A lithology identification model training method, a lithology identification method and related devices

Abstract

The application discloses a lithology identification model training method, a lithology identification method and related devices, and the method comprises the following steps: clustering training well logging data to obtain clustering clusters of various lithology samples; and oversampling samples to be supplemented corresponding to the oversampling clusters, so that the number of lithology samples of all categories in the training well logging data is balanced. The well logging data with balanced samples is used to train a lithology identification model, the lithology identification model does not produce prediction deviation, and identification is accurate.

Description

Technical Field

[0001] This invention relates to a lithology identification model training method, a lithology identification method, and related devices, belonging to the field of geological exploration. Background Technology

[0002] Lithology identification plays an important role in reservoir evaluation. With the widespread application of machine learning methods, intelligent lithology identification is receiving increasing attention from geologists. In the process of intelligent lithology identification, the quality of the identification model often determines the identification effect. Due to the cyclic nature of formation sedimentation, the lithology labels of well logging data are usually imbalanced between classes. Models trained with well logging data that are imbalanced between classes will produce prediction bias, leading to inaccurate identification. For example, if there are more sandstone samples and fewer limestone samples in the well logging data, the accuracy of the trained model in identifying limestone will be low. Summary of the Invention

[0003] This invention provides a lithology identification model training method, a lithology identification method, and a related device, which solves the problems disclosed in the background art.

[0004] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0005] A method for training a lithology identification model includes:

[0006] Obtain the number of various lithological samples in the training logging data;

[0007] The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are used as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0008] Clustering of training logging data yields clusters of lithological samples;

[0009] For each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The samples that need to be supplemented corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the training logging data until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the training logging data is not greater than the threshold A.

[0010] The lithology identification model was trained using training well logging data with new samples added.

[0011] Clustering of training logging data yields clusters of lithological samples, including:

[0012] K-means clustering was used to cluster the training logging data to obtain clusters of lithological samples.

[0013] Oversampling is performed on the samples that need to be supplemented corresponding to the oversampled clusters, including:

[0014] The SMOTE method is used to oversample the samples that need to be supplemented corresponding to the oversampled clusters.

[0015] The lithology identification model is a cascaded forest.

[0016] A training device for a lithology identification model, comprising:

[0017] The training sample quantity module retrieves the quantity of various lithological samples from the training logging data used for training.

[0018] The training requires supplementing the sample module. The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are considered as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0019] The clustering module is trained to cluster the training well logging data to obtain clusters of lithological samples;

[0020] The training oversampling module, for each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The supplementary samples corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the training logging data until the difference between the number of standard lithological samples and the number of supplementary samples is not greater than the threshold A.

[0021] The training module uses training well logging data with added new samples to train the lithology identification model.

[0022] A lithology identification method, comprising:

[0023] Obtain the number of various lithological samples in the well logging data to be identified;

[0024] The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are used as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0025] Clustering is performed on the well logging data to be identified to obtain clusters of lithological samples;

[0026] For each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The samples that need to be supplemented corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the logging data to be identified until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the logging data to be identified is not greater than the threshold A.

[0027] The well logging data to be identified after adding new samples is input into a pre-trained lithology identification model to obtain lithology identification results; wherein, the trained lithology identification model is trained using the method described in any one of claims 1 to 4.

[0028] A lithology identification system, comprising:

[0029] The sample quantity identification module obtains the quantity of various lithological samples in the well logging data to be identified.

[0030] The module for identifying samples that need to be supplemented uses the lithological sample with the largest number as the standard lithological sample and the other lithological samples as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are identified as samples that need to be supplemented. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0031] The clustering module identifies and clusters the logging data to be identified, obtaining clusters of lithological samples.

[0032] The oversampling identification module identifies each cluster as an oversampling cluster if the proportion of the most numerous lithological samples in the cluster is greater than threshold B and the most numerous lithological samples are samples that need to be supplemented. The cluster is then oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the logging data to be identified until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the logging data to be identified is no greater than threshold A.

[0033] The identification module inputs the well logging data to be identified after adding new samples into a pre-trained lithology identification model to obtain lithology identification results; wherein the trained lithology identification model is trained using the method described in any one of claims 1 to 4.

[0034] A computer-readable storage medium storing one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform a lithology identification model training method or lithology identification.

[0035] A computer device includes one or more processors and one or more memories, wherein one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, the one or more programs including instructions for performing a lithology identification model training method or lithology identification.

[0036] The beneficial effects achieved by this invention are as follows: This invention clusters training logging data to obtain clusters of various lithological samples, and oversamples the samples that need to be supplemented corresponding to the oversampled clusters, so that the number of lithological samples of all categories in the training logging data is balanced. The lithological identification model is trained using logging data with balanced samples, so that the lithological identification model does not produce prediction bias and the identification is accurate. Attached Figure Description

[0037] Figure 1 This is a flowchart of the lithology identification model training method;

[0038] Figure 2 This is a schematic diagram of clustering.

[0039] Figure 3 This is a schematic diagram of a cascading forest.

[0040] Figure 4 This is a flowchart of a lithology identification method. Detailed Implementation

[0041] The present invention will be further described below with reference to the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solution of the present invention, and should not be used to limit the scope of protection of the present invention.

[0042] like Figure 1 As shown, a lithology identification model training method includes the following steps:

[0043] Step 1: Obtain the number of various lithological samples in the training logging data;

[0044] Step 2: Use the lithological sample with the largest number as the standard lithological sample, and the other lithological samples as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than the threshold A are considered as supplementary samples; where the difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0045] Step 3: Cluster the training logging data to obtain clusters of lithological samples;

[0046] Step 4: For each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The samples that need to be supplemented corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the training logging data until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the training logging data is not greater than the threshold A.

[0047] Step 5: Train the lithology identification model using the training logging data after adding new samples.

[0048] The above method clusters the training logging data to obtain clusters of various lithological samples. Oversampling is performed on the samples that need to be supplemented corresponding to the oversampled clusters to make the number of lithological samples of all categories in the training logging data balanced. The lithological identification model is trained using logging data with balanced samples. The lithological identification model does not produce prediction bias and the identification is accurate.

[0049] The training of lithology identification models mainly uses well logging data, which generally contains various types of lithology samples. However, real well logging data is often not ideal and often exhibits an imbalance in lithology labels, meaning that the number of various lithology samples varies significantly. Therefore, in order to obtain balanced training data, it is necessary to supplement the lithology samples with fewer samples.

[0050] Based on the above analysis, we can first determine the number of various lithological samples in the training logging data, take the lithological sample with the largest number as the standard lithological sample, and take the other lithological samples as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than the threshold A are taken as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0051] For example, if the well logging data includes sandstone, mudstone, and limestone samples, with 100 sandstone samples, 30 mudstone samples, and 20 limestone samples, and the threshold A is 5, then the mudstone and limestone samples in the well logging data are both samples that need to be supplemented.

[0052] Due to instrument or human factors, noisy samples are unavoidable in well logging data. To reduce the interference of noise samples and enhance the robustness of the model, the training well logging data can be clustered to obtain clusters of lithological samples. Here, K-means clustering is used, and the specific process is as follows:

[0053] 1) Randomly select samples as initial cluster centers;

[0054] 2) For the cluster centers selected in 1), calculate the distance from all samples to each center based on Euclidean distance, and cluster the samples near the cluster centers into clusters;

[0055] 3) Calculate the new cluster centers after clustering, and use them as the new class centers;

[0056] 4) Then repeat steps 2) to 3) until the clustering results no longer change.

[0057] For each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as an oversampling cluster of samples that need to be supplemented. The samples that need to be supplemented corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the training logging data until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the training logging data is not greater than the threshold A.

[0058] Taking sandstone, mudstone, and limestone samples as examples, see the specific results below. Figure 2 The left cluster has the most mudstone samples, accounting for 7 / 9; the middle cluster has the most sandstone samples, accounting for 22 / 25; and the right cluster has the most limestone samples, accounting for 8 / 9. All of these exceed the threshold B. Since only mudstone and limestone samples need to be supplemented, the left and right clusters are oversampled to supplement mudstone and limestone samples to no less than 95%.

[0059] This invention employs the SMOTE method for oversampling. Specifically, it performs linear interpolation on the samples to be supplemented corresponding to the sampling clusters to generate new samples. The type of the new samples is consistent with the type of the samples to be supplemented. The formula for generating new samples using the SMOTE method is as follows:

[0060] X new =x + random(0,1) × |xx n |

[0061] In the formula, X new For the newly generated sample, random(0,1) is a random number between 0 and 1, x and x n Select two additional samples from the same cluster.

[0062] The lithology identification model is trained using well logging data with newly added samples, i.e., using well logging data with balanced samples. The lithology identification model of this invention employs a cascaded forest, see [link to relevant documentation]. Figure 3 The cascaded forest is composed of multiple sets of random forests and completely random forests cascaded together. Each forest unit is equivalent to a neuron in a neural network. Its cascaded structure gives the cascaded forest a deep structure similar to that of a deep neural network. As training progresses, the original logging sequence (logging data) is continuously enhanced and trained until the effect no longer improves. The purpose of using two tree models, random forest and completely random forest, is to ensure the diversity of the lithology identification model and improve the overall robustness.

[0063] The training and prediction process of cascaded forests is as follows:

[0064] S1) Input the original well logging sequence into the first group of random forest and completely random forest, and the four base models will make predictions on the sequence;

[0065] S2) The four sets of prediction results from S1) are concatenated with the original well logging sequence and input into the next set of random forest and completely random forest to continue prediction; the prediction results serve as the enhancement features of the sequence.

[0066] The cascade structure will continuously repeat the process S1) to S2) until the effect no longer improves or the effect improvement reaches its set value, then the cascade ends.

[0067] S1) Predict the final enhanced sequence after the cascading is completed. The prediction result will be obtained by voting from two random forests and two completely random forests.

[0068] This invention supplements unbalanced samples to balance various lithological samples in well logging data. The model trained with balanced well logging data exhibits stronger generalization performance and noise resistance.

[0069] Based on the same technical solution, this invention also discloses a corresponding virtual device (such as software) for the above method, and a training device for a lithology identification model, comprising:

[0070] The training sample quantity module retrieves the quantity of various lithological samples from the training logging data used for training.

[0071] The training requires supplementing the sample module. The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are used as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0072] The training clustering module clusters the training well logging data to obtain clusters of lithological samples.

[0073] The training oversampling module, for each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The supplementary samples corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the training logging data until the difference between the number of standard lithological samples and the number of supplementary samples is no greater than the threshold A.

[0074] The training module uses training well logging data with added new samples to train the lithology identification model.

[0075] Furthermore, such as Figure 4 As shown, the present invention also discloses a lithology identification method, comprising:

[0076] A1) Obtain the number of various lithological samples in the well logging data to be identified;

[0077] A2) The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than the threshold A are used as supplementary samples; where the difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0078] A3) Cluster the logging data to be identified to obtain clusters of lithological samples;

[0079] A4) For each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The samples that need to be supplemented corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the logging data to be identified until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the logging data to be identified is not greater than the threshold A.

[0080] A5) Input the logging data to be identified after adding new samples into the pre-trained lithology identification model to obtain lithology identification results; wherein, the trained lithology identification model is trained using the above-mentioned lithology identification model training method.

[0081] During the identification process, steps A1) to A4) are similar to the first four steps of the lithology identification model training method. They all involve supplementing the unbalanced samples and finally inputting the supplemented well logging data into the lithology identification model to obtain accurate identification results. This method can be well applied to lithology identification in areas with unbalanced lithology samples.

[0082] Based on the same technical solution, this invention also discloses a corresponding virtual device (such as software) for the above method, and a lithology identification system, comprising:

[0083] The sample quantity identification module obtains the quantity of various lithological samples in the well logging data to be identified.

[0084] The module for identifying samples that need to be supplemented uses the lithological sample with the largest number as the standard lithological sample and the other lithological samples as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are identified as samples that need to be supplemented. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples.

[0085] The clustering module identifies and clusters the logging data to be identified, obtaining clusters of lithological samples.

[0086] The oversampling identification module identifies each cluster as an oversampling cluster if the proportion of the most numerous lithological samples in the cluster is greater than threshold B and the most numerous lithological samples are samples that need to be supplemented. The cluster is then oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the logging data to be identified until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the logging data to be identified is no greater than threshold A.

[0087] The identification module inputs the well logging data to be identified after adding new samples into a pre-trained lithology identification model to obtain lithology identification results; wherein the trained lithology identification model is trained using the method described in any one of claims 1 to 4.

[0088] Based on the same technical solution, the present invention also discloses a computer-readable storage medium that stores one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform a lithology identification model training method or lithology identification.

[0089] Based on the same technical solution, the present invention also discloses a computer device, including one or more processors and one or more memories, wherein one or more programs are stored in the one or more memories and configured to be executed by the one or more processors, and the one or more programs include instructions for executing a lithology identification model training method or lithology identification.

[0090] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention 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.

[0091] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. 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 illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0092] 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.

[0093] 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.

[0094] The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of the claims of the present invention pending approval.

Claims

1. A method for training a lithology identification model, characterized in that, include: Obtain the number of various lithological samples in the training logging data; The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are used as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples. Clustering of training logging data yields clusters of lithological samples; For each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The samples that need to be supplemented corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the training logging data until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the training logging data is not greater than the threshold A. The lithology identification model was trained using well logging data with new samples added. The lithology identification model was a cascaded forest, which consisted of multiple sets of random forests and completely random forests. Each forest unit was equivalent to a neuron in a neural network. Its cascaded structure gave the cascaded forest the deep structure of a deep neural network. As training progressed, the original well logging data was continuously enhanced and trained until the effect could no longer be improved.

2. The lithology identification model training method according to claim 1, characterized in that, Clustering of training logging data yields clusters of lithological samples, including: K-means clustering was used to cluster the training logging data to obtain clusters of lithological samples.

3. The lithology identification model training method according to claim 1, characterized in that, Oversampling is performed on the samples that need to be supplemented corresponding to the oversampled clusters, including: The SMOTE method is used to oversample the samples that need to be supplemented corresponding to the oversampled clusters.

4. A training device for a lithology identification model, characterized in that, include: The training sample quantity module retrieves the quantity of various lithological samples from the training logging data used for training. The training requires supplementing the sample module. The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are considered as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples. The clustering module is trained to cluster the training well logging data to obtain clusters of lithological samples; The training oversampling module, for each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The supplementary samples corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the training logging data until the difference between the number of standard lithological samples and the number of supplementary samples is not greater than the threshold A. The training module uses training logging data with added new samples to train the lithology identification model. The lithology identification model is a cascaded forest, which is composed of multiple sets of random forests and completely random forests. Each forest unit is equivalent to a neuron in a neural network. Its cascaded structure gives the cascaded forest the deep structure of a deep neural network. As training progresses, the original logging data is continuously enhanced and trained until the effect no longer improves.

5. A lithology identification method, characterized in that, include: Obtain the number of various lithological samples in the well logging data to be identified; The lithological sample with the largest number is used as the standard lithological sample, and the other lithological samples are used as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are used as supplementary samples. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples. Clustering is performed on the well logging data to be identified to obtain clusters of lithological samples; For each cluster, if the proportion of the most numerous lithological samples in the cluster is greater than the threshold B, and the most numerous lithological samples are the samples that need to be supplemented, the cluster is regarded as the oversampling cluster of the samples that need to be supplemented. The samples that need to be supplemented corresponding to the oversampling cluster are oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the logging data to be identified until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the logging data to be identified is not greater than the threshold A. The well logging data to be identified after adding new samples is input into a pre-trained lithology identification model to obtain lithology identification results; wherein, the trained lithology identification model is trained using the method described in any one of claims 1 to 3.

6. A lithology identification system, characterized in that, include: The sample quantity identification module obtains the quantity of various lithological samples in the well logging data to be identified. The module for identifying samples that need to be supplemented uses the lithological sample with the largest number as the standard lithological sample and the other lithological samples as non-standard lithological samples. Non-standard lithological samples with a difference in number greater than a threshold A are identified as samples that need to be supplemented. The difference in number is the difference between the number of standard lithological samples and the number of non-standard lithological samples. The clustering module identifies and clusters the logging data to be identified, obtaining clusters of lithological samples. The oversampling identification module identifies each cluster as an oversampling cluster if the proportion of the most numerous lithological samples in the cluster is greater than threshold B and the most numerous lithological samples are samples that need to be supplemented. The cluster is then oversampled to generate new samples of the same type as the samples that need to be supplemented. The new samples are added to the logging data to be identified until the difference between the number of standard lithological samples and the number of samples that need to be supplemented in the logging data to be identified is no greater than threshold A. The identification module inputs the well logging data to be identified after adding new samples into a pre-trained lithology identification model to obtain lithology identification results; wherein the trained lithology identification model is trained using the method described in any one of claims 1 to 3.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores one or more programs, the one or more programs including instructions that, when executed by a computing device, cause the computing device to perform any of the methods described in claims 1 to 3 and 5.

8. A computer device, characterized in that, include: One or more processors and one or more memories, one or more programs stored in the one or more memories and configured to be executed by the one or more processors, the one or more programs including instructions for performing any of the methods of claims 1 to 3, 5.

Images