Super resolution with deep learning for seismic image resolution enhancement

Abstract

A method for enhancing a resolution of a seismic image with reference well log measurements includes receiving a first low-resolution seismic image. The method also includes generating one or more pseudo well logs based upon the first low-resolution seismic image. The method also includes generating synthetic seismic traces. The synthetic seismic traces include (1) synthetic low-resolution seismic traces that are based upon the first low-resolution seismic image and the one or more pseudo well logs and (2) synthetic high-resolution seismic traces that are based upon a targeted high-resolution seismic image and the one or more pseudo well logs. The method also includes training a convolutional neural network to map the synthetic low-resolution seismic traces to the synthetic high-resolution seismic traces. The method also includes generating the targeted high-resolution seismic image using the trained convolutional neural network.

Description

SUPER RESOLUTION WITH DEEP LEARNING EOR SEISMIC IMAGERESOLUTION ENHANCEMENTCross-Reference to Related Applications

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 620,418, filed on January 12, 2024, which is hereby incorporated by reference in its entirety.Background

[0002] Seismic images are used extensively for understanding subsurface environments. Useful information may be extracted from fault and horizon interpretations, facies modeling, seismic attribute analysis, and property predictions.

[0003] The quality of the seismic images is dependent of several aspects, among which are the sampling rate, which is the regular time interval (e.g., millisecond) each signal is recorded to by the instrument, and the intensity of the amplitudes in comparison with the background environment. A geological layer may be defined on the recorded seismic section by a pair of positive (e g., peak) and negative (e.g., trough) events, where the thickness of the layer is proportional to the length of this pair. The vertical resolution is defined as the finest feature that can be observed on the seismic event.

[0004] Unfortunately, while traveling through Earth layers, the seismic amplitude is altered by the filtering effect. The high-frequency content, which determines the vertical resolutions of the seismic records, is suppressed when acoustic signals pass through absorptive layers. The recorded signals, therefore, lack information that may be used to identify the fine events. Furthermore, the recorded seismic data are bandlimited. The high-frequency components in these bandlimited data may be further removed before being fed to the seismic data processing algorithms due to computation resource limitations, resulting in relatively low-resolution subsurface seismic images. Accordingly, a solution is needed to overcome one or more of the deficiencies identified above.Summary

[0005] A method for enhancing a resolution of a seismic image with reference well log measurements is disclosed. The method includes receiving a first low-resolution seismic image. The method also includes generating one or more pseudo well logs based upon the first low-resolution seismic image. The method also includes generating synthetic seismic traces. The synthetic seismic traces include (1) synthetic low-resolution seismic traces that are based upon the first low-resolution seismic image and the one or more pseudo well logs and (2) synthetic high- resolution seismic traces that are based upon a targeted high-resolution seismic image and the one or more pseudo well logs. The method also includes training a convolutional neural network to map the synthetic low-resolution seismic traces to the synthetic high-resolution seismic traces. The method also includes generating the targeted high-resolution seismic image using the trained convolutional neural network.

[0006] A computing system is also disclosed. The computing system includes one or more processors and a memory system. The memory system includes one or more non-transitory computer-readable media storing instructions that, when executed by at least one of the one or more processors, cause the computing system to perform operations. The operations include receiving one or more well logs corresponding to one or more wells. The operations also include receiving a first low-resolution seismic image, which covers a trajectory of the one or more wells. The operations also include pre-processing the one or more well logs to produce pre-processed well logs. The operations also include pre-processing the first low-resolution seismic image to produce a pre-processed low-resolution seismic image. The operations also include generating pseudo well logs based upon the pre-processed well logs and the pre-processed low-resolution seismic image. The pseudo well logs are generated by converting at least a portion of the pre- processed low-resolution seismic image into the pseudo well logs using a convolutional neural network guided by the pre-processed well logs. The operations also include generating synthetic seismic traces. The synthetic seismic traces include synthetic low-resolution seismic traces that are based upon the pre-processed low-resolution seismic image and the pseudo well logs. The synthetic seismic traces also include synthetic high-resolution seismic traces that are based upon a targeted high-resolution seismic image and the pseudo well logs. The operations also include training the convolutional neural network to map the synthetic low-resolution seismic traces to the synthetic high-resolution seismic traces. The operations also include generating the targeted high- resolution seismic image using the trained convolutional neural network. The targeted high- resolution seismic image is generated by inputting traces from the pre-processed low-resolution seismic image or a second low-resolution seismic image into the trained convolutional neural network.

[0007] A non-transitory computer-readable medium is also disclosed. The medium stores instructions that, when executed by one or more processors of a computing system, cause the computing system to perform operations. The operations include receiving well logs corresponding to one or more wells. The operations also include receiving a first low-resolution seismic image, which covers a trajectory of the one or more wells. The operations also include pre-processing the well logs to produce pre-processed well logs. The operations also include preprocessing the first low-resolution seismic image to produce a pre-processed low-resolution seismic image. The operations also include generating pseudo well logs based upon the pre- processed well logs and the pre-processed low-resolution seismic image. The pseudo well logs are generated by converting at least a portion of the pre-processed low-resolution seismic image into the pseudo well logs using a convolutional neural network guided by the pre-processed well logs. The pseudo well logs include reflection coefficients that represent one or more acoustic properties of subsurface layers. The one or more acoustic properties include an acoustic wave impedance. The operations also include generating synthetic seismic traces. The synthetic seismic traces include synthetic low-resolution seismic traces that are based upon the pre-processed low- resolution seismic image and the pseudo well logs. The synthetic seismic traces also include synthetic high-resolution seismic traces that are based upon a targeted high-resolution seismic image and the pseudo well logs. The synthetic low-resolution seismic traces and the synthetic high-resolution seismic traces include reflection seismic signals. The synthetic low-resolution seismic traces and the synthetic high-resolution seismic traces are generated by applying first and second low-pass filters to the pseudo well logs. The first low-pass filter include a first cutoff frequency that is determined based upon the pre-processed low-resolution seismic image. The second low-pass filter includes a second cutoff frequency that is determined based upon the targeted high-resolution seismic image. The second cutoff frequency is less than a Nyquist frequency and greater than a highest frequency of the first low-resolution seismic image. The synthetic low-resolution seismic traces and the synthetic high-resolution seismic traces are also or instead generated by convolving a pre-designed low-frequency seismic wavelet with the pseudo well logs to produce the synthetic low-resolution seismic traces, and convolving a pre-designed high-frequency seismic wavelet with the pseudo well logs to produce the synthetic high-resolution seismic traces. The pre-designed low-frequency seismic wavelet is extracted from the pre- processed low-resolution seismic image. The pre-designed high-frequency seismic wavelet isdetermined based upon the targeted high-resolution seismic image. The operations also include training the convolutional neural network to map the synthetic low-resolution seismic traces to the synthetic high-resolution seismic traces. The operations also include generating the targeted high- resolution seismic image using the trained convolutional neural network. The targeted high- resolution seismic image is generated by inputting traces from the pre-processed low-resolution seismic image or a second low-resolution seismic image into the trained convolutional neural network. The second low-resolution seismic image is from a same seismic survey as the first low- resolution seismic image.

[0008] It will be appreciated that this summary is intended merely to introduce some aspects of the present methods, systems, and media, which are more fully described and / or claimed below. Accordingly, this summary is not intended to be limiting.Brief Description of the Drawings

[0009] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the present teachings and together with the description, serve to explain the principles of the present teachings. In the figures:

[0010] Figure 1 illustrates an example of a system that includes various management components to manage various aspects of a geologic environment, according to an embodiment.

[0011] Figure 2 is a flowchart of a method for enhancing resolution of a seismic image with reference well log measurements, according to an embodiment.

[0012] Figure 3 illustrates a schematic view of the method of Figure 2, according to an embodiment.

[0013] Figure 4 is a flowchart of another method for enhancing resolution of a seismic image with reference well log measurements, according to an embodiment.

[0014] Figure 5 illustrates input data including one or more well logs and a first low-resolution seismic image, according to an embodiment.

[0015] Figure 6 illustrates a schematic view of pseudo well log generation (e.g., network training stage), according to an embodiment.

[0016] Figure 7 illustrates a schematic view of pseudo well log generation (e.g., network inferencing stage), according to an embodiment.

[0017] Figure 8 illustrates a first example of generating low-resolution seismic traces and high- resolution seismic traces, according to an embodiment.

[0018] Figure 9 illustrates a second example of generating the low-resolution seismic traces and the high-resolution seismic traces, according to an embodiment.

[0019] Figure 10 illustrates a schematic view of training convolutional neural network, according to an embodiment.

[0020] Figure 11 illustrates a schematic view of generating targeted high-resolution seismic image, according to an embodiment.

[0021] Figure 12 is a flowchart of a method for enhancing resolution of a seismic image with referenced high-resolution 2D image(s), according to an embodiment.

[0022] Figure 13 illustrates a schematic view of the method of Figure 12, according to an embodiment.

[0023] Figures 14A and 14B illustrate the amplitude maps extracted at top of reservoir to show the improvements of amplitude contrasts after ML application in the first method (Figures 2-11), according to an embodiment.

[0024] Figure 15 illustrates the improvements of high frequency content in the ML enhanced image in the first method (Figures 2-11), according to an embodiment.

[0025] Figures 16A and 16B illustrate the comparison of seismic sections after high-pass frequency filtering to show the extra fine features generated by the ML are real and fully supported by the background geology in the first method (Figures 2-11), according to an embodiment.

[0026] Figure 17 illustrates a schematic view of a computing system for performing at least a portion of the method(s) described herein, according to an embodiment.Detailed Description

[0027] Reference will now be made in detail to embodiments, examples of which are illustrated in the accompanying drawings and figures. In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be apparent to one of ordinary skill in the art that the invention may be practiced without these specific details. In other instances, well-known methods, procedures, components, circuits, and networks have not been described in detail so as not to unnecessarily obscure aspects of the embodiments.

[0028] It will also be understood that, although the terms first, second, etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another. For example, a first object or step could be termed a second object or step, and, similarly, a second object or step could be termed a first object or step, without departing from the scope of the present disclosure. The first object or step, and the second object or step, are both, objects or steps, respectively, but they are not to be considered the same object or step.

[0029] The terminology used in the description herein is for the purpose of describing particular embodiments and is not intended to be limiting. As used in this description and the appended claims, the singular forms “a,” “an” and “the” are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will also be understood that the term “and / or” as used herein refers to and encompasses any possible combinations of one or more of the associated listed items. It will be further understood that the terms “includes,” “including,” “comprises” and / or “comprising,” when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. Further, as used herein, the term “if’ may be construed to mean “when” or “upon” or “in response to determining” or “in response to detecting,” depending on the context.

[0030] Attention is now directed to processing procedures, methods, techniques, and workflows that are in accordance with some embodiments. Some operations in the processing procedures, methods, techniques, and workflows disclosed herein may be combined and / or the order of some operations may be changed.

[0031] The smallest identifiable events on the seismic sections may be in meter scale. Meanwhile, well logs recorded from instruments along the wellbores provide information of much higher resolution (e.g., feet or inches), due to the very fine acquisition interval. A machine learning neural network can be designed to recover the high-frequency content of seismic signals by exploiting the information from well log data at wells located within the seismic boundary. The direct mapping from the low-resolution traces extracted from the seismic image to the high- resolution well log data using a neural network often introduces bias and errors when the number of available well logs is limited (e.g., tens of wells or even fewer). The machine learning workflow disclosed herein is designed to handle these scenarios. This workflow is applicable to improve theseismic vertical resolution by retrieving the high -resolution information from a few available well logs without suffering from the well-known generalization problem of supervised learning when the training data are not sufficiently representative due to the limited number of samples.System Overview

[0032] Figure 1 illustrates an example of a system 100 that includes various management components 110 to manage various aspects of a geologic environment 150 (e.g., an environment that includes a sedimentary basin, a reservoir 151, one or more faults 153-1, one or more geobodies 153-2, etc.). For example, the management components 110 may allow for direct or indirect management of sensing, drilling, injecting, extracting, etc., with respect to the geologic environment 150. In turn, further information about the geologic environment 150 may become available as feedback 160 (e.g., optionally as input to one or more of the management components 110).

[0033] In the example of Figure 1, the management components 110 include a seismic data component 112, an additional information component 114 (e.g., well / logging data), a processing component 116, a simulation component 120, an attribute component 130, an analysis / visualization component 142 and a workflow component 144. In operation, seismic data and other information provided per the components 112 and 114 may be input to the simulation component 120.

[0034] In an example embodiment, the simulation component 120 may rely on entities 122. Entities 122 may include earth entities or geological objects such as wells, surfaces, bodies, reservoirs, etc. In the system 100, the entities 122 can include virtual representations of actual physical entities that are reconstructed for purposes of simulation. The entities 122 may include entities based on data acquired via sensing, observation, etc. (e.g., the seismic data 112 and other information 114). An entity may be characterized by one or more properties (e.g., a geometrical pillar grid entity of an earth model may be characterized by a porosity property). Such properties may represent one or more measurements (e.g., acquired data), calculations, etc.

[0035] In an example embodiment, the simulation component 120 may operate in conjunction with a software framework such as an object-based framework. In such a framework, entities may include entities based on pre-defined classes to facilitate modeling and simulation. A commercially available example of an object-based framework is the MICROSOFT® .NET®framework (Redmond, Washington), which provides a set of extensible object classes. In the .NET® framework, an object class encapsulates a module of reusable code and associated data structures. Object classes can be used to instantiate object instances for use in by a program, script, etc. For example, borehole classes may define objects for representing boreholes based on well data.

[0036] In the example of Figure 1, the simulation component 120 may process information to conform to one or more attributes specified by the attribute component 130, which may include a library of attributes. Such processing may occur prior to input to the simulation component 120 (e.g., consider the processing component 116). As an example, the simulation component 120 may perform operations on input information based on one or more attributes specified by the attribute component 130. In an example embodiment, the simulation component 120 may construct one or more models of the geologic environment 150, which may be relied on to simulate behavior of the geologic environment 150 (e.g., responsive to one or more acts, whether natural or artificial). In the example of Figure 1, the analysis / visualization component 142 may allow for interaction with a model or model-based results (e.g., simulation results, etc.). As an example, output from the simulation component 120 may be input to one or more other workflows, as indicated by a workflow component 144.

[0037] As an example, the simulation component 120 may include one or more features of a simulator such as the ECLIPSE™ reservoir simulator (SLB, Houston Texas), the INTERSECT1''1reservoir simulator (SLB, Houston Texas), etc. As an example, a simulation component, a simulator, etc. may include features to implement one or more meshless techniques (e.g., to solve one or more equations, etc.). As an example, a reservoir or reservoirs may be simulated with respect to one or more enhanced recovery techniques (e.g., consider a thermal process such as SAGD, etc ).

[0038] In an example embodiment, the management components 110 may include features of a commercially available framework such as the PETREL® seismic to simulation software framework (SLB, Houston, Texas). The PETREL® framework provides components that allow for optimization of exploration and development operations. The PETREL® framework includes seismic to simulation software components that can output information for use in increasing reservoir performance, for example, by improving asset team productivity. Through use of such a framework, various professionals (e.g., geophysicists, geologists, and reservoir engineers) candevelop collaborative workflows and integrate operations to streamline processes. Such a framework may be considered an application and may be considered a data-driven application (e.g., where data is input for purposes of modeling, simulating, etc.).

[0039] In an example embodiment, various aspects of the management components 110 may include add-ons or plug-ins that operate according to specifications of a framework environment. For example, a commercially available framework environment marketed as the OCEAN® framework environment (SLB, Houston, Texas) allows for integration of add-ons (or plug-ins) into a PETREL® framework workflow. The OCEAN® framework environment leverages .NET® tools (Microsoft Corporation, Redmond, Washington) and offers stable, user-friendly interfaces for efficient development. In an example embodiment, various components may be implemented as add-ons (or plug-ins) that conform to and operate according to specifications of a framework environment (e.g., according to application programming interface (API) specifications, etc.).

[0040] Figure 1 also shows an example of a framework 170 that includes a model simulation layer 180 along with a framework services layer 190, a framework core layer 195 and a modules layer 175. The framework 170 may include the commercially available OCEAN® framework where the model simulation layer 180 is the commercially available PETREL® model-centric software package that hosts OCEAN® framework applications. In an example embodiment, the PETREL® software may be considered a data-driven application. The PETREL® software can include a framework for model building and visualization.

[0041] As an example, a framework may include features for implementing one or more mesh generation techniques. For example, a framework may include an input component for receipt of information from interpretation of seismic data, one or more attributes based at least in part on seismic data, log data, image data, etc. Such a framework may include a mesh generation component that processes input information, optionally in conjunction with other information, to generate a mesh.

[0042] In the example of Figure 1, the model simulation layer 180 may provide domain objects 182, act as a data source 184, provide for rendering 186 and provide for various user interfaces 188. Rendering 186 may provide a graphical environment in which applications can display their data while the user interfaces 188 may provide a common look and feel for application user interface components.

[0043] As an example, the domain objects 182 can include entity objects, property objects and optionally other objects. Entity objects may be used to geometrically represent wells, surfaces, bodies, reservoirs, etc., while property objects may be used to provide property values as well as data versions and display parameters. For example, an entity object may represent a well where a property object provides log information as well as version information and display information (e.g., to display the well as part of a model).

[0044] In the example of Figure 1, data may be stored in one or more data sources (or data stores, generally physical data storage devices), which may be at the same or different physical sites and accessible via one or more networks. The model simulation layer 180 may be configured to model projects. As such, a particular project may be stored where stored project information may include inputs, models, results and cases. Thus, upon completion of a modeling session, a user may store a project. At a later time, the project can be accessed and restored using the model simulation layer 180, which can recreate instances of the relevant domain objects.

[0045] In the example of Figure 1, the geologic environment 150 may include layers (e.g., stratification) that include a reservoir 151 and one or more other features such as the fault 153-1, the geobody 153-2, etc. As an example, the geologic environment 150 may be outfitted with any of a variety of sensors, detectors, actuators, etc. For example, equipment 152 may include communication circuitry to receive and to transmit information with respect to one or more networks 155. Such information may include information associated with downhole equipment 154, which may be equipment to acquire information, to assist with resource recovery, etc. Other equipment 156 may be located remote from a well site and include sensing, detecting, emitting or other circuitry. Such equipment may include storage and communication circuitry to store and to communicate data, instructions, etc. As an example, one or more satellites may be provided for purposes of communications, data acquisition, etc. For example, Figure 1 shows a satellite in communication with the network 155 that may be configured for communications, noting that the satellite may additionally or instead include circuitry for imagery (e.g., spatial, spectral, temporal, radiometric, etc.).

[0046] Figure 1 also shows the geologic environment 150 as optionally including equipment 157 and 158 associated with a well that includes a substantially horizontal portion that may intersect with one or more fractures 159. For example, consider a well in a shale formation that may include natural fractures, artificial fractures (e.g., hydraulic fractures) or a combination ofnatural and artificial fractures. As an example, a well may be drilled for a reservoir that is laterally extensive. In such an example, lateral variations in properties, stresses, etc. may exist where an assessment of such variations may assist with planning, operations, etc. to develop a laterally extensive reservoir (e.g., via fracturing, injecting, extracting, etc.). As an example, the equipment 157 and / or 158 may include components, a system, systems, etc. for fracturing, seismic sensing, analysis of seismic data, assessment of one or more fractures, etc.

[0047] As mentioned, the system 100 may be used to perform one or more workflows. A workflow may be a process that includes a number of worksteps. A workstep may operate on data, for example, to create new data, to update existing data, etc. As an example, a may operate on one or more inputs and create one or more results, for example, based on one or more algorithms. As an example, a system may include a workflow editor for creation, editing, executing, etc. of a workflow. In such an example, the workflow editor may provide for selection of one or more predefined worksteps, one or more customized worksteps, etc. As an example, a workflow may be a workflow implementable in the PETREL® software, for example, that operates on seismic data, seismic attribute(s), etc. As an example, a workflow may be a process implementable in the OCEAN® framework. As an example, a workflow may include one or more worksteps that access a module such as a plug-in (e.g., external executable code, etc.).Exemplary Methods

[0048] The present disclosure includes two multi-stage methods for different problems / scenarios. The first method is designed to enhance the vertical resolution of an input 3D seismic volume using referenced well log measurements. The first method enhances a seismic volume, and several well with property logs (e.g., acoustic log and density log) whose locations and trajectories are covered within the seismic volume. Meanwhile, the second method is designed to enhance the vertical resolution of an input 3D seismic volume using referenced high-resolution image(s). Thus, the second method enhances a seismic volume, and one or several 2D images of higher resolution than input 3D volume.First Method (Embodiment One)

[0049] Figure 2 is a flowchart of a method 200 for enhancing resolution of a seismic image with reference well log measurements, according to an embodiment. This method 200 involves aseismic volume to be enhanced, and several well with property logs (e.g., acoustic log and density log) whose locations and trajectories are covered within the seismic volume. An illustrative order of the method 200 is provided below; however, one or more portions of the method 200 may be performed in a different order, simultaneously, repeated, or omitted. Figure 3 illustrates a schematic view of the method 200 for enhancing resolution of a seismic image with reference well log measurements using machine-learning algorithms, according to an embodiment.

[0050] The method 200 may include receiving a well log and a low-resolution seismic image, as at 210. The low-resolution seismic image covers a trajectory of the well log.

[0051] The method 200 may also include pre-processing the well log and the low-resolution seismic image to produce a pre-processed well log and a pre-processed low-resolution image, as at 220. This may be referred to as Stage 0. The pre-processing may include implementing well log quality control (QC) to remove outliers and to fix other errors in the well log. The preprocessing may also or instead include implement well log reconstruction when the data in the well log is partially missing. The pre-processing may also or instead include implementing a welltie process. The pre-processing may also or instead include pre-conditioning of seismic images to remove noises. This may include performing a depth-to-time conversion if the seismic images are in the depth domain.

[0052] The method 200 may also include generating pseudo well logs based upon the pre- processed well log and the pre-processed low-resolution seismic image, as at 230. This may be referred to as Stage 1, as shown in Figure 3. In one embodiment, the purpose of Stage 1 is to tackle the data limitation when less than a predetermined amount of well logs are available within the seismic dataset by creating pseudo well logs to represent the geological features of the targeted subsurface environment. This may be performed by implementing a deep learning network to capture the approximate relationship between the seismic traces extracted at the well locations and the acoustic and density logs measured at the well locations, and then populating this relationship away from the well to generate pseudo log curves. These pseudo logs may be an approximation to the real logs. An encoder-decoder format network structure may be implemented, where the encoder performs the mapping from the extracted seismic trace to the well log curve, and the decoder performs the mapping from the well log curve to the input seismic trace. The encoderdecoder format may ensure that the log curves generated with the encoder preserve the relationship between the seismic and log curves at the well locations.

[0053] Additionally, Stage 1 (see Figure 3) may also include a supervised machine-learning (ML) algorithm that inputs the extracted seismic trace along the well bores, and outputs (a) a seismic trace with the same dimension(s) and sampling interval(s) with the input seismic trace, and / or (b) the reflection coefficient log or a log curve that can be used to derive the acoustic reflection coefficients at the same or finer sampling interval containing higher frequency content (e.g., acoustic impedance). Both encoder and decoder structures may include multiple layers of convolutional neural networks for feature extractions. Illustrative networks include the U-Net, deep convolutional network, cGAN, GAN, etc.

[0054] The encoder may input the extracted seismic traces along the wellbore, and output a log that is fed into the decoder to generate a reconstructed seismic trace. The network optimization process may be performed iteratively to update the parameters (e.g., weights and / or biases) of both the encoder and decoders to minimize a loss function that measures the difference between generated values and target labels (e.g., log values and input seismic traces). The successfully trained network may accurately generate the log values from the encoder and fully reconstruct the input seismic trace from the decoder at the locations where the few well logs are available.

[0055] The method 200 may also include generating synthetic low-resolution seismic traces and / or synthetic high-resolution seismic traces based upon the pseudo well logs, as at 240. This may be referred to as Stage 2 in Figure 3. The synthetic seismic traces may be generated using the model trained in Stage I (see Figure 3). The seismic traces may be at the pseudo-well locations of the pseudo well logs. In one embodiment, the purpose of Stage 2 in Figure 3 is to generate extra training data in other locations where no well logs are available by applying the trained network in Stage 1 in Figure 3. Many seismic traces from a subset of the input seismic volumes may be input into the network in Stage 1 in Figure 3 to create the synthetic well log curves and the corresponding synthetic seismic traces.

[0056] The pseudo well log curves may later be used to generate the reflection coefficient curves, and convolved with a pre-designed high-resolution wavelet to create synthetic high-resolution seismic traces. The pseudo reflection coefficient curves may also be convolved with a predesigned low-resolution wavelet to create synthetic low-resolution seismic traces. In another embodiment, the synthetic low-resolution traces may be obtained from the output of the network. The pairs of the low-resolution traces and the high-resolution traces may be used for the second network training in Stage 3.

[0057] The method 200 may also include mapping the high-resolution seismic traces, as at 250. This may be referred to as Stage 3 in Figure 3. Stage 3 may include a supervised machine-learning algorithm that inputs the many pairs of synthetic low-resolution seismic traces and the synthetic high-resolution traces generated in Stage 2 in Figure 3. The synthetic low-resolution traces may be the input to the network, and the high-resolution traces serve as the ground truth (i.e., label) for the network training. The algorithm can be any trace-to-trace or image-to-image transformation structure, such as U-Net, deep convolution network, cGan, GAN, etc. The network optimization process may be performed iteratively to update the hyper-parameters (e.g., weights and bias) of the network to minimize a loss function that measures the difference between network’s generated values at each iteration and the label (e.g., high-resolution traces).

[0058] The method 200 may also include generating a higher-resolution version of the low- resolution seismic image, as at 260. The higher-resolution version may be generated based upon the mapping between the low-resolution trace and the high-resolution trace established by the network trained above.

[0059] The method 200 may also include displaying the outputs, as at 270. This may be or include the pseudo well logs, the synthetic low-resolution seismic traces, the synthetic high- resolution seismic traces, the mapped high-resolution seismic traces, the higher-resolution version of the well log (e.g., seismic image), or a combination thereof.

[0060] The method 200 may also include performing a wellsite action, as at 280. The wellsite action may be based upon the pseudo well logs, the synthetic low-resolution seismic traces, the synthetic high-resolution seismic traces, the mapped high-resolution seismic traces, the higher- resolution version of the well log (e.g., seismic image), or a combination thereof. The wellsite action may be or include generating and / or transmitting a signal (e.g., using a computing system) that causes a physical action to occur at a wellsite. The wellsite action may also or instead include performing the physical action at the wellsite. The physical action may include selecting where to drill a wellbore, drilling the wellbore, varying a weight and / or torque on a drill bit that is drilling the wellbore, varying a drilling trajectory of the wellbore, varying a concentration and / or flow rate of a fluid pumped into the wellbore, or the like.First Method (Embodiment Two)

[0061] Figure 4 is a flowchart of a method 400 for enhancing resolution of a seismic image with reference well log measurements, according to an embodiment. An illustrative order of the method 400 is provided below; however, one or more portions of the method 400 may be performed in a different order, simultaneously, repeated, or omitted. Figures 5-11 illustrate schematic views of the method 400 for enhancing resolution of a seismic image with reference well log measurements using machine-learning algorithms, according to an embodiment.

[0062] The method 400 may include receiving a first well log corresponding to a first well, as at 405. Figure 5 illustrates input data including one or more well logs (two are shown: 510A, 510B), according to an embodiment.

[0063] The method 400 may also include receiving a first low-resolution seismic image, as at 410. Figure 5 also illustrates input data including a first low-resolution seismic image 520, according to an embodiment. The first low-resolution seismic image 520 may include a portion of the first well (e.g., a trajectory of the first well).

[0064] The method 400 may also include pre-processing the first well log(s) 510A, 510B to produce one or more pre-processed well log(s), as at 415.

[0065] The method 400 may also include pre-processing the first low-resolution seismic image 520 to produce a pre-processed low-resolution seismic image, as at 420.

[0066] The method 400 may also include generating a pseudo well log (i.e., pseudo reflection coefficient profiles) based upon the pre-processed well log and the pre-processed low-resolution seismic image, as at 425. The pseudo well logs may be generated by converting at least a portion of the pre-processed low-resolution seismic image 520 into the pseudo well logs using a convolutional neural network (CNN) guided by the pre-processed well logs 510A, 510B. The pseudo well log may include one or more reflection coefficients that represent one or more acoustic properties of subsurface layers. The one or more acoustic properties may be or include an acoustic wave impedance.

[0067] Figure 6 illustrates a schematic view of pseudo well log generation (e.g., network training stage), according to an embodiment. One or more low-resolution seismic traces 610A, 610B may be extracted from the (e.g., pre-processed) low-resolution seismic image 520 at the well locations. The traces 610A, 610B may be used to train the convolutional neural network 620 bymapping the low-resolution seismic traces 61 OA, 61 OB to the corresponding (e.g., pre-processed) well logs 510A, 51 OB.

[0068] Figure 7 illustrates a schematic view of pseudo well log generation (e.g., network inferencing stage), according to an embodiment. One or more additional low-resolution seismic traces 61 OC-61 OF may be extracted from the low-resolution seismic image 520 at random locations (e.g., not at the well locations). The low-resolution seismic traces 610C-610F may be input into the trained convolutional neural network 620, which may then output the corresponding pseudo well logs 710A-710D.

[0069] The method 400 may also include generating low-resolution seismic traces, as at 430. The low-resolution seismic traces may be based upon the (e.g., pre-processed) low-resolution seismic image 520 and / or the pseudo-well logs 710A-710D.

[0070] The method 400 may also include generating synthetic high-resolution seismic traces, as at 435. The synthetic high-resolution seismic traces may be based upon a targeted high-resolution seismic image and / or the pseudo-well logs 710A-710D. The synthetic low-resolution seismic traces and / or the synthetic high-resolution seismic traces may include reflection seismic signals.

[0071] Figure 8 illustrates a first example of generating the low-resolution seismic traces 810A- 810D and the high-resolution seismic traces 820A-820D, according to an embodiment. The synthetic low-resolution seismic traces 810A-810D and the synthetic high-resolution seismic traces 820A-820D may be generated by applying first and second low-pass filters 830, 840 to the pseudo well logs 710A-710D. The first low-pass filter 830 may be or include a first cutoff frequency that is determined based upon the (e.g., pre-processed) low-resolution seismic image 520. The second low-pass filter 840 may be or include a second cutoff frequency that is determined based upon (e.g., a bandwidth of) the targeted high-resolution seismic image. The second cutoff frequency may be less than the Nyquist frequency and / or greater than a highest frequency of the first low-resolution seismic image 520.

[0072] Figure 9 illustrates a second example of generating the low-resolution seismic traces 810A-810D and the high-resolution seismic traces 820A-820D, according to an embodiment. The synthetic low-resolution seismic traces 810A-810D and the synthetic high-resolution seismic traces 820A-820D may be generated by convolving a pre-designed low-frequency seismic wavelet 910 with the pseudo well logs 710A-710D to produce the synthetic low-resolution seismic traces 810A-810D, and convolving a pre-designed high-frequency seismic wavelet 920 with the pseudowell logs 710A-710D to produce the synthetic high-resolution seismic traces 820A-820D. The pre-designed low-frequency seismic wavelet 910 may be extracted from the (e.g., pre-processed) low-resolution seismic image 520. The pre-designed high-frequency seismic wavelet 920 may be determined based upon the targeted high-resolution seismic image.

[0073] The method 400 may also include training the convolutional neural network 620 to map the synthetic low-resolution seismic traces 810A-810D to the synthetic high-resolution seismic traces 820A-820D, as at 440. Figure 10 illustrates a schematic view of training the convolutional neural network 620, according to an embodiment. The input of the convolutional neural network 620 may be the low-resolution seismic traces 810A-810D, and the output of the convolutional neural network 620 may be the high-resolution seismic traces 820A-820D.

[0074] The method 400 may also include generating the targeted high-resolution seismic image using the trained convolutional neural network 620, as at 445. The targeted high-resolution seismic image may be generated by inputting traces 610C-610F from the pre-processed low- resolution seismic image 520 or a second low-resolution seismic image into the trained convolutional neural network 620. The second low-resolution seismic image may be from a same (or similar) seismic survey as the first low-resolution seismic image.

[0075] Figure 11 illustrates a schematic view of generating the targeted high-resolution seismic image 1110, according to an embodiment. As mentioned above, the low-resolution seismic traces 610C-610F may be extracted from the low-resolution seismic image 520 one by one and input into the trained convolutional neural network 620 to predict corresponding high-resolution seismic traces 1110A-1 HOD. Eventually, the predicted high-resolution seismic traces 1110A-1110D may be merged together to produce the targeted high-resolution seismic image 1120.

[0076] The method 400 uses a portion of the low-resolution traces 610C-610F extracted from the low -resolution seismic image 520, because the purpose is to train the network. In Figure 11, one objective is to generate a corresponding high-resolution image 1120. In other words, for every low-resolution trace 610C-610F extracted from the low-resolution seismic image 520, the method 400 may generate the corresponding high-resolution trace 1110A-1 HOD, so the method 400 can merge these traces 1110A-1 HOD to form a complete high-resolution seismic image 1120 with the same size of the low-resolution seismic image 520.

[0077] The method 400 may also include displaying the targeted high-resolution seismic image 1120, as at 450.

[0078] The method 400 may also include performing a wellsite action based upon or in response to the targeted high-resolution seismic image 1120, as at 455. The wellsite action may be or include generating and / or transmitting a signal (e.g., using a computing system) that instructs or causes a physical action to occur at a wellsite. The wellsite action may also or instead include performing the physical action at the wellsite. The physical action may include selecting where to drill a wellbore, drilling the wellbore, varying a weight and / or torque on a drill bit that is drilling the wellbore, varying a drilling trajectory of the wellbore, varying a concentration and / or flow rate of a fluid pumped into the wellbore, or the like.Second Method

[0079] Figure 12 is a flowchart of a method 1200 for enhancing resolution of a seismic image with referenced high-resolution 2D image(s), according to an embodiment. This method 1200 involves 2D high-resolution image(s) and the 3D volume to be resolution enhanced. An illustrative order of the method 1200 is provided below; however, one or more portions of the method 1200 may be performed in a different order, simultaneously, repeated, or omitted. Figure 13 illustrates a schematic view of the method 1200 for enhancing resolution of a seismic image with referenced high-resolution 2D image(s) using machine-learning algorithms, according to an embodiment.

[0080] The method 1200 may include receiving a seismic image from a 3D volume, as at 1210. The seismic image(s) may be referred to as input 2D high-resolution image(s). This is shown on the top left of Figure 13.

[0081] The method 1200 may also include converting the input 2D high-resolution seismic image(s) into 2D low-resolution seismic image(s), as at 1220. This may be referred to as a data preparation stage. This is shown on the bottom left of Figure 13. The 2D low-resolution image(s) may have the same resolution as the 3D volume to be enhanced. The image processing can be a spectral shaping via frequency filtering, or image-to-image transformation with ML, or any applicable process that ensures the output 2D image(s) express the same frequency spectrum with the input 3D volume for enhancement.

[0082] The method 1200 may also include extracting low-resolution seismic traces from the 2D low-resolution image(s), as at 1230, and extracting high-resolution seismic traces from the 2D high-resolution image(s), as at 1240. This is shown by the two arrows pointing to the right in Figure 13. As described below, the extracted traces may be used to train an ML model.

[0083] The method 1200 may also include mapping the high-resolution seismic traces, as at 1250. The ML application for high-resolution mapping from the low-resolution traces includes a supervised ML algorithm that converts from low-resolution to high-resolution traces. The low- resolution traces may be extracted from the output 2D image(s). The mapping between the low- resolution traces and high-resolution traces can be established by designing and training a convolutional neural network. Illustrative networks may include the U-Net, deep convolutional network, cGAN, GAN, etc.

[0084] The method 1200 may also include generating an output high-resolution seismic image of the 3D volume based upon the mapped high-resolution seismic traces, as at 1260. More particularly, once the supervised algorithm is trained, it may be applied back to the original 3D volume to generate the higher-resolution volume whose vertical resolution is similar to the referenced 2D high-resolution image(s) input. In addition, once the ML algorithm is properly trained, it may also or instead be applied to the original low-resolution seismic images to create resolution-enhanced seismic images.

[0085] The method 1200 may also include displaying the outputs, as at 1270. This may be or include the 2D low-resolution image(s), the low-resolution traces, the high-resolution traces, the mapped high-resolution seismic traces, the higher-resolution version of the 3D volume, or a combination thereof.

[0086] The method 1200 may also include performing a wellsite action, as at 1280. The wellsite action may be based upon the 2D low-resolution image(s), the low-resolution traces, the high- resolution traces, the mapped high-resolution seismic traces, the higher-resolution version of the 3D volume, or a combination thereof. The wellsite action may be or include generating and / or transmitting a signal (e.g., using a computing system) that causes a physical action to occur at a wellsite. The wellsite action may also or instead include performing the physical action at the wellsite. The physical action may include selecting where to drill a wellbore, drilling the wellbore, varying a weight and / or torque on a drill bit that is drilling the wellbore, varying a drilling trajectory of the wellbore, varying a concentration and / or flow rate of a fluid pumped into the wellbore, or the like.Amplitude Maps

[0087] Figures 14A and 14B illustrate the amplitude maps extracted at a top of a reservoir to show the improvements of amplitude contrasts after the ML application in the first method (Figures 2-11), according to an embodiment. More particularly, Figure 14A illustrates an amplitude map with an input seismic volume, and Figure 14B illustrates the amplitude map with a high(er)-resolution seismic volume due to the ML application.Improvements of High Frequency Content

[0088] Figure 15 illustrates the improvements of high frequency content in the ML enhanced image in the first method (Figures 2-11), according to an embodiment. More particularly, Figure 15 illustrates a frequency spectrum comparison showing the improvement of the high- frequency content while still retaining the low frequency content.Seismic Sections after High-Pass Frequency Filtering

[0089] Figures 16A and 16B illustrate the comparison of seismic sections after high-pass frequency filtering to show the extra fine features generated by the ML are real and fully supported by the background geology in the first method (Figures 2-11), according to an embodiment. More particularly, Figures 16A and 16B illustrate cross-sections of images to show the validation of the extra events due to the high-frequency contents of the ML enhanced images. Figure 16A illustrates the input seismic volume, and Figure 16B illustrates a high(er)-resolution seismic volume due to the high-pass frequency filtering.

[0090] A high-pass frequency filtering may suppress parts of the signals whose frequency content is smaller than the desired frequency threshold. In one embodiment, Stage 2 may not be applied to the whole image volume. In addition, it may be implemented in a trace-by-trace (or several traces) manner. In Stage 3, the input to the network may be synthetic low-resolution traces, and the output may be high-resolution traces.

[0091] Current available approaches are either based on (a) synthetic models, which involve tedious model generation and then data computation; or (b) pseudo well log generations based on random sampling of the available dataset, which introduces biases that reduce the generality of the algorithm when applying to areas with complex geological setting. The methods described herein do not suffer the data deficiency and the model generality. These are tackled by Stage 1 and / orStage 2, where the well logs are properly interpolated and extrapolated using the seismic images, which contain geological information. The methods can be used primarily to improve the resolution of seismic images with available well logs. This is a pre-requisite in the high-resolution reservoir interpretation workflow.

[0092] In addition, the algorithm in Stage 1 can be used for pseudo well generations or property model calculations from seismic images. The methods directly tackle the well log deficiency problem that currently available approaches are facing. The pseudo well log generation is strongly calibrated by the seismic images, which contain the geological information of the subsurface in the whole seismic survey region instead of some specific locations where the well logs are available. Thus, the methods provide a more reliable resolution-enhanced images that are geologically supported by the input seismic images.Exemplary Computing System

[0093] In some embodiments, the methods of the present disclosure may be executed by a computing system. Figure 17 illustrates an example of such a computing system 1700, in accordance with some embodiments. The computing system 1700 may include a computer or computer system 1701 A, which may be an individual computer system 1701 A or an arrangement of distributed computer systems. The computer system 1701A includes one or more analysis modules 1702 that are configured to perform various tasks according to some embodiments, such as one or more methods disclosed herein. To perform these various tasks, the analysis module 1702 executes independently, or in coordination with, one or more processors 1704, which is (or are) connected to one or more storage media 1706. The processor(s) 1704 is (or are) also connected to a network interface 1707 to allow the computer system 1701A to communicate over a data network 1709 with one or more additional computer systems and / or computing systems, such as 1701B, 1701C, and / or 1701D (note that computer systems 1701B, 1701C and / or 1701D may or may not share the same architecture as computer system 1701 A, and may be located in different physical locations, e.g., computer systems 1701A and 1701B may be located in a processing facility, while in communication with one or more computer systems such as 1701C and / or 1701D that are located in one or more data centers, and / or located in varying countries on different continents).

[0094] A processor may include a microprocessor, microcontroller, processor module or subsystem, programmable integrated circuit, programmable gate array, or another control or computing device.

[0095] The storage media 1706 may be implemented as one or more computer-readable or machine-readable storage media. Note that while in the example embodiment of Figure 17 storage media 1706 is depicted as within computer system 1701 A, in some embodiments, storage media 1706 may be distributed within and / or across multiple internal and / or external enclosures of computing system 1701A and / or additional computing systems. Storage media 1706 may include one or more different forms of memory including semiconductor memory devices such as dynamic or static random access memories (DRAMs or SRAMs), erasable and programmable read-only memories (EPROMs), electrically erasable and programmable read-only memories (EEPROMs) and flash memories, magnetic disks such as fixed, floppy and removable disks, other magnetic media including tape, optical media such as compact disks (CDs) or digital video disks (DVDs), BLURAY® disks, or other types of optical storage, or other types of storage devices. Note that the instructions discussed above may be provided on one computer-readable or machine-readable storage medium, or may be provided on multiple computer-readable or machine-readable storage media distributed in a large system having possibly plural nodes. Such computer-readable or machine-readable storage medium or media is (are) considered to be part of an article (or article of manufacture). An article or article of manufacture may refer to any manufactured single component or multiple components. The storage medium or media may be located either in the machine running the machine-readable instructions, or located at a remote site from which machine-readable instructions may be downloaded over a network for execution.

[0096] In some embodiments, computing system 1700 contains one or more super resolution module(s) 1708. In the example of computing system 1700, computer system 1701 A includes the super resolution module 1708. In some embodiments, a single super resolution module may be used to perform some aspects of one or more embodiments of the methods disclosed herein. In other embodiments, a plurality of super resolution modules may be used to perform some aspects of methods herein.

[0097] It should be appreciated that computing system 1700 is merely one example of a computing system, and that computing system 1700 may have more or fewer components than shown, may combine additional components not depicted in the example embodiment ofFigure 17, and / or computing system 1700 may have a different configuration or arrangement of the components depicted in Figure 17. The various components shown in Figure 17 may be implemented in hardware, software, or a combination of both hardware and software, including one or more signal processing and / or application specific integrated circuits.

[0098] Further, the steps in the processing methods described herein may be implemented by running one or more functional modules in information processing apparatus such as general purpose processors or application specific chips, such as ASICs, FPGAs, PLDs, or other appropriate devices. These modules, combinations of these modules, and / or their combination with general hardware are included within the scope of the present disclosure.

[0099] Computational interpretations, models, and / or other interpretation aids may be refined in an iterative fashion; this concept is applicable to the methods discussed herein. This may include use of feedback loops executed on an algorithmic basis, such as at a computing device (e.g., computing system 1700, Figure 17), and / or through manual control by a user who may make determinations regarding whether a given step, action, template, model, or set of curves has become sufficiently accurate for the evaluation of the subsurface three-dimensional geologic formation under consideration.

[0100] The foregoing description, for purpose of explanation, has been described with reference to specific embodiments. However, the illustrative discussions above are not intended to be exhaustive or limiting to the precise forms disclosed. Many modifications and variations are possible in view of the above teachings. Moreover, the order in which the elements of the methods described herein are illustrated and described may be re-arranged, and / or two or more elements may occur simultaneously. The embodiments were chosen and described in order to best explain the principles of the disclosure and its practical applications, to thereby enable others skilled in the art to best utilize the disclosed embodiments and various embodiments with various modifications as are suited to the particular use contemplated.

Claims

CLAIMSWhat is claimed is:

1. A method for enhancing a resolution of a seismic image with reference well log measurements, the method comprising: receiving a first low-resolution seismic image; generating one or more pseudo well logs based upon the first low-resolution seismic image; generating synthetic seismic traces, wherein the synthetic seismic traces comprise: synthetic low-resolution seismic traces that are based upon the first low-resolution seismic image and the one or more pseudo well logs; and synthetic high-resolution seismic traces that are based upon a targeted high- resolution seismic image and the one or more pseudo well logs; training a convolutional neural network to map the synthetic low-resolution seismic traces to the synthetic high -re solution seismic traces; and generating the targeted high-resolution seismic image using the trained convolutional neural network.

2. The method of Claim 1, further comprising receiving one or more well logs corresponding to one or more wells, wherein the one or more pseudo well logs are also based upon the one or more well logs.

3. The method of Claim 2, further comprising: pre-processing the one or more well logs to produce pre-processed well logs; and pre-processing the first low-resolution seismic image to produce a pre-processed low- resolution seismic image, wherein the one or more pseudo well logs are based upon the pre-processed well logs and the pre-processed low-resolution seismic image.

4. The method of Claim 2, wherein the one or more pseudo well logs are generated by converting at least a portion of the first low-resolution seismic image into the one or more pseudo well logs using the convolutional neural network guided by the one or more well logs.

5. The method of Claim 1, wherein the one or more pseudo well logs comprise reflection coefficients that represent one or more acoustic properties of subsurface layers, and wherein the one or more acoustic properties comprise an acoustic wave impedance.

6. The method of Claim 1, wherein the synthetic low-resolution seismic traces and the synthetic high-resolution seismic traces comprise reflection seismic signals.

7. The method of Claim 1, wherein the targeted high-resolution seismic image is generated by inputting traces from the first low-resolution seismic image into the trained convolutional neural network.

8. The method of Claim 1, wherein the targeted high-resolution seismic image is generated by inputting traces from a second low-resolution seismic image into the trained convolutional neural network, and wherein the second low-resolution seismic image is from a same seismic survey as the first low-resolution seismic image, or the second low-resolution seismic image is similar to the first low-resolution seismic image in terms of seismic attributes including bandwidth, source wavelet signature, noise characteristics, and / or geological features.

9. The method of Claim 1, further comprising displaying the targeted high-resolution seismic image.

10. The method of Claim 1, further comprising performing a wellsite action based upon or in response to the targeted high-resolution seismic image.

11. A computing system, comprising: one or more processors; and a memory system comprising one or more non-transitory computer-readable media storing instructions that, when executed by at least one of the one or more processors, cause the computing system to perform operations, the operations comprising: receiving one or more well logs corresponding to one or more wells;receiving a first low-resolution seismic image, which covers a trajectory of the one or more wells; pre-processing the one or more well logs to produce pre-processed well logs; pre-processing the first low-resolution seismic image to produce a pre-processed low-resolution seismic image; generating pseudo well logs based upon the pre-processed well logs and the pre- processed low-resolution seismic image, wherein the pseudo well logs are generated by converting at least a portion of the pre-processed low-resolution seismic image into the pseudo well logs using a convolutional neural network guided by the pre-processed well logs; generating synthetic seismic traces, wherein the synthetic seismic traces comprise synthetic low-resolution seismic traces that are based upon the pre-processed low- resolution seismic image and the pseudo well logs, wherein the synthetic seismic traces also comprise synthetic high-resolution seismic traces that are based upon a targeted high- resolution seismic image and the pseudo well logs; training the convolutional neural network to map the synthetic low-resolution seismic traces to the synthetic high-resolution seismic traces; and generating the targeted high-resolution seismic image using the trained convolutional neural network, wherein the targeted high-resolution seismic image is generated by inputting traces from the pre-processed low-resolution seismic image or a second low-resolution seismic image into the trained convolutional neural network.

12. The computing system of Claim 11, wherein the synthetic low-resolution seismic traces and the synthetic high-resolution seismic traces are generated by: applying a first low-pass filter to the pseudo well logs, wherein the first low-pass filter comprises a first cutoff frequency that is determined based upon the pre-processed low-resolution seismic image; and applying a second low-pass filter to the pseudo well logs, wherein the second low-pass filter comprises a second cutoff frequency that is determined based upon the targeted high- resolution seismic image.

13. The computing system of Claim 12, wherein the second cutoff frequency is less than a Nyquist frequency and greater than a highest frequency of the first low-resolution seismic image.

14. The computing system of Claim 11, wherein the synthetic low-resolution seismic traces and the synthetic high-resolution seismic traces are generated by: convolving a pre-designed low-frequency seismic wavelet with the pseudo well logs to produce the synthetic low-resolution seismic traces; and convolving a pre-designed high-frequency seismic wavelet with the pseudo well logs to produce the synthetic high-resolution seismic traces.

15. The computing system of Claim 14, wherein the pre-designed low-frequency seismic wavelet is extracted from the pre-processed low-resolution seismic image, and wherein the predesigned high-frequency seismic wavelet is determined based upon the targeted high-resolution seismic image.

16. A non-transitory computer-readable medium storing instructions that, when executed by one or more processors of a computing system, cause the computing system to perform operations, the operations comprising: receiving well logs corresponding to one or more wells; receiving a first low-resolution seismic image, which covers a trajectory of the one or more wells; pre-processing the well logs to produce pre-processed well logs; pre-processing the first low-resolution seismic image to produce a pre-processed low- resolution seismic image; generating pseudo well logs based upon the pre-processed well logs and the pre-processed low-resolution seismic image, wherein the pseudo well logs are generated by converting at least a portion of the pre-processed low-resolution seismic image into the pseudo well logs using a convolutional neural network guided by the pre-processed well logs, wherein the pseudo well logs comprise reflection coefficients that represent one or more acoustic properties of subsurface layers, and wherein the one or more acoustic properties comprise an acoustic wave impedance;generating synthetic seismic traces, wherein the synthetic seismic traces comprise synthetic low-resolution seismic traces that are based upon the pre-processed low-resolution seismic image and the pseudo well logs, wherein the synthetic seismic traces also comprise synthetic high- resolution seismic traces that are based upon a targeted high-resolution seismic image and the pseudo well logs, wherein the synthetic low-resolution seismic traces and the synthetic high- resolution seismic traces comprise reflection seismic signals, and wherein the synthetic low- resolution seismic traces and the synthetic high-resolution seismic traces are generated by: applying first and second low-pass filters to the pseudo well logs, wherein the first low-pass filter comprises a first cutoff frequency that is determined based upon the pre- processed low-resolution seismic image, wherein the second low-pass filter comprises a second cutoff frequency that is determined based upon the targeted high-resolution seismic image, wherein the second cutoff frequency is less than a Nyquist frequency and greater than a highest frequency of the first low-resolution seismic image; or convolving a pre-designed low-frequency seismic wavelet with the pseudo well logs to produce the synthetic low-resolution seismic traces, and convolving a pre-designed high-frequency seismic wavelet with the pseudo well logs to produce the synthetic high- resolution seismic traces, wherein the pre-designed low-frequency seismic wavelet is extracted from the pre-processed low-resolution seismic image, and wherein the predesigned high-frequency seismic wavelet is determined based upon the targeted high- resolution seismic image; training the convolutional neural network to map the synthetic low-resolution seismic traces to the synthetic high-resolution seismic traces; and generating the targeted high-resolution seismic image using the trained convolutional neural network, wherein the targeted high-resolution seismic image is generated by inputting traces from the pre-processed low-resolution seismic image or a second low-resolution seismic image into the trained convolutional neural network, wherein the second low-resolution seismic image is from a same seismic survey as the first low-resolution seismic image.

17. The non-transitory computer-readable medium of Claim 16, wherein the operations further comprise displaying the targeted high-resolution seismic image.

18. The non-transitory computer-readable medium of Claim 16, wherein the operations further comprise performing a wellsite action based upon or in response to the targeted high-resolution seismic image.

19. The non-transitory computer-readable medium of Claim 18, wherein the wellsite action comprises generating and / or transmitting a signal that instructs or causes a physical action to occur at a wellsite.

20. The non-transitory computer-readable medium of Claim 19, wherein the physical action comprises selecting where to drill a wellbore, drilling the wellbore, varying a weight and / or torque on a drill bit that is drilling the wellbore, varying a drilling trajectory of the wellbore, or varying a concentration and / or flow rate of a fluid pumped into the wellbore.

Images