep learning and XRF, resulting in a one-dimensional convolutional neural network lithology identification model based on channel attention. ;

[0021] The identification module performs X-ray fluorescence spectroscopy analysis on mixed rock fragment samples to obtain the content ratio of various elements in the rock fragment samples, and then inputs it into the one-dimensional convolutional neural network lithology identification model based on channel attention trained by the network module. In this process, lithological classification results are obtained.

[0022] Thirdly, a computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the steps of the above-described method for identifying rock cuttings combining deep learning and XRF.

[0023] Fourthly, embodiments of the present invention provide a computer-readable storage medium including a computer program, which, when executed by a processor, implements the steps of the above-described method for identifying rock cuttings combining deep learning and XRF.

[0024] Compared with the prior art, the present invention has at least the following beneficial effects:

[0025] This invention combines deep learning and XRF for rock debris identification. It utilizes deep learning to identify rock debris element content. To ensure the model focuses on significant rock debris element features while ignoring less important information, an improved SE module is embedded between the expanded and compressed channels. This ensures good coverage of various rock debris element contents, preventing the neglect of elements with low abundance. The proposed model can autonomously learn the importance of each channel in the feature map, assigning different weights to each channel after channel changes. This strengthens features relevant to the current task, enabling the network to acquire significant features for training. Simultaneously, it avoids introducing task-irrelevant nonlinear features during channel changes. Compared to existing methods, this invention significantly shortens sample identification time and improves accuracy by embedding the improved SE module into a one-dimensional convolutional neural network.

[0026] Furthermore, the data size of elemental content in individual rock fragments is Based on past experience, traditional machine learning methods such as Support Vector Machines, Nearest Neighbors, and GBDT gradient descent trees are typically chosen first. However, due to the simplicity of the data and the lack of obvious features, the accuracy of traditional machine learning methods is not high. Therefore, a one-dimensional convolutional neural network is used for rock debris lithology identification. Since the rock debris element content input to the model is single-channel data, it contains too few features. Therefore, channel expansion is used to increase features, followed by channel compression for feature fusion. If too many channels are expanded, too many useless features will be introduced, while too few will not introduce important features. Therefore, after experiments, five one-dimensional convolutional modules were selected as the backbone. After feature extraction by the five one-dimensional convolutional modules, the learned features are further learned through two fully connected layers. Here, too many fully connected layers may lead to overfitting; too few will result in insufficient learning and a decline in accuracy. Therefore, this invention selects two fully connected layers for learning. The meaning of the softmax layer is that it no longer uniquely determines a certain maximum value, but outputs the probability value of each classification result. The model of this invention uses this as the classification output of the rock debris identification model.

[0027] Furthermore, by embedding the improved SE module into a one-dimensional convolutional neural network, the sample recognition time was greatly shortened and the recognition accuracy was improved.

[0028] Furthermore, the value of the dimensionality reduction coefficient should be set according to the actual situation. It should not be too small, as this will lead to a significant increase in the amount of computation. If the dimensionality reduction coefficient is too large, it will lead to a large number of model parameters. Considering both the number of model parameters and the amount of computation, a dimensionality reduction coefficient of 16 is the optimal value.

[0029] Furthermore, during the channel modification process, excessive amplification at once will introduce a large number of invalid features. To ensure the effectiveness of the feature map, a gradual channel amplification approach is adopted, and the same principle applies during feature fusion. Therefore, this invention employs five one-dimensional convolutional modules to perform channel modification, and the channel modification process is as follows: .

[0030] Furthermore, this model significantly enhances the learning ability of the neural network by introducing convolutional layers and activation layers, leveraging the learning capabilities of convolutional layers and the filtering capabilities of activation layers. The batch regularization layer following each convolutional layer accelerates training and improves the model's generalization ability, while the activation layer uses a linear rectified unit (RTU) layer function. The RTU layer function removes neurons with values ​​less than 0, filtering out effective features and avoiding the gradient explosion problem.

[0031] Furthermore, due to the imbalance in the number of samples used for training, the following approach is adopted: If so, It balances the importance of positive and negative samples, and does not distinguish between easily separable and difficult-to-separate samples, allowing the model to focus on training difficult-to-separate negative samples, which can solve the problem of imbalanced sample numbers.

[0032] It is understood that the beneficial effects of the second to fourth aspects mentioned above can be found in the relevant descriptions in the first aspect mentioned above, and will not be repeated here.

[0033] In summary, this invention not only improves the accuracy of rock cuttings sample identification, but also can be quickly transferred to other similar one-dimensional data identification, demonstrating strong generalization ability and providing inspiration for other similar tasks.

[0034] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0035] Figure 1 This is a flowchart of the rock cuttings identification method combining deep learning and XRF of the present invention;

[0036] Figure 2 This is a flowchart of the Conv-SE module of the present invention;

[0037] Figure 3 This is a network structure diagram of the rock cuttings identification method combining deep learning and XRF of the present invention;

[0038] Figure 4 This is a diagram illustrating the channel change process of the rock cuttings identification method combining deep learning and XRF according to the present invention.

[0039] Figure 5 This is a schematic diagram of a computer device provided according to an embodiment of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] In the description of this invention, it should be understood that the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0042] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0043] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations. For example, A and / or B can represent three cases: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this document generally indicates that the preceding and following objects have an "or" relationship.

[0044] It should be understood that although terms such as first, second, third, etc., may be used in the embodiments of the present invention to describe the preset range, these preset ranges should not be limited to these terms. These terms are only used to distinguish the preset ranges from one another. For example, without departing from the scope of the embodiments of the present invention, the first preset range may also be referred to as the second preset range, and similarly, the second preset range may also be referred to as the first preset range.

[0045] Depending on the context, the word "if" as used here can be interpreted as "when," "when," "in response to determination," or "in response to detection." Similarly, depending on the context, the phrase "if determination" or "if detection (of the stated condition or event)" can be interpreted as "when determination," "in response to determination," "when detection (of the stated condition or event)," or "in response to detection (of the stated condition or event)."

[0046] The accompanying drawings illustrate various structural schematic diagrams according to embodiments disclosed in this invention. These drawings are not to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0047] This invention provides a rock cuttings identification method combining deep learning and XRF, focusing on the problem of lithological identification based on rock cuttings elemental content. Addressing the limitations of traditional machine learning methods, this chapter proposes a novel method combining deep learning and XRF for rock cuttings identification. This method utilizes channel expansion to increase the nonlinear features of the original data, followed by channel compression and feature fusion to extract features beneficial for lithological identification. More importantly, an innovative improved SE module is embedded after the one-dimensional convolutional module with channel modification, making the features extracted by the neural network more targeted. Experimental results demonstrate that this algorithm exhibits better accuracy than other methods mentioned.

[0048] Please see Figure 1 This invention discloses a method for identifying rock cuttings that combines deep learning and XRF, comprising the following steps:

[0049] S1. Improve the SE module in SE-Net (Squeeze-and-Excitation Networks) and embed the improved SE module into a one-dimensional convolutional neural network to construct a rock cuttings recognition model that combines deep learning and XRF. And training;

[0050] Please see Figure 2 Each one-dimensional convolutional module consists of a one-dimensional convolutional layer, a batch regularization layer, and a ReLU layer. The output of the linear rectified layer is connected to the input of the subsequent improved SE module. Each improved SE module is composed of an adaptive average pooling layer and two one-dimensional convolutional layers, where the first one-dimensional convolutional layer is followed by a linear rectified unit layer, and the second one-dimensional convolutional layer is followed by a sigmoid layer. Finally, a shortcut connection is used to weight the outputs of the one-dimensional convolutional modules and the sigmoid layers, achieving feature recalibration of the one-dimensional convolutional modules after changing the feature map channels. This allows for the improvement of useful features and the suppression of features that are not very useful for the current lithology identification task through autonomous learning.

[0051] Please see Figure 3 The backbone of a one-dimensional convolutional neural network is constructed using five one-dimensional convolutional modules (Conv). Each one-dimensional convolutional module contains a one-dimensional convolutional layer, a batch regularization layer, and a ReLU layer. The middle three one-dimensional convolutional modules are each followed by an improved SE module to form Conv-SE. Each improved SE module consists of an adaptive average pooling layer and two one-dimensional convolutional layers, with the first one-dimensional convolutional layer followed by a linear rectified unit layer and the second one-dimensional convolutional layer followed by a sigmoid layer. Two fully connected layers and a softmax layer are added after the fifth one-dimensional convolutional module to output probabilities for the predicted rock debris lithology, thus constructing a rock debris recognition model combining deep learning and XRF. .

[0052] Please see Figure 4 All five one-dimensional convolutional modules undergo channel modification, and the channel modification process is as follows: For a raw elemental data input of 1×10×1, a one-dimensional convolutional module conv1 with a kernel size of 3 is first used to expand the channels to obtain a feature map of size 1×8×16. Then, in the three cascaded conv-SE units, the kernel size of the one-dimensional convolutions is 2, and the kernel size of the two one-dimensional convolutional layers in the improved SE module of conv-SE is 1, with a dimensionality reduction coefficient of r=16. After three conv-SE units, the feature map changes from 1×8×16 to 1×5×32. Then, a one-dimensional convolutional module conv5 with a kernel size of 2 is used to compress the channels to 1×4×16. Finally, two fully connected layers are used to reduce the feature map size from 1×1×64 to 1×1×32 for subsequent classification probability calculations.

[0053] Specifically, the two one-dimensional convolutional layers in each improved SE module serve to reduce and increase dimensionality, respectively, with a dimensionality reduction coefficient of 16.

[0054] Rock debris element content data are fed into a rock debris recognition model that combines deep learning and XRF. Model training was performed; to address the significant differences in sample sizes, the loss function was modified to FocalLoss for the supervised lithology identification model. Training is performed; the FocalLoss formula is as follows:

[0055]

[0056] in, This is an adjustable focusing parameter. To balance the weighting factors, This represents the predicted probability value.

[0057] S2. Perform X-ray fluorescence spectroscopy (XRF) analysis on the mixed rock cuttings samples collected from logging to obtain the content percentage of various elements in the rock cuttings samples, and then input them into the one-dimensional convolutional neural network lithology identification model based on channel attention established in step S1. In this process, lithological classification results are obtained.

[0058] In another embodiment of the present invention, a rock cuttings identification system combining deep learning and XRF is provided. This system can be used to implement the above-mentioned rock cuttings identification method combining deep learning and XRF. Specifically, the rock cuttings identification system combining deep learning and XRF includes modules, modules, modules, modules, and modules.

[0059] The network module improves upon the SE module in SE-Net and embeds the improved SE module into a one-dimensional convolutional neural network to construct a rock cuttings recognition model that combines deep learning and XRF. And training;

[0060] The identification module performs X-ray fluorescence spectroscopy analysis on mixed rock fragment samples to obtain the content ratio of various elements in the rock fragment samples, and then inputs it into the one-dimensional convolutional neural network lithology identification model based on channel attention trained by the network module. In this process, lithological classification results are obtained.

[0061] In another embodiment of the present invention, a terminal device is provided, comprising a processor and a memory. The memory stores a computer program, which includes program instructions. The processor executes the program instructions stored in the computer storage medium. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. It is the computing and control core of the terminal, suitable for implementing one or more instructions, specifically suitable for loading and executing one or more instructions to achieve a corresponding method flow or corresponding function. The processor described in this embodiment can be used for the operation of a rock cuttings identification method combining deep learning and XRF, including:

[0062] The SE module in SE-Net is improved and embedded into a one-dimensional convolutional neural network to construct a rock cuttings recognition model that combines deep learning and XRF. The model is then trained; X-ray fluorescence spectroscopy analysis is performed on mixed rock fragment samples to obtain the content ratio of various elements in the rock fragment samples, and then input into the trained one-dimensional convolutional neural network lithology identification model based on channel attention. In this process, lithological classification results are obtained.

[0063] In another embodiment of the present invention, a storage medium is provided, specifically a computer-readable storage medium (Memory). This computer-readable storage medium is a memory device in a terminal device used to store programs and data. It is understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and extended storage media supported by the terminal device. The computer-readable storage medium provides storage space that stores the terminal's operating system. Furthermore, this storage space also stores one or more instructions suitable for loading and execution by a processor. These instructions can be one or more computer programs (including program code). It should be noted that the computer-readable storage medium here can be high-speed RAM or non-volatile memory, such as at least one disk storage device.

[0064] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the rock cuttings identification method combining deep learning and XRF in the above embodiments; one or more instructions in the computer-readable storage medium are loaded and executed by the processor in the following steps:

[0065] The SE module in SE-Net is improved and embedded into a one-dimensional convolutional neural network to construct a rock cuttings recognition model that combines deep learning and XRF. The model is then trained; X-ray fluorescence spectroscopy analysis is performed on mixed rock fragment samples to obtain the content ratio of various elements in the rock fragment samples, and then input into the trained one-dimensional convolutional neural network lithology identification model based on channel attention. In this process, lithological classification results are obtained.

[0066] Please see Figure 5 The computer device 60 in this embodiment includes a processor 61, a memory 62, and a computer program 63 stored in the memory 62 and executable on the processor 61. When the processor 61 executes the computer program 63, it implements the rock cuttings identification method combining deep learning and XRF as described in this embodiment. To avoid repetition, details are omitted here. Alternatively, when the processor 61 executes the computer program 63, it implements the functions of each model / unit in the rock cuttings identification system combining deep learning and XRF as described in this embodiment. To avoid repetition, details are omitted here.

[0067] Computer device 60 can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. Computer device 60 may include, but is not limited to, a processor 61 and a memory 62. Those skilled in the art will understand that... Figure 5This is merely an example of computer device 60 and does not constitute a limitation on computer device 60. It may include more or fewer components than shown, or combine certain components, or different components. For example, computer device may also include input / output devices, network access devices, buses, etc.

[0068] The processor 61 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0069] The memory 62 can be an internal storage unit of the computer device 60, such as a hard disk or RAM of the computer device 60. The memory 62 can also be an external storage device of the computer device 60, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc. equipped on the computer device 60.

[0070] Furthermore, the memory 62 may include both internal storage units of the computer device 60 and external storage devices. The memory 62 is used to store computer programs and other programs and data required by the computer device. The memory 62 can also be used to temporarily store data that has been output or will be output.

[0071] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0072] A total of 2934 sets of data from seven types of rock debris were used for model training, and 326 sets of test samples from seven types were used for testing. Three traditional machine learning methods were selected for comparative experiments. The specific experimental results are as follows:

[0073]

[0074] As can be seen from the table above, the overall accuracy rate proposed in this invention is 91.35%.

[0075] In summary, the present invention provides a rock cuttings identification method and system that combines deep learning and XRF. By using deep learning algorithms to train a model to identify rock cuttings elemental content data, it achieves better generalization, and can achieve the effects of rapid and accurate identification and cost savings.

[0076] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0077] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0078] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0079] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0080] The above content is only for illustrating the technical concept of the present invention and should not be construed as limiting the scope of protection of the present invention. Any modifications made to the technical solution based on the technical concept proposed in this invention shall fall within the scope of protection of the claims of this invention.

Claims

1. A method for cutting identification combining deep learning and XRF, characterized in that, Comprising the steps of: S1, improve the SE module in the SE-Net, embed the improved SE module in a one-dimensional convolutional neural network, construct a lithology identification model combining deep learning and XRF, and train to obtain a one-dimensional convolutional neural network lithology identification model based on channel attention ; specifically: The backbone of the one-dimensional convolutional neural network is built by using five one-dimensional convolution modules, an improved SE module (Conv-SE) is embedded after each of the three one-dimensional convolution modules in the middle, two fully connected layers and a Softmax layer are added after the fifth one-dimensional convolution module, which is used for probability output of the predicted lithic rock types, and a lithic rock identification model combining deep learning and XRF is constructed The channel change process is Each one-dimensional convolution module comprises a one-dimensional convolution layer, a batch normalization layer and a ReLU layer. Each improved SE module includes an adaptive average pooling layer and two one-dimensional convolution layers, a linear rectifier unit layer is connected after the first one-dimensional convolution layer, a Sigmoid layer is connected after the second one-dimensional convolution layer, and the dimension reduction coefficients of the two one-dimensional convolution layers are 16; Adopting Loss function for cutting recognition model combining deep learning and XRF Training, The loss function is specifically: wherein, is an adjustable focusing parameter, is a balancing weight factor, is a predicted probability value; S2, performing X-ray fluorescence spectrum analysis on the hybrid rock sample to obtain the content proportion of various elements in the rock sample, and then inputting the content proportion into the one-dimensional convolution neural network lithology identification model based on channel attention trained in step S1 In this way, the lithology classification result is obtained.

2. A system for cutting identification combining deep learning with XRF, characterized in that, Comprising: The network module improves the SE module in the SE-Net, embeds the improved SE module into a one-dimensional convolutional neural network, constructs a lithic identification model combining deep learning and XRF, and trains to obtain a one-dimensional convolutional neural network lithology identification model based on channel attention ; Specifically: The backbone of the one-dimensional convolutional neural network is built by using five one-dimensional convolution modules, an improved SE module (Conv-SE) is embedded after each of the three one-dimensional convolution modules in the middle, two fully connected layers and a Softmax layer are added after the fifth one-dimensional convolution module, which is used for probability output of the predicted lithic rock types, and a lithic rock identification model combining deep learning and XRF is constructed The channel change process is Each one-dimensional convolution module comprises a one-dimensional convolution layer, a batch normalization layer and a ReLU layer. Each improved SE module includes an adaptive average pooling layer and two one-dimensional convolution layers, a linear rectifier unit layer is connected after the first one-dimensional convolution layer, a Sigmoid layer is connected after the second one-dimensional convolution layer, and the dimension reduction coefficients of the two one-dimensional convolution layers are 16; Adopting Loss function for cutting recognition model combining deep learning and XRF Training, The loss function is specifically: wherein, is an adjustable focusing parameter, is a balancing weight factor, is a predicted probability value; The recognition module performs X-ray fluorescence spectrum analysis on the mixed rock sample to obtain the content proportion of various elements in the rock sample, and then inputs a one-dimensional convolutional neural network lithology recognition model based on channel attention trained by the network module The lithology classification result is obtained.

3. A computer-readable storage medium storing one or more programs, the one or more programs comprising instructions that when executed by a computer cause the computer to perform a method comprising: The one or more programs include instructions that, when executed by a computing device, cause the computing device to perform the method of claim 1.

4. A computing device, comprising: Comprising: One or more processors, memory, and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the one or more processors, the one or more programs include instructions for performing the method of claim 1.

Images