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 .
Under 35 USC § 101
Although claims 1, 8 and 15 include abstract ideas, claims 1, 8 and 15 also recite
additional elements such as “deploying an azimuthal borehole measurement tool into the borehole; obtaining at least one azimuthal borehole image utilizing the azimuthal borehole measurement tool and generating a synthetic image by sparse convolution of a weight function and a plurality of feature kernels”. The inclusion of these additional elements integrates the identified judicial exception into a practical application that effects a meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim(s) as a whole is more than a drafting effort designed to monopolize the exception. Therefore, claims 1-20 are considered eligible under 35 U.S.C. § 101.
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, 2, 7-9, and 14-16 are rejected under 35 U.S.C. 103 as being
unpatentable over Gelman (Pub. No. US 2018/0176499) in view of Pereg et al. (NPL: “Convolutional Sparse Coding Fast Approximation with Application to Seismic Reflectivity Estimation” (2021) (hereinafter Pereg).
As per claims 1, 2, 8, 9, 15, and 16, Gelman teaches deploying an azimuthal borehole measurement tool into the borehole; obtaining at least one azimuthal borehole image utilizing the azimuthal borehole measurement tool (see ¶¶ [0019]-[0020] and [0023], downhole measurement system including sensors 208 for acquiring measurements of a subterranean formations, The acquired measurements are processed into borehole image data, including image block 321, representing formation properties as functions of borehole positions, ¶ [0029] discloses identifying geological features appearing in the borehole image, including sinusoidal dip features corresponding to geological structures, ¶ [0027] discloses the claimed acquisition of an azimuthal borehole image using a downhole tool) ; generating a synthetic image by sparse convolution of a weight function and a plurality of feature kernels (see Fig. 3 and ¶¶ [0014], [0015], and [0030]-[0031], reconstructing a borehole image using weighted dictionary functions. Specifically, wavelet transform functions 327 constitute dictionary elements, and coefficients 328 specify their respective contributions to a reconstructed image).
However, Gelman fails to explicitly teach generating a synthetic image by sparse convolution operation (emphasis underlined).
Pereg teaches convolutional sparse coding in which measured data are reconstructed using a convolutional dictionary formed from shifted local filters and corresponding sparse coefficients (see pages 4-5, Section II-C, equations (6)-(7). Pereg also teaches applying convolutional sparse coding to geophysical measurements, particularly seismic reflectivity estimation (see Section III).
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to modify the weighted dictionary-based borehole image reconstruction method Gelman to employ the convolutional sparse coding technique taught by Pereg because both Gelman and Pereg teach reconstructing measured information using dictionary functions and associated coefficients, and Pereg teaches using sparse convolutional representations for geophysical data, thereby providing an alternative reconstruction technique in which geophysical image features are represented by localized kernels and sparse weight functions.
determining an optimal weight function that minimizes a difference between the synthetic image and the at least one azimuthal borehole image (see Gelman Fig. 3 and ¶¶ [0014], [0015], and [0030]-[0031], reconstructing a borehole image using weighted dictionary functions. Specifically, wavelet transform functions 327 constitute dictionary elements, and coefficients 328 specify their respective contributions to a reconstructed image).
Pereg teaches determining sparse coefficients by minimizing an objective function containing a reconstruction-error term and a sparsity-promoting regularization term (see Section II-B).
min
X
1
2
y
-
D
X
2
2
+
λ
X
1
,
Where y represents measured data, DX represents reconstructed data, and X represents the sparse coefficients.
Pereg further teaches using convolutional dictionary constructed from shifted local filters (see Section II-C, equations (6)-(7). In such a representation, the sparse coefficients specify the locations and contributions of the dictionary elements.
The examiner notes that in this combination, the sparse coefficient maps would constitute the weight functions associated with the geological feature kernels. Optimizing these coefficients against the measured azimuthal borehole image would produce a synthetic image whose reconstruction error is reduced relative to the measured image.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to employ Pereg’s sparse coefficient optimization technique in the borehole image reconstruction method of Gelman because Gelman teaches representing borehole images using weighted dictionary functions and Pereg teaches optimizing sparse coefficients to reduce differences between reconstructed and measured geophysical information, thereby determining weights that produce a synthetic borehole image more closely approximating the measured azimuthal borehole image.
Regarding “wherein the optimal weight function is the corresponding weight function of the plurality of feature kernels which produces the synthetic image substantially matching the at least one azimuthal borehole image”, Gelman teaches reconstructing borehole images using dictionary functions and corresponding coefficients (see ¶ [0054], equation (6)).
Pereg teaches convolutional sparse coding in which a dictionary comprises convolution matrices associated with respective filters, and the sparse coefficient are associated with the corresponding dictionary elements (see Section II-C, equations (6)-(7)). Pereg further teaches determining optimized sparse coefficients by minimizing a regularized objective comprising a reconstruction-error term between measured data and reconstructed data (see Section II-B). Thus, in the proposed combination, the convolutional dictionary filters would correspond to the claimed plurality of geological feature kernels, and the optimized sparse coefficients associated with those filters would constitute the corresponding optimal weight functions. Applying the optimized coefficients to their corresponding feature kernels would generate a reconstructed, or synthetic, azimuthal borehole image approximating the measured azimuthal borehole image.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to determine and use such optimized weight functions in the borehole image reconstruction method of Gelman because Pereg teaches optimizing coefficients associated with convolutional dictionary filters using an objective function that account for both reconstruction error and sparsity, thereby providing a known technique for determining weight functions associated with geological feature kernels and generating a synthetic borehole image that approximates the measured azimuthal borehole image while maintaining a sparse feature representation..
Regarding “overlaying the synthetic image with the plurality of feature kernels scaled by corresponding weight functions on the at least one azimuthal borehole image”, Gelman teaches reconstructing borehole images from weighted dictionary functions (see ¶ [0054], equation (6)). The reconstructed image is derived from image representation functions whose contributions are controlled by corresponding coefficients.
Pereg teaches constructing a reconstructed representation using convolutional dictionary elements and sparse coefficients and determining the coefficients by comparing the reconstructed representation with measured data (see Sections II-B to II-C).
The combination does not explicitly disclose “overlaying the synthetic image with the plurality of feature kernels scaled by corresponding weight functions on the at least one azimuthal borehole image” as recited, however, under the broadest reasonable interpretation consistent with the specification, the claimed overlaying operation encompasses computationally placing the synthetic image and measured azimuthal borehole image into corresponding spatial registration for comparison (see original specification of the present invention ¶¶ [0040]-[0043]).
The proposed combination of Gelman and Pereg would perform this operation by generating the synthetic image from weighted feature kernels and evaluating the reconstructed image against the measured borehole image at corresponding image locations during sparse-coefficient optimization.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to computationally overlay the synthetic image comprising weighted geological feature kernels with the measured azimuthal borehole image because Gelman teaches reconstructing borehole images using weighted dictionary functions and Pereg teaches optimizing convolutional sparse representations by comparing reconstructed and measured geophysical information, thereby enabling differences between corresponding image locations to be evaluated and the feature-kernel weights to be adjusted to produce a synthetic image substantially matching the measured borehole image.
Regarding “determining the one or more geological characteristics of the borehole based on the optimal weight function and corresponding feature kernel”, Gelman teaches that geological dip features appearing in azimuthal borehole images can be represented using sinusoidal functions (see ¶¶ [0035] and [0053]-[0054], equations (5)-(6), identifying parameters of cosine functions corresponding to geological features and reconstructing the image using weighted feature representations). Gelman further teaches that parameters of sinusoidal features, including amplitude and phase, are associated with geological dip magnitude and orientation (see ¶ [0029]).
Pereg teaches convolutional sparse coding in which a dictionary comprises convolution matrices associated with respective shifted filters, and sparse coefficients are associated with the corresponding dictionary elements (see Section II-C, equations (6)-(7)). Pereg further teaches determining sparse coefficients using a cost function comprising a squared reconstruction-error term,
y
-
D
X
2
2
,
and a regularization term,
λ
X
1
, which promotes sparsity in the coefficients (see Section II-A, equations (2)-(4) and Section II-B).
The examiner notes that in this combination the convolutional dictionary filters taught by Pereg would be implemented as geological feature kernels representing the sinusoidal geological features taught by Gelman. The optimization sparse coefficients associated with those filters would serve as the corresponding optimal weight functions, identifying the locations and contributions of the geological feature kernels in reconstructing the measured azimuthal borehole image. The geological characteristics would then be determined using the identified feature kernels and their corresponding optimized weight functions. For example, an optimized nonzero coefficient associated with a particular sinusoidal feature kernel would indicate the contribution of that kernel at a corresponding image location, while the geometric parameters of the kernel, including amplitude and phase, would provide information concerning the geological dip magnitude and orientation of the represented feature.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to employ the convolutional sparse coding and coefficient-optimization technique taught by Pereg in the borehole-image reconstruction method of Gelman because Gelman teaches representing geological dip features using parametrized sinusoidal functions and reconstructing borehole images using weighted dictionary functions, while Pereg teaches optimizing sparse coefficients associated with convolutional dictionary filters using a regularized reconstruction objective, thereby providing a technique for identifying the locations and contributions of geological feature kernels and using the optimized weight functions together with the corresponding kernel parameters to determine geological dip magnitude, dip orientation, and other geological characteristics.
Regarding “at least one processor” and “a memory” (claim 8) (see ¶ [0025]).
As per claims 7 and 14, the combination of Gelman and Pereg teaches the system as stated above. Gelman teaches representing geological dip features appearing in azimuthal borehole images using sinusoidal functions characterized by amplitude, phase, and depth location (see ¶¶ [0029], [0035], and [0043]-[0045]). Particularly, Gelman teaches a sinusoidal model expressed by equation (1), where the amplitude A corresponds to dip magnitude, the phase shift Ԏ corresponds to dip orientation, and the constant C defines the exact location of the dip within the image block and may be inferred from measured depth (see ¶¶ [0043]-[0045]). Gelman also teaches representing a reconstructing borehole images using dictionary atoms and corresponding numerical coefficients (see ¶¶ [0053]-[0054], equations (5)-(6)).
Pereg teaches convolutional sparse coding in which convolutional dictionary filters are associated with corresponding sparse coefficients (see Section II-C, equations (6)-(7)). Pereg further teaches optimizing sparse coefficients using an objective function comprising a reconstruction-error term and L1-norm regularization term (see Section II-B).
The examiner notes that the sinusoidal geological features taught by Gelman would be represented using a plurality of sinusoidal feature kernels having different candidate amplitudes, while the corresponding sparse weight representation taught by Pereg would be organized according to the sinusoidal phase, amplitude, and depth location of the represented geological features. The organized coefficients would identify the contributions of the respective sinusoidal feature kernels at corresponding positions within the azimuthal borehole image.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to employ sinusoidal feature kernels having different candidate amplitudes and associate their corresponding sparse weights with sinusoidal phase, amplitude, and depth in the borehole-image reconstruction method of Gelman, as modified by Pereg, because Gelman explicitly teaches representing geological dip features using sinusoidal amplitude, phase shift, and depth location (see ¶¶ [0043]-[0045]), while Pereg teaches associating convolutional dictionary filters with sparse coefficients and optimizing those coefficients, thereby providing a technique for representing geological features having different dip magnitudes, orientations, and depth locations using corresponding sinusoidal feature kernels and optimized weight functions.
Claims 4, 11, and 17 are rejected under 35 U.S.C. 103 as being unpatentable
over Gelman in view of Pereg and further in view of Sugiura et al. (Pub. No. US 2014/0291024) (hereinafter Sugiura).
As per claims 4, 11, and 17, the combination of Gelman and Pereg teaches the system as stated above. While Gelman recognizes the use of downhole measurements for formation evaluation, well placement, and geosteering (see ¶ [0020]), the combination of Gelman and Pereg fails to explicitly teach steering a downhole drilling tool based on the one or more geological characteristics (emphasis underlined).
Sugiura, however, teaches a close-loop geosteering method in which formation evaluation measurements are processed to determine borehole-image parameters that are subsequently used to obtain corrective steering tool settings for controlling the direction of drilling (see ¶¶ [0028]-[0030], [0048]-[0049], AND Fig. 2). Sugiura further teaches that the borehole -image parameters may include a relative dip angle between the borehole and a formation boundary and that azimuthally resolved measurements may be processed to determine a formation dip angle, which is subsequently processed to obtain a corrective steering tool setting (see Figs 7 and 9 and ¶¶ [0038]-[0041]). The corrective steering tool setting is applied to the steering tool to change the direction of drilling and maintain a desired trajectory relative to a geological formation (see ¶¶ [0006], [0023] and [0025]). Thus, Sugiura teaches steering a downhole drilling tool based on geological characteristics, including dip angle and the relative position of a formation boundary.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to use the geological characteristics determined by the borehole-image processing method of Gelman, as modified by Pereg, to control the steering of a downhole drilling tool as taught by Sugiura because Gelman teaches obtaining geological formation characteristics from borehole images during operations, while Sugiura teaches using formation dip and other borehole image parameters to determine corrective steering tool settings, thereby enabling the drilling direction to be adjusted in response to identified geological characteristics to maintain a desired wellbore trajectory relative to the surrounding formation.
Claims 6, 13, and 19 are rejected under 35 U.S.C. 103 as being unpatentable
over Gelman in view of Pereg and further in view of Freitas et al. (Pub. No. US 2018/0031718) (hereinafter Freitas).
As per claims 6, 13, and 19, the combination of Gelman and Pereg teaches the system as stated above except that the at least one azimuthal borehole image is preprocessed prior to determining the optimal weight function, wherein the at least one azimuthal borehole image is preprocessed using automated gain control and/or wavenumber filtering.
Freitas, however, teaches processing acoustic reflection images obtained during borehole logging to identify geological features corresponding to reflecting structural interfaces of a subsurface formation and explains that acoustic reflection image may be obtained from an original acoustic reflection image by applying a two-dimensional spatial wavenumber filter, which may be generated from polar segment of the image’s two-dimensional spatial wavenumber amplitude spectrum and that applying spatial Fourier transformation and band-pas filtering to image data before identifying geological patterns and features (see ¶¶ [0007]-[0008], [0069], and Fig. 2). Freitas further teaches that filtering process can improve identification of meaningful geological structures by suppressing unwanted image components, including noise and migration artifacts (see ¶¶ [0075] and [0115]). Thus, Freitas teaches using wavenumber filtering as an image-processing operation before subsequent geological-feature analysis.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to modify the borehole-image reconstruction method of Gelman, as modified by Pereg, to preprocess the measured borehole image using the two-dimensional spatial wavenumber filtering technique taught by Freitas because Freitas teaches that spatial wavenumber filtering can suppress unwanted components and improve the identification of meaningful geological features in borehole images, thereby providing a filtered measured image for subsequent sparse-convolution reconstruction and optimization of the feature-kernel weight functions.
Claims 3, 5, 10, 18, and 20 are rejected under 35 U.S.C. 103 as being
unpatentable over Gelman in view of Pereg and further in view of Yuan et al. (NPL: “Gradient Hard Thresholding Pursuit for Sparsity-Constrained Optimization” (2014)) (hereinafter Yuan).
As per claims 3, 10, and 20, the combination of Gelman and Pereg teaches the system as stated above except that the claimed iterative retention of selected weighted feature kernels.
Yuan, however, teaches an iterative gradient hard thresholding pursuit algorithm for sparse optimization (see Abstract and Section 2. Gradient Hard Thresholding Pursuit). In Algorithm 1, operations are repeated until a stopping condition is satisfied. Yuan also teaches applying the algorithm to squared-error minimization for sparse inverse problems (see Section 2. Gradient Hard Thresholding Pursuit). Accordingly, Yuan’s iterative support-selection procedure would retain the active weighted kernels identified by gradient-based update and optimize their coefficients to improve the reconstruction of the measured azimuthal borehole image.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to employ Yuan’s iterative gradient hard thresholding pursuit algorithm in the sparse convolutional reconstruction method of Gelman, as modified by Pereg, because Yuan teaches an iterative technique for selecting significant sparse coefficients and refining their values to minimize an objective function, thereby enabling the reconstruction process to retain a limited set of weighted geological feature kernels and improve the approximation of the measured azimuthal borehole image.
As per claims 5, 12, and 18, the combination of Gelman and Pereg teaches the system as stated above.
While the combination of Gelman and Pereg teaches sparse coefficient optimization for reconstructing measured data, the combination of Gelman and Pereg fails to explicitly teach solving the inversion equation using gradient hard thresholding pursuit.
Yuan, however, teaches Gradient Hard Thresholding Pursuit (GraHTP), an iterative algorithm for solving sparsity-constrained optimization problems (see Abstract and Section 2. Gradient Hard Thresholding Pursuit). Yuan explains that the algorithm alternates between a gradient descent operation and a hard truncation operation, with an optional debiasing step for refining the selected coefficients. Accordingly, Yuan teaches using gradient hard thresholding pursuit to determine sparse coefficients by iteratively updating candidate solutions and retaining a limited number of selected coefficients.
It would have been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to employ the Gradient Hard Thresholding Pursuit algorithm taught by Yuan to solve the sparse inverse problem arising from the borehole-image reconstruction method Gelman, as modified by Pereg, because Yuan teach Gradient Hard Thresholding Pursuit as an iterative method for solving sparsity-constraint optimization problems through gradient-based updates and hard truncation of coefficients, thereby providing a known algorithm for determining sparse weight functions associated with geological feature kernels while controlling the number of retained nonzero coefficients.
Prior art
The prior art made record and not relied upon is considered pertinent to applicant’s
disclosure:
Stockhausen et al. [‘537] discloses systems and methods for identifying formation boundaries without necessarily obtaining an azimuthal borehole image are provided. A downhole tool may be placed in a wellbore in a geological formation that has a formation boundary. First and second measurements may be obtained at a number of depths of the wellbore. The first measurement may have a first effective penetration length into the geological formation and the second measurement may have a second effective penetration length into the geological formation different from the first effective penetration length. Thus, the first measurement may detect the formation boundary at a first depth and the second measurement may detect the formation boundary at a second depth. Using a difference between the first depth and the second depth, an apparent relative angle between the wellbore and the formation boundary or an apparent formation dip, or both, may be obtained.
Aeron et al. [‘248] discloses methods and apparatus for waveform processing are disclosed. An example method includes determining shrinkage estimators in a Discrete Radon transform domain based on semblance of waveform data and de-noising the waveform data using a processor and the shrinkage estimators to enable the identification of weak signals in the waveform data.
Haugland [‘627] discloses a method for forming a borehole image of an azimuthally sensitive formation (or borehole) parameter. Such a borehole image may be formed, for example, by convolving logging sensor data acquired as a function of logging sensor azimuth with a window function. In one embodiment, the logging sensor data and the corresponding azimuths are measured during drilling of a borehole while a LWD tool rotates with the drill string in the borehole. The convolution is typically evaluated at one or more azimuthal positions about a circular horizon. The convolved logging sensor data may then be associated with subterranean depth, for example, to enable imaging.
Contact information
Any inquiry concerning this communication or earlier communications from the
examiner should be directed to MOHAMED CHARIOUI whose telephone number is (571)272-2213. The examiner can normally be reached Monday through Friday, from 9 am to 6 pm.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Andrew Schechter can be reached on (571) 272-2302. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
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.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free).
Mohamed Charioui
/MOHAMED CHARIOUI/Primary Examiner, Art Unit 2857
Prosecution Insights