DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statements (IDS) submitted on 5/1/2024 was in compliance with the provisions of 37 CFR 1.97. Accordingly, the IDS is being considered by the examiner.
Claim Objections
Claims 9, 15, 18, 21, 24 and 27 are objected to because of the following informalities:
In claim 9 lines 1-2, claim 15 lines 1-2, claim 18 lines 1-2, claim 24 lines 1-2, and claim 27 lines 1-2, “comprising generating a seismic image” should read “the operations comprising generating a seismic image.”
In claim 18 line 1 and in claim 21 line 1, “The method of claim 1,” should be removed.
Appropriate correction is required.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-27 are rejected under 35 U.S.C. 103 as being unpatentable over Colombo (US 2023/0125277 A1) in view of Araya-Polo et al., “Deep-learning tomography,” The Leading Edge, https://doi.org/10.1190/tle37010058.1, 1 January 2018.
As to claims 1, 4, and 7 Colombo teaches “[a] computer-implemented method (FIG. 10 computing system 100 configured to process seismic data; [0101] computing system 1000 implementing method)”, “a non-transitory computer-readable storage medium having executable code stored thereon (Abstract; FIG. 10 computer 1002 including memory 1010)”, “a system (FIG. 10 computer 1002)” “comprising” “a seismic data processor (FIG. 10 computer 1002); a non-transitory computer-readable storage memory accessible by the seismic data processor and having executable code stored thereon (FIG. 10 computer 1002 including memory 1010 accessible by processor 1008)” “for determining uphole velocities of an uphole velocity model for an uphole seismic survey (Abstract describing method including using uphole velocity data (uphole model effectively entailing uphole seismic survey); [0011]-[0012] interpolation used to extrapolate update velocity data) comprising an uphole seismic survey dataset ([0044] seismic data transmitted by sensor(s) 116 to computer 118) generated from a seismic receiver station configured to sense seismic signals originating from a seismic source station (FIG. 1A depicting seismic sensor(s) 116 configured to generate seismic data to be processed by computer 118 based on seismic signals originating from seismic source(s) 112; [0042], [0044]),” the method/operations comprising:
“obtaining the uphole seismic survey dataset comprising first break travel times ([0048] and [0062] received seismic trace data include first break/arrival picks (known in the art at travel times));
sorting the first break travel times into offset bins of a travel time attribute cube according to common midpoints (Abstract seismic trace data that per [0048] and [0062] including first break/arrivals distributed into common midpoint offset bins; FIG. 2 depicting XYO space (cube) 140, [0049] seismic travel time traces sorted in the CMP-offset bins 150; FIGS. 508 step 502, [0058]-[0059]) for refracted seismic wave travel between the seismic sources and the seismic receiver ([0049] binning configured to accommodate refracted waves such as depicted by FIG. 1B and surface waves 115 in FIG. 1A);
removing anomalous travel times from the sorted travel times in the offset bins to form a refined first break dataset (FIGS. 5-8 step 504, [0062] following formation of offset bins, reject outliers in first break data);
forming a travel times vs offset function based on the refined first break dataset ([0062] robust (outlier filtered) first break picks used to evaluate mean travel times per XYO bin forming travel time-offset functions);
obtaining uphole” [velocities] “associated with the uphole seismic survey dataset (FIG. 6 step 602, [0076] update vertical velocity data received in association with CMP-based velocity for combined processing by geostatic module), the uphole” [velocities] “comprising” [velocities] “vs depth (FIG. 8 steps 802 and 602 uphole velocities converted from travel time vs depth; [0056] travel time-depth converted to velocity-depth; [0033]-[0034] useful velocity profile depends on depth);
training a supervised machine learning model ([0053] machine learning used for interpolation/extrapolation; [0055] geostatics module that generates shallow velocity model applies machine learning; [0079] and [0082] machine learning model used for interpolation/extrapolation may be trained supervised model) using training data comprising the” [velocities] “vs offset function at a common midpoint (CMP) ([0083]-[0084] training data includes CMP velocity (CMP data characterized by offset bins)) based on an uphole location (CMP binning per [0059] associated with XY location), and the uphole” [velocities] ([0084] training data includes uphole velocity) “at the uphole location ([0076] uphole velocity data is vertical velocity data (inherently associated with an uphole location) and processed for interpolation/extrapolation of uphole velocities in association with CMP velocities (interpolation/extrapolation clearly implemented for a common location/area with CMP data such as e.g., [0035] describing interpolation incorporating spatial correlation); [0039] “localized” uphole velocities used for interpolation/extrapolation),”
“determining uphole” [velocities] “for the entire uphole seismic survey dataset using the trained machine learning model ([0076] geostatics module uses ML/DL to interpolate/extrapolate for determining the uphole velocities based on regionalized parameter distribution provided by CMP velocity model); and
transforming” “determined uphole times to uphole velocities (FIGS. 5-8 steps 518 and 520, [0073]-[0074] first break travel times translated to corresponding velocity model).”
Colombo discloses a method for applying machine learning modeling to velocity data rather than travel times from which the velocity data is generated and therefore does not expressly teach “obtaining uphole times associated with the uphole seismic survey dataset, the uphole times comprising travel times vs depth,” “training a supervised machine learning model using training data comprising the travel times vs offset function at a common midpoint (CMP) based on an uphole location, and the uphole times…” “determining uphole times for the entire uphole seismic survey dataset using the trained machine learning model,” and “transforming the determined uphole times to uphole velocities.”
While Colombo does not explicitly describe the manner of training the model in terms of labelling, Colombo discloses supervised machine learning modeling to interpolate/extrapolate uphole velocities in which the model is trained using uphole velocities and CMP velocities (offset binned velocity data). In a supervised machine learning context, training fundamentally entails labelling features with labels that represent the machine learning outputs as described by Araya-Polo (page 60, DNNs, paragraph beginning with “DNNs are composed of …” explaining use of training labels as corresponding to model outputs).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Araya-Polo’s teaching of training a machine learning model using labels that correspond to model outputs to the method taught by Colombo in which supervised machine learning model is trained to output uphole velocity data such that in combination the method includes the uphole velocities being the labels for the training data.
Such a combination would amount to selecting a known design option for training a model to achieve predictable results.
Colombo discloses, as was well known in the art, that seismic travel times are utilized as a fundamental parameter for determining subsurface (e.g., uphole) seismic velocities in terms of the travel times varying with respect to various downhole material compositions and that travel times are ultimately translated to corresponding velocities for purposes of surface material characterization (FIGS. 5-8 depicting translation of seismic travel time data to generate corresponding seismic velocity data; [0033]). In this manner, Colombo effectively discloses that travel times largely correspond seismic velocities in terms of characterizing uphole material compositions, which is consistent with the processing sequence in claim 1 in which uphole times are transformed to uphole velocities.
Furthermore, Colombo teaches that supervised machine learning may be used for estimating travel times ([0073] supervised learning used to train CNN to predict uphole travel times).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Colombo’s teaching of using a machine learning/training process for predicting uphole travel times to the method taught by Colombo as modified by Araya-Polo in which machine learning is used for interpolation/extrapolation of uphole velocity data using training data in the form of distributed velocities that is derived from a distribution of travel times in which uphole velocities are used as labels to derive a pseudo velocity model, such that in combination Colombo’s method includes using supervised machine learning to interpolate/extrapolate uphole times (determine uphole time for the entire seismic dataset) using the binned uphole travel time data (e.g., [0033] and [0047]-[0049]) as training data labeled with uphole times corresponding to Colombo’s disclosed uphole velocities (FIG. 8 uphole velocities 602 determined based on uphole time data 802), and in such configuration to transform the determined uphole times to uphole velocities as part of the process to generate the pseudo-3D velocity model.
Such a combination would have been obvious to try for one of ordinary skill due to the direct correspondence between the seismic travel time data and velocity data and in accounting for interpolation/extrapolation gaps that Colombo’s velocity-based machine learning interpolation/extrapolation is aimed at addressing such that using binned travel times and corresponding uphole time labels would be a readily apparent substitution.
As to claims 2, 5, and 8, the combination of Colombo and Araya-Polo teaches “wherein the supervised machine learning model comprises a fully-connected artificial neural network (ANN) (Colombo: [0082] model may be a CNN (a type of ANN having a fully connected final layer)), a convolutional neural network (CNN) (Colombo: [0082] model may be CNN) or a multivariate regression model.”
As to claims 3, 6, and 9 the combination of Colombo and Araya-Polo teaches “generating a seismic image using the uphole velocities (Colombo: FIG. 5 steps 520 and 522 full wave inversion (imaging) implemented based on pseudo velocity model; FIG. 6 steps 520, 606, 610, and 522 full wave inversion modeling (imaging) implemented based on pseudo velocity model; [0007]-[0008] and [0033]).”
As to claims 10, 13, and 16 Colombo teaches “[a] computer-implemented method (FIG. 10 computing system 100 configured to process seismic data; [0101] computing system 1000 implementing method)” “a non-transitory computer-readable storage medium having executable code stored thereon (Abstract; FIG. 10 computer 1002 including memory 1010)”, “a system (FIG. 10 computer 1002)” “comprising” “a seismic data processor (FIG. 10 computer 1002); a non-transitory computer-readable storage memory accessible by the seismic data processor and having executable code stored thereon (FIG. 10 computer 1002 including memory 1010 accessible by processor 1008)” “for determining uphole velocities of an uphole velocity model for an uphole seismic survey (Abstract describing method including using uphole velocity data (uphole model effectively entailing uphole seismic survey); [0011]-[0012] interpolation used to extrapolate update velocity data) comprising an uphole seismic survey dataset ([0044] seismic data transmitted by sensor(s) 116 to computer 118) generated from a seismic receiver station configured to sense seismic signals originating from a seismic source station (FIG. 1A depicting seismic sensor(s) 116 configured to generate seismic data to be processed by computer 118 based on seismic signals originating from seismic source(s) 112; [0042], [0044]),” the method/operations comprising:
“obtaining the uphole seismic survey dataset comprising first break travel times ([0048] and [0062] received seismic trace data include first break/arrival picks (known in the art at travel times));
sorting the first break travel times into offset bins of a travel time attribute cube according to common midpoints (Abstract seismic trace data that per [0048] and [0062] including first break/arrivals distributed into common midpoint offset bins; FIG. 2 depicting XYO space (cube) 140, [0049] seismic travel time traces sorted in the CMP-offset bins 150; FIGS. 508 step 502, [0058]-[0059]) for refracted seismic wave travel between the seismic sources and the seismic receiver ([0049] binning configured to accommodate refracted waves such as depicted by FIG. 1B and surface waves 115 in FIG. 1A);
removing anomalous travel times from the sorted travel times in the offset bins to form a refined first break dataset (FIGS. 5-8 step 504, [0062] following formation of offset bins, reject outliers in first break data);
forming a travel times vs offset function based on the refined first break dataset ([0062] robust (outlier filtered) first break picks used to evaluate mean travel times per XYO bin forming travel time-offset functions);
obtaining uphole velocities associated with the uphole seismic survey dataset (FIG. 6 step 602, [0076] update vertical velocity data received in association with CMP-based velocity for combined processing by geostatic module), the uphole velocities comprising interval velocity vs. depth (FIG. 8 steps 802 and 602 uphole velocities converted from travel time vs depth; [0056] travel time-depth converted to velocity-depth; [0033]-[0034] useful velocity profile depends on depth; [0100] interval velocities processed by interpolation);
training a supervised machine learning model ([0053] machine learning used for interpolation/extrapolation; [0055] geostatics module that generates shallow velocity model applies machine learning; [0079] and [0082] machine learning model used for interpolation/extrapolation may be trained supervised model) using training data comprising the travel times vs offset function at a common midpoint (CMP) ([0083]-[0084] training data includes CMP velocity (CMP data characterized by offset bins)) based on an uphole location (CMP binning per [0059] associated with XY location) and uphole velocities ([0084] training data includes uphole velocity) at the uphole location ([0076] uphole velocity data is vertical velocity data (inherently associated with an uphole location) and processed for interpolation/extrapolation of uphole velocities in association with CMP velocities (interpolation/extrapolation clearly implemented for a common location/area with CMP data such as e.g., [0035] describing interpolation incorporating spatial correlation); [0039] “localized” uphole velocities used for interpolation/extrapolation),” “and
determining uphole velocities for the uphole seismic survey dataset using the trained machine learning model ([0076] geostatics module uses ML/DL to interpolate/extrapolate for determining the uphole velocities based on regionalized parameter distribution provided by CMP velocity model).”
While Colombo does not explicitly describe the manner of training the model in terms of labelling, Colombo discloses supervised machine learning modeling to interpolate/extrapolate uphole velocities in which the model is trained using uphole velocities and CMP velocities (offset binned velocity data). In a supervised machine learning context, training fundamentally entails labelling features with labels that represent the machine learning outputs as described by Araya-Polo (page 60, DNNs, paragraph beginning with “DNNs are composed of …” explaining use of training labels as corresponding to model outputs).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Araya-Polo’s teaching of training a machine learning model using labels that correspond to model outputs to the method taught by Colombo in which supervised machine learning model is trained to output uphole velocity data such that in combination the method includes the uphole velocities being the labels for the training data.
Such a combination would amount to selecting a known design option for training a model to achieve predictable results.
As to claims 11, 14, and 17, the combination of Colombo and Araya-Polo teaches “wherein the supervised machine learning model comprises a fully-connected artificial neural network (ANN) (Colombo: [0082] model may be a CNN (a type of ANN having a fully connected final layer)), a convolutional neural network (CNN) (Colombo: [0082] model may be CNN) or a multivariate regression model.”
As to claims 12, 15, and 18, the combination of Colombo and Araya-Polo teaches “generating a seismic image using the uphole velocities (Colombo: FIG. 5 steps 520 and 522 full wave inversion (imaging) implemented based on pseudo velocity model; FIG. 6 steps 520, 606, 610, and 522 full wave inversion modeling (imaging) implemented based on pseudo velocity model; [0007]-[0008] and [0033]).”
As to claims 19, 22, and 25, Colombo teaches “[a] computer-implemented method (FIG. 10 computing system 100 configured to process seismic data; [0101] computing system 1000 implementing method)” “a non-transitory computer-readable storage medium having executable code stored thereon (Abstract; FIG. 10 computer 1002 including memory 1010)”, “a system (FIG. 10 computer 1002)” “comprising” “a seismic data processor (FIG. 10 computer 1002); a non-transitory computer-readable storage memory accessible by the seismic data processor and having executable code stored thereon (FIG. 10 computer 1002 including memory 1010 accessible by processor 1008)” “for determining uphole velocities of an uphole velocity model for an uphole seismic survey (Abstract describing method including using uphole velocity data (uphole model effectively entailing uphole seismic survey); [0011]-[0012] interpolation used to extrapolate update velocity data) comprising an uphole seismic survey dataset ([0044] seismic data transmitted by sensor(s) 116 to computer 118) generated from a seismic receiver station configured to sense seismic signals originating from a seismic source station (FIG. 1A depicting seismic sensor(s) 116 configured to generate seismic data to be processed by computer 118 based on seismic signals originating from seismic source(s) 112; [0042], [0044]),” the method/operations comprising:
“obtaining the uphole seismic survey dataset comprising first break travel times ([0048] and [0062] received seismic trace data include first break/arrival picks (known in the art at travel times));
sorting the first break travel times into offset bins of a travel time attribute cube according to common midpoints (Abstract seismic trace data that per [0048] and [0062] including first break/arrivals distributed into common midpoint offset bins; FIG. 2 depicting XYO space (cube) 140, [0049] seismic travel time traces sorted in the CMP-offset bins 150; FIGS. 508 step 502, [0058]-[0059]) for refracted seismic wave travel between the seismic sources and the seismic receiver ([0049] binning configured to accommodate refracted waves such as depicted by FIG. 1B and surface waves 115 in FIG. 1A);
removing anomalous travel times from the sorted travel times in the offset bins to form a refined first break dataset (FIGS. 5-8 step 504, [0062] following formation of offset bins, reject outliers in first break data);
forming a travel times vs offset function based on the refined first break dataset ([0062] robust (outlier filtered) first break picks used to evaluate mean travel times per XYO bin forming travel time-offset functions);
inverting the travel-times vs offset function to obtain a velocity model for first break waves, wherein the velocity model comprises seismic velocities vs. depth (FIGS. 5-8 steps 518 and 520, [0073]-[0074] CMP-based travel time offset for first break travel times are converted/translated into a corresponding velocity depth function/model via 1-D inversion);
obtaining uphole velocities associated with the uphole seismic survey dataset (FIG. 6 step 602, [0076] update vertical velocity data received in association with CMP-based velocity for combined processing by geostatic module), the uphole velocities comprising interval velocity vs. depth (FIG. 8 steps 802 and 602 uphole velocities converted from travel time vs depth; [0056] travel time-depth converted to velocity-depth; [0033]-[0034] useful velocity profile depends on depth; [0100] interval velocities processed by interpolation);
training a supervised machine learning model ([0053] machine learning used for interpolation/extrapolation; [0055] geostatics module that generates shallow velocity model applies machine learning; [0079] and [0082] machine learning model used for interpolation/extrapolation may be trained supervised model) using training data comprising the seismic velocities vs. depth ([0083]-[0084] training data includes CMP velocity (CMP data characterized by offset bins)) at an uphole location (CMP binning per [0059] associated with XY location) and the uphole velocities at the uphole location ([0076] uphole velocity data is vertical velocity data (inherently associated with an uphole location) and processed for interpolation/extrapolation of uphole velocities in association with CMP velocities (interpolation/extrapolation clearly implemented for a common location/area with CMP data such as e.g., [0035] describing interpolation incorporating spatial correlation); [0039] “localized” uphole velocities used for interpolation/extrapolation),” “and
determining uphole velocities for the uphole seismic survey dataset using the trained machine learning model ([0076] geostatics module uses ML/DL to interpolate/extrapolate for determining the uphole velocities based on regionalized parameter distribution provided by CMP velocity model).
While Colombo does not explicitly describe the manner of training the model in terms of labelling, Colombo discloses supervised machine learning modeling to interpolate/extrapolate uphole velocities in which the model is trained using uphole velocities and CMP velocities (offset binned velocity data). In a supervised machine learning context, training fundamentally entails labelling features with labels that represent the machine learning outputs as described by Araya-Polo (page 60, DNNs, paragraph beginning with “DNNs are composed of …” explaining use of training labels as corresponding to model outputs).
It would have been obvious to one of ordinary skill in the art before the effective filing date, to have applied Araya-Polo’s teaching of training a machine learning model using labels that correspond to model outputs to the method taught by Colombo in which supervised machine learning model is trained to output uphole velocity data such that in combination the method includes the uphole velocities being the labels for the training data.
Such a combination would amount to selecting a known design option for training a model to achieve predictable results.
As to claims 20, 23, and 26, the combination of Colombo and Araya-Polo teaches “wherein the supervised machine learning model comprises a fully-connected artificial neural network (ANN) (Colombo: [0082] model may be a CNN (a type of ANN having a fully connected final layer)), a convolutional neural network (CNN) (Colombo: [0082] model may be CNN) or a multivariate regression model.”
As to claims 21, 24, and 27, the combination of Colombo and Araya-Polo teaches “generating a seismic image using the uphole velocities (Colombo: FIG. 5 steps 520 and 522 full wave inversion (imaging) implemented based on pseudo velocity model; FIG. 6 steps 520, 606, 610, and 522 full wave inversion modeling (imaging) implemented based on pseudo velocity model; [0007]-[0008] and [0033]).”
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MATTHEW W BACA whose telephone number is (571)272-2507. The examiner can normally be reached Monday - Friday 8:00 am - 5:30 pm.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Andrew Schechter can be reached at (571) 272-2302. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MATTHEW W. BACA/Examiner, Art Unit 2857
/ANDREW SCHECHTER/Supervisory Patent Examiner, Art Unit 2857
Prosecution Insights