Well logging sedimentary microfacies division method and device based on target detection

Abstract

The present invention discloses a method and device for well logging sedimentary microfacies division based on target detection, the method comprising: adopting a multi-scale sliding window method, traversing a well logging gamma curve to be detected, and intercepting a plurality of frames to be detected; extracting curve feature information of each frame to be detected, inputting it into a pre-trained sedimentary microfacies classifier, and outputting the probability that each frame to be detected corresponds to different sedimentary microfacies; determining the sedimentary microfacies with the highest probability as the sedimentary microfacies to which each frame to be detected belongs, and counting the position information of the frames to be detected corresponding to different sedimentary microfacies; merging the regions where the same sedimentary microfacies corresponds to the frames to be detected and have a positional overlap relationship; determining the regions where different sedimentary microfacies correspond to the frames to be detected and have a positional overlap relationship as new detection frames, extracting curve feature information of the new detection frame, and determining the sedimentary microfacies to which the new detection frame belongs in combination with the vertical correlation information obtained by pre-statistication. The present invention can greatly improve the efficiency and accuracy of sedimentary microfacies division.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas exploration, and in particular to a method and device for dividing well logging sedimentary microfacies based on target detection. Background Art

[0002] This section is intended to provide a background or context to the embodiments of the invention recited in the claims. No admission is made that the description herein is prior art by inclusion in this section.

[0003] Sedimentary facies refers to the combination of sedimentary environment and the characteristics of sedimentary rocks (sediments) formed in this environment. Sedimentary facies control the lithological distribution and reservoir heterogeneity of the strata to a certain extent. Therefore, the study of sedimentary facies is the basis and focus of oil and gas exploration and development. Core data can directly indicate the sedimentary information of the strata and is the best data for studying sedimentary facies. Due to the high cost and limited quantity of core data, well logging data with rich data are widely used for phase belt identification and division in sedimentary research. Through the calibration of sedimentary facies types with core data, the indicative relationship between the morphology of well logging curves and the corresponding sedimentary microfacies can be established. At present, the division of sedimentary microfacies based on well logging data mainly relies on manual interpretation based on the morphology of well logging curves (such as the box type, bell type, and funnel type in the well logging gamma curve represent different sedimentary environments and sedimentary cycles). This method is affected by manual experience and proficiency. It is not only time-consuming and labor-intensive, but also has certain subjectivity and uncertainty in the interpretation results.

[0004] Among the many geological information detections based on well logging data, there are a lot of studies on the automatic division of well logging lithofacies. For example, the automatic division of lithofacies based on well logging data is based on supervised and unsupervised machine learning methods. The automatic division of lithofacies based on supervised machine learning includes discriminant analysis, neural network, Bayesian, etc. The automatic division of lithofacies based on unsupervised machine learning includes cluster analysis.

[0005] However, there is no method for automatic division of sedimentary microfacies based on well logging data in the existing technology. Summary of the invention

[0006] The embodiment of the present invention provides a method for dividing sedimentary microfacies based on target detection, which is used to solve the technical problem that the prior art uses an artificial method to divide sedimentary microfacies according to well logging data, which is time-consuming, labor-intensive and has inaccurate results. The method comprises: using a multi-scale sliding window method to traverse the well logging gamma curve to be detected and intercept multiple frames to be detected; extracting curve feature information of each frame to be detected; inputting the curve feature information of each frame to be detected into a pre-trained sedimentary microfacies classifier, and outputting the probability that each frame to be detected corresponds to different sedimentary microfacies; determining the sedimentary microfacies with the highest probability as the sedimentary microfacies to which each frame to be detected belongs, and counting the position information of the frames to be detected corresponding to different sedimentary microfacies; merging the areas where the same sedimentary microfacies has a positional overlap with the frames to be detected; determining the areas where different sedimentary microfacies have a positional overlap with the frames to be detected as new detection frames, extracting the curve feature information of the new detection frame, and determining the sedimentary microfacies to which the new detection frame belongs in combination with the vertical correlation information of different sedimentary microfacies obtained in advance.

[0007] The present invention also provides a device for dividing sedimentary microfacies based on target detection in a well logging method, which is used to solve the technical problem that the prior art uses a manual method to divide sedimentary microfacies according to well logging data, which is time-consuming, labor-intensive and inaccurate. The device includes: a well logging gamma curve cutting module, which uses a multi-scale sliding window method to traverse the well logging gamma curve to be detected and intercept multiple frames to be detected; a feature information extraction module, which is used to extract curve feature information of each frame to be detected; a sedimentary microfacies identification module, which is used to input the curve feature information of each frame to be detected into a pre-trained sedimentary microfacies classifier, and output the information of each frame to be detected corresponding to different sedimentary microfacies. probability; a sedimentary microfacies position information determination module, which is used to determine the sedimentary microfacies with the highest probability as the sedimentary microfacies to which each to-be-detected frame belongs, and to count the position information of the to-be-detected frames corresponding to different sedimentary microfacies; a same sedimentary microfacies overlapping region processing module, which is used to merge the regions with overlapping positions of the same sedimentary microfacies corresponding to the to-be-detected frames; a different sedimentary microfacies overlapping region processing module, which is used to determine the regions with overlapping positions of different sedimentary microfacies corresponding to the to-be-detected frames as new detection frames, extract the curve feature information of the new detection frame, and determine the sedimentary microfacies to which the new detection frame belongs in combination with the vertical correlation information of different sedimentary microfacies obtained in advance.

[0008] A computer device is also provided in an embodiment of the present invention to solve the technical problem that the prior art uses a manual method to divide sedimentary microfacies according to logging data, which is time-consuming, labor-intensive and has inaccurate results. The computer device includes a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned target detection-based logging sedimentary microfacies division method is implemented.

[0009] A computer-readable storage medium is also provided in an embodiment of the present invention to solve the technical problem that the prior art uses manual methods to divide sedimentary microfacies according to logging data, which is time-consuming, labor-intensive and has inaccurate results. The computer-readable storage medium stores a computer program for executing the above-mentioned target detection-based logging sedimentary microfacies division method.

[0010] In an embodiment of the present invention, a multi-scale sliding window method is adopted to traverse the well logging gamma curve to be detected, and multiple frames to be detected are intercepted to extract the curve feature information of each frame to be detected, and based on a pre-trained sedimentary microfacies classifier, the probability of each frame to be detected corresponding to different sedimentary microfacies is obtained, and the sedimentary microfacies with the highest probability is determined as the sedimentary microfacies to which each frame to be detected belongs, and the position information of the frames to be detected corresponding to different sedimentary microfacies is counted, and finally, the areas where the same sedimentary microfacies and the frames to be detected have a positional overlap relationship are merged, and the areas where different sedimentary microfacies and the frames to be detected have a positional overlap relationship are determined as new detection frames, and the curve feature information of the new detection frame is extracted, and the sedimentary microfacies to which the new detection frame belongs is determined in combination with the vertical correlation information of different sedimentary microfacies obtained in advance. Compared with the technical solution of manually interpreting sedimentary microfacies in the prior art, the embodiment of the present invention realizes automatic division of sedimentary microfacies based on target detection, which can greatly improve the efficiency and accuracy of sedimentary microfacies division. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the prior art descriptions. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0012] Figure 1 A flow chart of a method for well logging sedimentary microfacies division based on target detection provided in an embodiment of the present invention;

[0013] Figure 2 A schematic diagram comparing the least squares gradient calculation scheme provided in the prior art and the gradient calculation method combining the random sampling consensus algorithm and the least squares method provided in an embodiment of the present invention;

[0014] Figure 3 A schematic diagram of an original gamma curve and its corresponding random simulation curve provided in an embodiment of the present invention;

[0015] Figure 4 A schematic diagram of a well logging sedimentary microfacies target detection strategy provided in an embodiment of the present invention;

[0016] Figure 5A flowchart for implementing a method for well logging sedimentary microfacies division based on target detection provided in an embodiment of the present invention;

[0017] Figure 6 Schematic diagram of sedimentary microfacies division results obtained by respectively using the sedimentary microfacies manual interpretation scheme provided by the prior art and the sedimentary microfacies automatic division scheme provided in the embodiments of the present invention;

[0018] Figure 7 It is a schematic diagram of a well logging sedimentary microfacies classification device based on target detection provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0019] To make the purpose, technical solution and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below in conjunction with the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0020] The present invention provides a method for dividing well logging sedimentary microfacies based on target detection. Figure 1 Flow chart of a method for well logging sedimentary microfacies division based on target detection provided in an embodiment of the present invention, such as Figure 1 As shown, the method comprises the following steps:

[0021] S101, using a multi-scale sliding window method, traverse the well logging gamma curve to be detected, and intercept multiple frames to be detected.

[0022] It should be noted that the multi-scale sliding window method refers to a method of using multiple sliding windows of preset sizes to slide in equidistant steps to form a large number of windows. In an embodiment of the present invention, for the curve to be detected, a multi-scale sliding window method is applied to obtain a large number of frames to be detected. For example, using a window with a length of 2m, traverse the curve to be detected in a step of 0.5m, and intercept a series of frames to be detected; then use a window with a length of 2m to traverse the curve to be detected in a step of 0.75m to obtain a series of frames to be detected; continue to use a window with a length of 2m to traverse the curve in a step of 1m until the step length reaches the defined maximum value. Then, increase the window length and use a window with a length of 2.5m for similar traversal. Continue to increase the window length for similar traversal until the window length reaches the defined maximum value.

[0023] S102: extract curve feature information of each frame to be detected.

[0024] It should be noted that the curve characteristic information extracted in the embodiment of the present invention refers to characteristic information extracted from the logging gamma curve that can characterize the sedimentary information to distinguish and describe different sedimentary microfacies. The extracted curve characteristic information includes but is not limited to: mean, amplitude, relative centroid and gradient information.

[0025] Assume that the logging gamma curve sequence is a i (i=1,.....n)(a i represents the i-th value of the logging gamma curve sequence, and n is the total number of points in the sequence, i.e., the length). The extraction method of each feature information is as follows:

[0026] ① Mean:

[0027]

[0028] ②Amplitude:

[0029]

[0030] Where Median represents the median of the logging sequence. The amplitude characteristics of the gamma curve reflect the changes in the sedimentary environment energy.

[0031] ③Relative center of mass:

[0032]

[0033] The relative centroid can reflect the shape characteristics of the logging curve. rb When f ≥0.5, the curve is bell-shaped; rb ≈0.5, the curve is box-shaped; when f rb When ≤0.5, the curve is funnel-shaped.

[0034] ④ Gradient:

[0035] The overall gradient of the logging curve is the most important feature for identifying the curve shape and classifying the sedimentary microfacies types. For example, when the sedimentary facies is an estuary bar, the corresponding gamma curve is funnel-shaped with a gradient greater than 90°; when the sedimentary facies is an underwater distributary channel, the corresponding gamma curve is box-shaped or bell-shaped with a gradient close to 90° or less than 90°.

[0036] It should be noted that in the embodiment of the present invention, a random sampling consensus algorithm is applied to accurately calculate the overall gradient of the logging curve. Usually, when calculating the gradient of a logging curve segment, the top and bottom data of the curve are easily outliers that affect the accuracy of the gradient calculation. The random sampling consensus algorithm can eliminate the interference of noise data, obtain the main trend of the data, and thus accurately extract the overall gradient of the curve. Figure 2 A schematic diagram showing a comparison between the least squares gradient calculation scheme provided in the prior art and the gradient calculation method combining the random sampling consensus algorithm and the least squares method provided in an embodiment of the present invention, as shown in FIG. Figure 2 As shown, the line shown in icon 201 is the result of fitting all curve points using the least squares method; the line shown in icon 202 is the result of first extracting representative points using a random sampling consensus algorithm and then fitting all representative points using the least squares method. Figure 2 The "star" points on the curve shown in are points representing the main trend extracted by applying the random sampling consensus algorithm. It can be seen that the line shown in icon 202 can more accurately reflect the overall gradient of the curve segment than the line shown in icon 201.

[0037] S103, inputting the curve feature information of each to-be-detected frame into a pre-trained sedimentary microfacies classifier, and outputting the probability that each to-be-detected frame corresponds to a different sedimentary microfacies.

[0038] In specific implementation, the target detection-based logging sedimentary microfacies classification method provided in the embodiment of the present invention can be trained to obtain a sedimentary microfacies classifier through the following steps: obtaining training sample data for machine learning of the sedimentary microfacies classifier; and obtaining the sedimentary microfacies classifier through machine learning training based on the obtained training sample data.

[0039] It should be noted that the number of samples of each type is basically the same, which is the basis for achieving good machine learning results. However, in the same sedimentary facies belt, the number of different sedimentary microfacies is different. Therefore, in order to achieve better machine learning results, the embodiment of the present invention can perform data expansion on sedimentary microfacies with a small number of samples before using sample data to perform machine learning on the sedimentary microfacies classifier.

[0040] Therefore, before executing the above S103, the embodiment of the present invention generates training sample data for machine learning of the sedimentary microfacies classifier through the following steps: collecting sample data of different sedimentary microfacies in the same sedimentary facies belt; expanding the sample data of sedimentary microfacies with insufficient sample quantity to make the sample data of different sedimentary microfacies in the same sedimentary facies belt consistent; and generating training sample data for machine learning of the sedimentary microfacies classifier based on the expanded sample data.

[0041] In the specific implementation, the sample data is expanded through the following steps: obtaining the original logging gamma curve of the sedimentary microfacies to be expanded; based on the original logging gamma curve of the sedimentary microfacies to be expanded, using the Monte Carlo spatial correlation simulation method to generate multiple random simulation curves of the sedimentary microfacies to be expanded, wherein each random simulation curve is consistent with the length of the original logging gamma curve.

[0042] It should be noted that applying the Monte Carlo spatial correlation simulation method to expand the data can generate a series of random simulated gamma curves, such as Figure 3 As shown, six simulation curves are generated for the original logging gamma curve.

[0043] The specific steps are as follows:

[0044] ① Smooth the original gamma curve to obtain its depth domain trend m, and calculate the residual between it and the original gamma curve;

[0045] ②Establish the residual variogram based on the Gaussian model (h) (h=0,1,2,3.....d max ), and calculate the corresponding spatially symmetric covariance matrix C:

[0046]

[0047] ③Perform Cholesky decomposition on C and obtain the lower triangular matrix R:

[0048] C=RR T (5)

[0049] ④ Generate a random vector U (U~N(0,1)), whose length is consistent with the original data;

[0050] ⑤ According to the formula V=m+RU, random simulation can be performed to generate a series of random simulation curves, which can be used as a supplement to the original sample data and participate in subsequent machine learning.

[0051] S104, determining the sedimentary microfacies with the highest probability as the sedimentary microfacies to which each to-be-detected frame belongs, and counting the position information of the to-be-detected frames corresponding to different sedimentary microfacies.

[0052] It should be noted that after the probability of each to-be-detected frame corresponding to each sedimentary microfacies is output by the sedimentary microfacies classifier, the sedimentary microfacies with the highest probability is determined as the sedimentary microfacies to which the to-be-detected frame belongs, and then the sedimentary microfacies to which all to-be-detected frames belong are obtained. The to-be-detected frame corresponding to each sedimentary microfacies can be obtained by statistics. Since the position information of the to-be-detected frame on the to-be-detected well logging gamma curve is known, the position information of each sedimentary microfacies distribution can be obtained, such as Figure 4 (a) shows the position information of the detection boxes for the three sedimentary microfacies.

[0053] In one embodiment, after obtaining the position information of the to-be-detected frames corresponding to different sedimentary microphases, the non-maximum suppression method can also be applied to process the position information of each sedimentary microphase distribution to remove redundant to-be-detected frames. Figure 4 The position information of the three sedimentary microfacies to be detected is processed to obtain Figure 4 (b) shows the position information of the detection boxes of the three sedimentary microfacies after removing redundancy.

[0054] S105, merging the regions of the same deposited micro-phase corresponding to the to-be-detected frame that have a positional overlap relationship.

[0055] It should be noted that by merging the overlapping areas of the to-be-detected frames of each sedimentary microfacies type, the distribution location information of each sedimentary microfacies type can be obtained. Figure 4 (c) shows the result of merging the overlapping areas of the detection frames of the same sedimentary microfacies type.

[0056] S106, determining the area with overlapping position relationship between different sedimentary microfacies and corresponding to the to-be-detected frames as a new detection frame, extracting curve feature information of the new detection frame, and combining the vertical correlation information of different sedimentary microfacies obtained in advance to determine the sedimentary microfacies to which the new detection frame belongs.

[0057] It should be noted that for the overlapping areas between different sedimentary microfacies, a new frame to be detected (i.e., a new detection frame) is formed, and the characteristic information of the curve corresponding to the new detection frame is extracted. Combined with the vertical correlation information of the sedimentary microfacies, the sedimentary microfacies types to which the different new detection frames belong can be determined. Figure 4 (d) shows the new detection frame formed; Figure 4 (e) shows the final target detection result.

[0058] In one embodiment, the target detection-based logging sedimentary microfacies division method provided in the embodiment of the present invention can also determine the vertical correlation information of different sedimentary microfacies through the following steps: statistically analyzing the vertical proximity probabilities of different sedimentary microfacies; and determining the vertical correlation information of different sedimentary microfacies based on the statistical results.

[0059] Since different sedimentary microfacies have certain proximity relationships in the vertical direction, it is necessary to calculate the vertical proximity probabilities of different sedimentary microfacies. Taking the target layer with estuary bar (MB), interchannel (IC) and underwater distributary channel (UDC) as an example, the matrix M is a conditional probability density matrix based on vertical correlation statistics.

[0060]

[0061] Among them, P(F|f k ) indicates that there is a kth phase type f k When , the conditional probability density of sedimentary microfacies F appears. The addition of the vertical correlation matrix of sedimentary microfacies can improve the accuracy of sedimentary microfacies machine learning.

[0062] From the above, it can be seen that the embodiment of the present invention is based on the morphological characteristics of well logging sedimentary microfacies, converts the manually interpreted sedimentary microfacies into target detection, and comprehensively applies machine learning and computer vision technology to realize the automatic division of well logging sedimentary microfacies. Based on the principle of sedimentation, in the embodiment of the present invention, four attributes are preferably used as the characteristics to describe the sedimentary microfacies. A random sampling consensus algorithm is applied to accurately calculate the overall gradient of the well logging curve to extract accurate curve features for subsequent machine learning. The recognition accuracy of machine learning is improved by adding vertical correlation of sedimentary microfacies. A target detection strategy for sedimentary microfacies is designed in a targeted manner. Compared with the scheme of manually interpreting sedimentary microfacies division in the prior art, the well logging sedimentary microfacies division method based on target detection provided in the embodiment of the present invention greatly improves the accuracy and efficiency of sedimentary microfacies division.

[0063] In specific implementation, the embodiment of the present invention can be Figure 5 This can be achieved by following the process shown in the figure. Figure 5 As shown in the figure, after collecting sample data (gamma curves) of different sedimentary microfacies, data expansion is performed on the sedimentary microfacies with a small number of samples, and the characteristic information of the gamma curves corresponding to different sedimentary microfacies is extracted. A sedimentary microfacies classifier is obtained through machine learning training, and the trained sedimentary microfacies classifier is used to perform target detection on the logging gamma curve to be detected, and the automatic division result of the sedimentary microfacies corresponding to the logging gamma curve to be detected is obtained.

[0064] In one embodiment, the target detection-based logging sedimentary microfacies division method provided in the embodiment of the present invention is applied to the 8th section of the Upper Triassic Yanchang Formation in a certain work area of ​​Huanxian-Heshui in the Ordos Basin, and the three types of sedimentary microfacies, namely estuary bar, inter-channel and underwater diversion channel, are automatically divided based on the logging gamma curve. More than 130 core records from 9 wells were obtained in this work area. The data was expanded by applying the Monte Carlo spatial correlation simulation method to form 200 samples, of which 70% were used as training samples and 30% were used as test samples. After testing with multiple machine learning methods, the naive Bayes classifier is more suitable for identifying sedimentary microfacies, and this method does not require a large number of samples to obtain good classification results.

[0065] The vertical relationship between the three sedimentary microfacies in the target layer of the work area was statistically analyzed, and a vertical correlation matrix was established:

[0066]

[0067] MB, IC, and UDC in the matrix represent estuary bars, interchannels, and underwater diversion channels, respectively. For example, in this work area, the probability of interchannels developing above estuary bars is 57%, and the probability of underwater diversion channels appearing above estuary bars is 43%.

[0068] Combining the trained classifier, sedimentary microfacies vertical relationship matrix and target detection method, the sedimentary microfacies of the target layer of the gamma curve of well 1 and well 2 are automatically divided. The results are as follows: Figure 6 As shown, microfacies 1, microfacies 2, and microfacies 3 represent estuary bars, inter-channels, and underwater distributary channels, respectively. In well 1, the manual sedimentary microfacies interpretation results of the long 8 sections and the automatic division results of the embodiment of the present invention are highly similar. The reason for the good recognition effect is that the gamma curve responses of different sedimentary microfacies in the target layer section in well 1 are clearly distinguished. In well 2, there is a significant difference between the manual sedimentary microfacies division results of the long 8 sections and the automatic division results of the present invention. This is because the gamma curve characteristics of some sections of the target layer of well 2 are not significant. For example, the gamma curve at the top of well 2 can be interpreted as an estuary bar or an inter-channel. In fact, if the present invention is based on good machine learning training results and rigorous vertical correlation statistics of sedimentary microfacies, the boundaries of the automatic division are more reasonable and objective than manual interpretation. At the bottom of well 2, the estuary bar boundary automatically divided by the present invention is more reasonable than the manual division. On the contrary, there is no quantitative standard for reference in the interpretation of artificial sedimentary microfacies, so there is a certain degree of subjectivity, especially in the face of well sections with atypical curve characteristics.

[0069] Based on the same inventive concept, an embodiment of the present invention also provides a well logging sedimentary microphase division device based on target detection, as described in the following embodiments. Since the principle of solving the problem by the device is similar to that of the well logging sedimentary microphase division method based on target detection, the implementation of the device can refer to the implementation of the well logging sedimentary microphase division method based on target detection, and the repeated parts will not be repeated.

[0070] Figure 7 Schematic diagram of a well logging sedimentary microfacies classification device based on target detection provided in an embodiment of the present invention, such as Figure 7 As shown, the device includes: a logging gamma curve cutting module 71, a feature information extraction module 72, a sedimentary microfacies identification module 73, a sedimentary microfacies position information determination module 74, a same sedimentary microfacies overlapping area processing module 75 and a different sedimentary microfacies overlapping area processing module 76.

[0071] Among them, the logging gamma curve cutting module 71 is used to adopt a multi-scale sliding window method to traverse the logging gamma curve to be detected and intercept multiple frames to be detected; the feature information extraction module 72 is used to extract the curve feature information of each frame to be detected; the sedimentary microfacies identification module 73 is used to input the curve feature information of each frame to be detected into a pre-trained sedimentary microfacies classifier, and output the probability that each frame to be detected corresponds to different sedimentary microfacies; the sedimentary microfacies position information determination module 74 is used to determine the sedimentary microfacies with the highest probability as the sedimentary microfacies to which each frame to be detected belongs, and to count the position information of the frames to be detected corresponding to different sedimentary microfacies; the same sedimentary microfacies overlapping area processing module 75 is used to merge the areas with overlapping position relationship of the same sedimentary microfacies corresponding to the frames to be detected; the different sedimentary microfacies overlapping area processing module 76 is used to determine the areas with overlapping position relationship of different sedimentary microfacies corresponding to the frames to be detected as new detection frames, extract the curve feature information of the new detection frame, and determine the sedimentary microfacies to which the new detection frame belongs in combination with the vertical correlation information of different sedimentary microfacies obtained in advance.

[0072] In one embodiment, the target detection-based well logging sedimentary microfacies classification device provided in the embodiment of the present invention may also include: a machine learning module 77, which is used to obtain training sample data for machine learning of the sedimentary microfacies classifier, and obtain the sedimentary microfacies classifier through machine learning training based on the acquired training sample data.

[0073] In one embodiment, the target detection-based well logging sedimentary microfacies classification device provided in the embodiment of the present invention may also include: a sample data expansion module 78, which is used to collect sample data of different sedimentary microfacies in the same sedimentary facies belt, and expand the sample data of sedimentary microfacies with insufficient sample quantity, so that the sample data of different sedimentary microfacies in the same sedimentary facies belt are consistent, and then generate training sample data for machine learning of the sedimentary microfacies classifier based on the expanded sample data.

[0074] Optionally, the sample data expansion module 78 is also used to obtain the original logging gamma curve of the sedimentary microfacies to be expanded; and based on the original logging gamma curve of the sedimentary microfacies to be expanded, a Monte Carlo spatial correlation simulation method is used to generate multiple random simulation curves of the sedimentary microfacies to be expanded, wherein each random simulation curve is consistent with the length of the original logging gamma curve.

[0075] In one embodiment, the target detection-based well logging sedimentary microfacies division device provided in the embodiment of the present invention may also include: a sedimentary microfacies vertical correlation analysis module 79, which is used to count the vertical proximity probabilities of different sedimentary microfacies; and determine the vertical correlation information of different sedimentary microfacies based on the statistical results.

[0076] Based on the same inventive concept, a computer device is also provided in an embodiment of the present invention to solve the technical problem that the prior art uses a manual method to divide sedimentary microfacies according to logging data, which is time-consuming, labor-intensive and has inaccurate results. The computer device includes a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned target detection-based logging sedimentary microfacies division method is implemented.

[0077] Based on the same inventive concept, a computer-readable storage medium is also provided in an embodiment of the present invention to solve the technical problem that the prior art uses manual methods to divide sedimentary microfacies according to logging data, which is time-consuming, labor-intensive and has inaccurate results. The computer-readable storage medium stores a computer program for executing the above-mentioned target detection-based logging sedimentary microfacies division method.

[0078] In summary, the embodiments of the present invention provide a method, device, computer equipment and computer-readable storage medium for well logging sedimentary microfacies division based on target detection. The multi-scale sliding window method is adopted to traverse the well logging gamma curve to be detected, and multiple frames to be detected are intercepted to obtain the curve feature information of each frame to be detected. Based on the pre-trained sedimentary microfacies classifier, the probability of each frame to be detected corresponding to different sedimentary microfacies is obtained, and the sedimentary microfacies with the highest probability is determined as the sedimentary microfacies to which each frame to be detected belongs, and the position information of the frames to be detected corresponding to different sedimentary microfacies is counted. Finally, the areas where the same sedimentary microfacies has a positional overlap relationship with the frames to be detected are merged, and the areas where different sedimentary microfacies have a positional overlap relationship with the frames to be detected are determined as new detection frames, and the curve feature information of the new detection frame is extracted. Combined with the vertical correlation information of different sedimentary microfacies obtained in advance, the sedimentary microfacies to which the new detection frame belongs is determined. Compared with the technical solution of manually interpreting sedimentary microfacies in the prior art, the embodiments of the present invention realize automatic division of sedimentary microfacies based on target detection, which can greatly improve the efficiency and accuracy of sedimentary microfacies division.

[0079] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented 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.

[0080] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0081] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0082] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0083] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for dividing well logging sedimentary microfacies based on target detection, characterized in that: include: A multi-scale sliding window method is used to traverse the gamma curve of the well logging to be detected and intercept multiple frames to be detected; The multi-scale sliding window method is a method of using multiple sliding windows of preset sizes to slide in equidistant steps to form a large number of windows; Extract the curve feature information of each frame to be detected; The extracted curve characteristic information is characteristic information extracted from the well logging gamma curve that can characterize the sedimentary information. The extracted curve characteristic information includes: mean value, amplitude, relative centroid and gradient information; The curve feature information of each to-be-detected frame is input into a pre-trained sedimentary microfacies classifier, and the probability that each to-be-detected frame corresponds to a different sedimentary microfacies is output; The sedimentary microfacies with the highest probability is determined as the sedimentary microfacies to which each frame to be detected belongs, and the position information of the frames to be detected corresponding to different sedimentary microfacies is counted; Merge the regions with overlapping positions of the same deposited micro-phase and the corresponding to-be-detected frames; The area where different sedimentary microfacies correspond to the to-be-detected frames with overlapping positions is determined as a new detection frame, and the curve feature information of the new detection frame is extracted. Combined with the vertical correlation information of different sedimentary microfacies obtained in advance, the sedimentary microfacies to which the new detection frame belongs is determined.

2. The method according to claim 1, characterized in that The curve characteristic information includes: mean value, amplitude, relative centroid and gradient information.

3. The method according to claim 2, characterized in that The random sampling consensus algorithm is used to extract the gradient information of the logging gamma curve.

4. The method according to claim 1, characterized in that The method further comprises: Acquiring training sample data for machine learning of the sedimentary microfacies classifier; The sedimentary microfacies classifier is obtained through machine learning training based on the acquired training sample data.

5. The method according to claim 4, characterized in that Acquiring training sample data for machine learning of the sedimentary microfacies classifier, including: Collect sample data of different sedimentary microphases in the same sedimentary facies belt; The sample data of sedimentary microfacies with insufficient sample quantity are expanded to make the sample data of different sedimentary microfacies in the same sedimentary facies belt consistent; Based on the expanded sample data, training sample data for machine learning of the sedimentary microfacies classifier is generated.

6. The method according to claim 5, characterized in that The sample data of sedimentary microfacies with insufficient sample quantity are expanded to make the sample data of different sedimentary microfacies in the same sedimentary facies belt consistent, including: Obtain the original logging gamma curve of the sedimentary microfacies to be expanded; According to the original logging gamma curve of the sedimentary microfacies to be expanded, a Monte Carlo spatial correlation simulation method is used to generate multiple random simulation curves of the sedimentary microfacies to be expanded, wherein the length of each random simulation curve is consistent with that of the original logging gamma curve.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: Statistical analysis of the vertical proximity probabilities of different sedimentary microfacies; Based on the statistical results, the vertical correlation information of different sedimentary microfacies is determined.

8. A well logging sedimentary microfacies division device based on target detection, characterized in that: include: The well logging gamma curve cutting module is used to traverse the well logging gamma curve to be detected by adopting a multi-scale sliding window method and intercept multiple frames to be detected; The multi-scale sliding window method is a method of using multiple sliding windows of preset sizes to slide in equidistant steps to form a large number of windows; A feature information extraction module is used to extract curve feature information of each frame to be detected; The extracted curve characteristic information is characteristic information extracted from the well logging gamma curve that can characterize the sedimentary information. The extracted curve characteristic information includes: mean value, amplitude, relative centroid and gradient information; The sedimentary microfacies identification module is used to input the curve feature information of each to-be-detected frame into a pre-trained sedimentary microfacies classifier, and output the probability that each to-be-detected frame corresponds to a different sedimentary microfacies; A sedimentary microfacies position information determination module is used to determine the sedimentary microfacies with the highest probability as the sedimentary microfacies to which each to-be-detected frame belongs, and to count the position information of the to-be-detected frames corresponding to different sedimentary microfacies; The same sedimentary microfacies overlapping region processing module is used to merge the regions of the same sedimentary microfacies corresponding to the to-be-detected frames that have overlapping positions; The module for processing overlapping areas of different sedimentary microfacies is used to determine the area with overlapping position relationship between different sedimentary microfacies and corresponding to the to-be-detected frames as a new detection frame, extract the curve feature information of the new detection frame, and determine the sedimentary microfacies to which the new detection frame belongs in combination with the vertical correlation information of different sedimentary microfacies obtained in advance.

9. The device according to claim 8, characterized in that The device also includes: The machine learning module is used to obtain training sample data for machine learning of the sedimentary microfacies classifier, and obtain the sedimentary microfacies classifier through machine learning training based on the acquired training sample data.

10. The device according to claim 9, characterized in that The device also includes: The sample data expansion module is used to collect sample data of different sedimentary microphases in the same sedimentary facies belt, expand the sample data of sedimentary microphases with insufficient sample quantity, make the sample data of different sedimentary microphases in the same sedimentary facies belt consistent, and then generate training sample data for machine learning of sedimentary microphase classifier based on the expanded sample data.

11. The device according to claim 10, characterized in that The sample data expansion module is also used to obtain the original logging gamma curve of the sedimentary microfacies to be expanded; and based on the original logging gamma curve of the sedimentary microfacies to be expanded, a Monte Carlo spatial correlation simulation method is used to generate multiple random simulation curves of the sedimentary microfacies to be expanded, wherein each random simulation curve is consistent with the length of the original logging gamma curve.

12. The device according to any one of claims 8 to 11, characterized in that The device also includes: The sedimentary microfacies vertical correlation analysis module is used to count the vertical proximity probabilities of different sedimentary microfacies; and to determine the vertical correlation information of different sedimentary microfacies based on the statistical results.

13. A computer 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 well logging sedimentary microfacies division based on target detection as described in any one of claims 1 to 7 is implemented.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program for executing the method for well logging sedimentary microfacies division based on target detection as described in any one of claims 1 to 7.

Images