Seismic exploration big data sample collection and labeling method and device

Abstract

The present application provides a kind of seismic exploration big data sample collection marking method and device, the seismic exploration big data sample collection marking method includes: step 1, load the seismic data of work area, and carry out the pre-processing of seismic data;Step 2, interactive setting digital type sample parameter and picture type sample parameter;Step 3, interactive drawing digital type sample collection boundary, and interactive setting picture type collection range and sample size;Step 4, carry out digital type sample collection and picture type sample collection;Step 5, mark digital type sample and picture type sample;Step 6, release seismic exploration big data sample.The seismic exploration big data sample collection marking method and device have the innovation of seismic exploration industry, have multiple application scene universality, high collection efficiency, user operation is simple and the like advantages.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent oil and gas exploration, and specifically relates to a method and device for collecting and labeling large seismic exploration data samples. Background Art

[0002] Currently, intelligent applications in seismic exploration are primarily focused on seismic data processing and interpretation. Key application scenarios include seismic structural interpretation, structural modeling, seismic phase identification, reservoir parameter prediction, noise suppression and signal enhancement, seismic wavefield reconstruction, seismic inversion, seismic velocity picking, first arrival picking, and comprehensive interpretation. While a small number of these applications generate training sample sets through forward modeling of three-dimensional geological models, the vast majority generate training and prediction sample sets through attribute or image extraction of seismic exploration data. The primary representation methods for these samples include digital and image types. Current big data sample annotation methods generally use general industry annotation methods, which differ significantly from the annotation requirements for seismic exploration big data samples. Seismic exploration big data sample collection is slow, interactive annotation of seismic data in a visual manner is impossible, and big data sample formats are inconsistent and cannot be shared.

[0003] In the Chinese patent application with application number: CN201911308462.9, a microseismic event detection method and system are involved. The steps of the method are: collecting microseismic signals from multiple monitoring stations during the fracturing process to establish a training data set and a test data set; establishing a convolutional neural network model, inputting samples of the training data set into the convolutional neural network model for training, and inputting samples of the test data set into the trained convolutional neural network model to test the performance of the trained convolutional neural network model; storing the parameters of the convolutional neural network model after training; collecting microseismic signals from multiple monitoring stations during real-time fracturing to establish a data set to be tested; inputting the data of the data set to be tested into the trained convolutional neural network model for detection, obtaining sample classification results of the waveform data of each monitoring station, and determining whether there is a microseismic event based on the sample classification results.

[0004] In the Chinese patent application with application number: CN201911140333.3, a method for first-arrival picking of seismic data is involved, which includes: obtaining multiple sample data labeled with semantic classification, wherein the semantic classification includes a background area class and an effective reflection area class; using the sample data labeled with semantic classification to train a semantic segmentation network model; using the trained semantic segmentation network model to perform semantic segmentation on the seismic data to be picked for first-arrival picking to obtain semantic classification; and determining the first-arrival position according to the semantic classification.

[0005] In the Chinese patent application with application number CN202010164500.4, a method for processing earthquake fault images is disclosed, which includes the following steps: obtaining earthquake fault image data sets; preprocessing the earthquake fault image data sets; constructing an earthquake fault image processing network; training the earthquake fault image processing network; obtaining earthquake fault image data to be processed; and processing the earthquake fault image.

[0006] The above existing technologies are significantly different from the present invention and fail to solve the technical problems we want to solve. To this end, we have invented a new method and device for collecting and labeling seismic exploration big data samples to solve the above technical problems. Summary of the Invention

[0007] The purpose of the present invention is to provide a method and device for collecting and labeling seismic exploration big data samples for intelligent application research in oil field exploration.

[0008] The purpose of the present invention can be achieved by the following technical measures: a method for collecting and labeling large data samples of seismic exploration, which comprises:

[0009] Step 1: Load the seismic data of the work area and preprocess the seismic data;

[0010] Step 2: interactively set the sample parameters of the digital type and the sample parameters of the image type;

[0011] Step 3: interactively draw the sample collection boundary for digital type, and interactively set the collection range and sample size for image type;

[0012] Step 4: Collect digital type samples and picture type samples;

[0013] Step 5: Label the digital type samples and the picture type samples;

[0014] Step 6: Release seismic exploration big data samples.

[0015] The purpose of the present invention can also be achieved by the following technical measures:

[0016] In step 1, based on the large file segmentation and merging technology, seismic data of three different dimensions, X, Y, and Z, are generated and redundantly stored to improve the efficiency of reading seismic exploration data and the efficiency of collecting seismic exploration big data samples.

[0017] In step 2, interactively select the work area, seismic data volume, and survey line type, and interactively set the sample scene.

[0018] In step 3, when interactively drawing the digital type sample collection boundary, two methods of drawing the sample collection boundary are supported: rectangle and polygon.

[0019] In step 3, interactively set the image type acquisition range and sample size, including interactively setting the range and step size of the main survey line (Inline), crossline (Crossline), and horizontal time slice (Time), and interactively set the sample size and eigenvalue name.

[0020] In step 4, digital type sample collection and picture type sample collection are implemented based on the set sample parameters and the drawn sample collection boundary.

[0021] In step 5, select a scene, select a seismic data volume, select a sample boundary list, specify a label, and perform digital type seismic exploration big data sample labeling.

[0022] In step 5, select a picture type data sample set, select the sample to be calibrated, and perform image type seismic exploration big data sample labeling.

[0023] In step 6, the seismic exploration big data samples are managed, uploaded, exported, and published.

[0024] The purpose of the present invention can also be achieved by the following technical measures: a seismic exploration big data sample collection and labeling device, the seismic exploration big data sample collection and labeling device comprising:

[0025] The earthquake pre-processing module generates earthquake data in three different dimensions: X, Y, and Z, and stores them redundantly.

[0026] Collection boundary interactive drawing module, interactively drawing the collection boundary of digital type samples;

[0027] The sample size collection area setting module interactively sets the sample size and collection area of ​​the image type big data sample;

[0028] Seismic sample annotation module, which annotates the collected digital and image type big data samples;

[0029] The automatic seismic sample acquisition module is connected to the seismic preprocessing module, the acquisition boundary interactive drawing module, the sample size acquisition area setting module and the seismic sample labeling module. Based on the preprocessed seismic data and the interactively drawn digital type big data sample boundaries, the module extracts seismic attributes to form digital type big data samples and transmits them to the seismic sample labeling module; based on the preprocessed seismic data and the size and area of ​​the interactively set picture type big data samples, the module extracts seismic images to form picture type big data samples and transmits them to the seismic sample labeling module.

[0030] The purpose of the present invention can also be achieved by the following technical measures:

[0031] The seismic exploration big data sample collection and labeling device further comprises a seismic loading module connected to the seismic preprocessing module, which loads the seismic data of the work area.

[0032] The seismic exploration big data sample collection and labeling device further comprises a sample publishing module connected to the seismic sample labeling module, which manages, uploads, exports and publishes the seismic exploration big data samples.

[0033] The seismic exploration big data sample collection and labeling method and device have industry innovation in seismic exploration, and the labeling tool has the advantages of multi-application scene universality, high collection efficiency and simple user operation. The following two effects are obvious.

[0034] (1) The digital type and picture type seismic exploration big data sample collection is efficiently and conveniently extracted from the seismic exploration data body, and the sample labeling is realized, which has universality in the field of seismic exploration intelligence.

[0035] (2) The format of the seismic exploration big data sample is unified, the distributed saving, efficient management and publishing of the sample are realized, and a good foundation is laid for the multi-scene application of the sample. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 The flowchart of a specific embodiment of the seismic exploration big data sample collection and labeling method of the present application is shown in the figure.

[0037] Figure 2 The system structure diagram of a specific embodiment of the seismic exploration big data sample collection and labeling device of the present application is shown in the figure. DETAILED DESCRIPTION

[0038] In order to make the features and advantages of the present application more obvious and easy to understand, a preferred embodiment is described in detail below with reference to the accompanying drawings.

[0039] The seismic exploration big data sample collection and labeling method of the present application uses the seismic exploration data of an oilfield, and constructs different big data sample sets according to different application scenarios. First, the seismic data is preprocessed, and on the basis of the big file splitting and merging technology based on the Hadoop distributed file system (HDFS), three different dimension redundant storages of the seismic data are generated to improve the collection efficiency of the seismic exploration big data sample. Then, the digital type big data sample collection and labeling is realized by setting the digital type sample parameters, drawing the sample collection boundary, collecting the sample and labeling the sample, and the picture type big data sample collection and labeling is realized by setting the picture type sample parameters, the sample size, collecting the sample and labeling the sample. Finally, the obtained seismic exploration big data sample is published to realize the sharing of the sample by multiple people and multiple applications.

[0040] The technical solution of the present invention is achieved through the following technical steps:

[0041] (1) Seismic data acquisition: Seismic data can be acquired from local files or OpenWorks interpretation results and loaded into the device.

[0042] (2) Seismic data preprocessing: Based on the large file segmentation and merging technology of Hadoop Distributed File System (HDFS), seismic data of three different dimensions, X, Y, and Z, are generated and redundantly stored to improve the efficiency of reading seismic exploration data and the efficiency of collecting seismic exploration big data samples.

[0043] (3) Collection of digital seismic exploration big data samples: interactively select the work area, seismic data volume, and survey line type, interactively set the sample scene, interactively draw rectangles and polygons to specify the sample collection boundary, save sample parameters and collection boundaries, and collect digital seismic exploration big data samples.

[0044] (4) Digital type seismic exploration big data sample labeling: Select the scene, select the seismic data volume, select the sample boundary list, specify the label, and label the digital type seismic exploration big data sample.

[0045] (5) Image type seismic exploration big data sample collection: interactively select the work area, seismic data volume, survey line type, interactively set the collection range, interactively set the sample size, characteristic value name, and conduct image type seismic exploration big data sample collection.

[0046] (6) Image type seismic exploration big data sample annotation: Select an image type data sample set, select one or more rows of samples, and perform image type seismic exploration big data sample annotation.

[0047] (7) Sample publishing and sharing: realize the management, uploading, exporting, and publishing of digital and image-type seismic exploration big data samples.

[0048] The following are several specific embodiments of the present invention.

[0049] Example 1:

[0050] In a specific embodiment 1 of the present invention, Figure 1 As shown, Figure 1 This is a flowchart of a specific embodiment of the method for collecting and labeling seismic exploration big data samples of the present invention. For digital seismic exploration big data samples, the method includes:

[0051] In step 110 , the seismic data for the work area is loaded. This can be loaded from a local file or from OpenWorks interpretation results. The process then proceeds to step 120 .

[0052] In step 120 , the seismic data is pre-processed using the Hadoop Distributed File System (HDFS) large file splitting and merging technology to generate seismic data in three different dimensions, X, Y, and Z, for redundant storage. The process then proceeds to step 130 .

[0053] In step 130 , the parameters of the digital sample are interactively set, including interactive selection of the work area, seismic data volume, survey line type, and interactive setting of the sample scene. The process proceeds to step 131 .

[0054] In step 131 , a digital sample collection boundary is interactively drawn, and both rectangular and polygonal drawing modes are supported. The process then proceeds to step 132 .

[0055] In step 132 , digital samples are collected based on the set sample parameters and the drawn sample collection boundary. The process then proceeds to step 133 .

[0056] In step 133 , digital samples are labeled by selecting a sample boundary list and assigning labels to implement digital sample labeling. The process then proceeds to step 150 .

[0057] In step 150, the seismic exploration big data sample is published. The management, uploading, exporting, and publishing of the seismic exploration big data sample are realized. The process ends.

[0058] Example 2:

[0059] In another specific embodiment 2 of the present invention, for a picture-type seismic exploration big data sample, the method further includes:

[0060] After step 120 , the process proceeds to step 140 .

[0061] In step 140 , the image type sample parameters are interactively set, including interactively selecting the work area, seismic data volume, and survey line type. The process proceeds to step 141 .

[0062] In step 141, the image type acquisition range and sample size are interactively set, including the range and step size of the main survey line (Inline), crossline (Crossline), and horizontal time slice (Time), as well as the sample size and feature value name. The process then proceeds to step 142.

[0063] In step 142 , a picture type sample is collected based on the set sample parameters and the set sample size. The process then proceeds to step 143 .

[0064] In step 143 , the image type samples are labeled, a image type data sample set is selected, samples to be calibrated are selected, and the image type seismic exploration big data samples are labeled. The process proceeds to step 150 .

[0065] Example 3:

[0066] In specific embodiment 3 of the present invention, Figure 2 This is a structural diagram of the seismic exploration big data sample collection and annotation device of the present invention. The seismic exploration big data sample collection and annotation device includes a seismic loading module 21, a seismic preprocessing module 22, a seismic sample automatic collection module 23, a collection boundary interactive drawing module 24, a sample size collection area setting module 25, a seismic sample annotation module 26, and a sample publishing module 27.

[0067] The seismic loading module 21 loads seismic data for the work area. The seismic preprocessing module 22 generates seismic data in three different dimensions: X, Y, and Z, for redundant storage. The automatic seismic sample acquisition module 23, connected to the seismic preprocessing module 22, the interactive acquisition boundary drawing module 24, the sample size acquisition area setting module 25, and the seismic sample annotation module 26, extracts seismic attributes based on the preprocessed seismic data and the interactively drawn digital big data sample boundaries to form digital big data samples, which are then transmitted to the seismic sample annotation module 26. Based on the preprocessed seismic data and the interactively set size and area of ​​the image big data samples, it extracts seismic images to form image big data samples, which are then transmitted to the seismic sample annotation module 26. The interactive acquisition boundary drawing module 24 interactively draws the acquisition boundaries for the digital samples. The sample size acquisition area setting module 25 interactively sets the sample size and acquisition area for the image big data samples. The seismic sample annotation module 26 annotates the collected digital and image big data samples. The sample publishing module 27 manages, uploads, exports, and publishes seismic exploration big data samples.

[0068] Finally, it should be noted that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art may modify the technical solutions described in the aforementioned embodiments or substitute equivalents for some of the technical features therein. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

[0069] Except for the technical features described in the specification, all other technical features are known technologies to those skilled in the art.

Claims

1. A method for collecting and labeling large seismic exploration data samples, characterized in that: The seismic exploration big data sample collection and annotation method includes: Step 1: Load the seismic data of the work area and preprocess the seismic data; Step 2: interactively set the sample parameters of the digital type and the sample parameters of the image type; Step 3: interactively draw the sample collection boundary for digital type, and interactively set the collection range and sample size for image type; Step 4: Collect digital type samples and picture type samples; Step 5: Label the digital type samples and the picture type samples; Step 6: Release the seismic exploration big data sample; In step 2, interactively select the work area, seismic data volume, and survey line type, and interactively set the sample scene; In step 3, when interactively drawing the digital type sample collection boundary, the sample collection boundary can be drawn in two ways: rectangle and polygon; In step 3, interactively set the image type acquisition range and sample size, including interactively setting the range and step size of the main survey line Inline, the crossline, and the horizontal time slice Time, and interactively set the sample size and eigenvalue name.

2. The method for collecting and labeling large seismic exploration data samples according to claim 1, characterized in that: In step 1, based on the large file segmentation and merging technology, seismic data of three different dimensions, X, Y, and Z, are generated and redundantly stored to improve the efficiency of reading seismic exploration data and the efficiency of collecting seismic exploration big data samples.

3. The method for collecting and labeling large seismic exploration data samples according to claim 1, characterized in that: In step 4, digital type sample collection and picture type sample collection are implemented based on the set sample parameters and the drawn sample collection boundary.

4. The method for collecting and labeling large seismic exploration data samples according to claim 1, characterized in that: In step 5, select a scene, select a seismic data volume, select a sample boundary list, specify a label, and perform digital type seismic exploration big data sample labeling.

5. The method for collecting and labeling large seismic exploration data samples according to claim 1, characterized in that: In step 5, select a picture type data sample set, select the sample to be calibrated, and perform image type seismic exploration big data sample labeling.

6. The method for collecting and labeling large seismic exploration data samples according to claim 1, characterized in that: In step 6, the seismic exploration big data samples are managed, uploaded, exported, and published.

7. A device for use in the method for collecting and labeling large seismic exploration data samples as claimed in claim 1, characterized in that: The seismic exploration big data sample collection and annotation device includes: The earthquake pre-processing module generates earthquake data in three different dimensions: X, Y, and Z, and stores them redundantly. Collection boundary interactive drawing module, interactively drawing the collection boundary of digital type samples; The sample size collection area setting module interactively sets the sample size and collection area of ​​the image type big data sample; Seismic sample annotation module, which annotates the collected digital and image type big data samples; The automatic seismic sample acquisition module is connected to the seismic preprocessing module, the acquisition boundary interactive drawing module, the sample size acquisition area setting module and the seismic sample labeling module. Based on the preprocessed seismic data and the interactively drawn digital type big data sample boundaries, the module extracts seismic attributes to form digital type big data samples and transmits them to the seismic sample labeling module; based on the preprocessed seismic data and the size and area of ​​the interactively set picture type big data samples, the module extracts seismic images to form picture type big data samples and transmits them to the seismic sample labeling module.

8. The seismic exploration big data sample collection and annotation device according to claim 7, characterized in that: The seismic exploration big data sample collection and annotation device also includes a seismic loading module, which is connected to the seismic preprocessing module and loads seismic data of the work area.

9. The seismic exploration big data sample collection and annotation device according to claim 7, characterized in that: The seismic exploration big data sample collection and annotation device also includes a sample publishing module, which is connected to the seismic sample annotation module to manage, upload, export and publish the seismic exploration big data samples.

Images