# US 11,443,149 B2 · Cuttings imaging for determining geological properties > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 11,443,149 B2 - Document ID: US11443149B2 - Status: Granted - Kind: B2 - Application number: 17/067,806 - Filed: 2020-10-12 - Published: 2021-10-14 - Granted: 2022-09-13 - Earliest claimed: 2020-04-14 - Assignees: Schlumberger Technology Corporation - Applicants: Schlumberger Technology Corporation - Inventors: Matthias Francois; Youssef Tamaazousti; Josselin Kherroubi - Web page: https://patentgrep.com/patent/US11443149B2 ## CPC - G01V 11/002: {Details, e.g. power supply systems for logging instruments, transmitting or recording data, specially adapted for well logging, also if the prospecting method is irrelevant (means for transmitting well survey signals E21B47/12; signal transmission systems in general G08C; transmission in general H04B)} - E21B 21/06: Arrangements for treating drilling fluids outside the borehole - G01N 15/088: {Investigating volume, surface area, size or distribution of pores; Porosimetry} - G01N 15/1433: using image recognition - G01N 21/255: {Details, e.g. use of specially adapted sources, lighting or optical systems} - G01N 33/24: Earth materials (G01N33/42 takes precedence) - G06F 18/214: Generating training patterns; Bootstrap methods, e.g. bagging or boosting - G06F 18/2413: based on distances to training or reference patterns - G06F 18/2431: Multiple classes - G06V 10/774: Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting - E21B 2200/22: Fuzzy logic, artificial intelligence, neural networks or the like - G01V 20/00: Geomodelling in general - G01V: GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS (means for indicating the location of accidentally buried, e.g. snow-buried, persons A63B29/02) - E21B: EARTH OR ROCK DRILLING; OBTAINING OIL, GAS, WATER, SOLUBLE OR MELTABLE MATERIALS OR A SLURRY OF MINERALS FROM WELLS - G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q) - G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N) - G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING ## Abstract Apparatus and methods for ascribing one of multiple predetermined sub-classes to multiple pixels of an image of an unknown rock sample retrieved from a geological formation. The ascription utilizes a deep learning model trained with an annotated training dataset. The annotated training dataset includes multi-pixel images of known rock samples and, for each known rock sample image, which sub-class corresponds to at least a subset of pixels of that image. For each pixel of the unknown rock sample image having an ascribed sub-class, which one of predetermined meta-classes is associated with that pixel is derived based on the sub-class ascribed to that pixel. The meta-classes represent different predetermined rock types. At least one property of the formation is predicted utilizing the ascription-derived meta-classes, including which rock type(s) are present in the formation. ## Claims ### Claim 1 A method comprising: ascribing one of a plurality of predetermined sub-classes to each of a plurality of pixels of an image of an unknown rock sample (“unknown rock sample image”) retrieved from a geological formation, wherein: the ascription utilizes a deep learning model that is trained with an annotated training dataset; and the annotated training dataset comprises: a plurality of multi-pixel images of known rock samples (“known rock sample images”); and for each known rock sample image, which of the sub-classes corresponds to at least a subset of pixels of that image; for each pixel of the unknown rock sample image having an ascribed sub-class, deriving which one of a plurality of predetermined meta-classes is associated with that pixel based on the sub-class ascribed to that pixel, wherein the meta-classes are each representative of a different one of a plurality of predetermined rock types; and predicting at least one property of the geological formation utilizing the ascription-derived meta-classes, wherein predicting the at least one property includes: predicting which one or more of the rock types are present in the geologic formation; and predicting a quantity of each of the one or more rock types predicted to be present in the geologic formation. ### Claim 2 (depends on 1) The method of claim 1 further comprising training the deep learning model with the annotated training dataset. ### Claim 3 (depends on 1) The method of claim 1 wherein the deep learning model includes a convolutional neural network. ### Claim 4 (depends on 3) The method of claim 3 wherein the deep learning model includes nodes each associated with at least a weight, and wherein the training includes optimizing the weights associated to each node using a loss function relative to the annotated training dataset. ### Claim 5 (depends on 1) The method of claim 1 wherein the deep learning model is defined to perform multi-scale processing. ### Claim 6 (depends on 1) The method of claim 1 further comprising defining a plurality of windows in the unknown rock sample image, wherein each window comprises a predetermined number of pixels and is of a predetermined shape, and wherein each known rock sample image of the annotated training dataset comprises the predetermined number of pixels and is of the predetermined shape. ### Claim 7 (depends on 1) The method of claim 1 further comprising: forming the annotated training dataset, including: obtaining the known rock sample images, wherein the known rock samples depicted in the known rock sample images collectively include one or more known rock types; and defining each sub-class via association with at least one subset of pixels of a corresponding one or more of the known rock sample images; and associating each defined sub-class to one of the meta-classes that corresponds to the known rock type of the corresponding at least one subset of pixels. ### Claim 8 (depends on 7) The method of claim 7 wherein at least one of the known rock sample images is an image of one of the known rock samples that comprises just one of the known rock types, and wherein the at least one subset of pixels of that image are associated with just one of the sub-classes. ### Claim 9 (depends on 1) The method of claim 1 wherein each sub-class is representative of one or more rock properties. ### Claim 10 (depends on 9) The method of claim 9 wherein the one or more rock properties include one or more of grain size, color, and porosity. ### Claim 11 (depends on 10) The method of claim 10 further comprising predicting one or more of grain size, color, and porosity of the geologic formation based on the ascriptions. ### Claim 12 (depends on 1) The method of claim 1 wherein the known rock sample comprises drill cuttings. ### Claim 13 (depends on 12) The method of claim 12 wherein the unknown rock sample is one of a plurality of unknown rock samples each retrieved from a different depth in the geological formation and depicted in a corresponding one of a plurality of unknown rock sample images, and wherein the method further comprises: performing the sub-class ascription, the meta-class derivation, and the at least one property prediction utilizing each of a plurality of unknown rock sample images; and generating a lithology profile of the geologic formation versus depth based on the property predictions. ### Claim 14 (depends on 1) The method of claim 1 further comprising: collecting a sample of drill cuttings retrieved from a wellbore while drilling into the formation to form the wellbore; and via operation of an imaging device, imaging the sample of drill cuttings to obtain the unknown rock sample image. ### Claim 15 (depends on 14) The method of claim 14 wherein parameters of the unknown rock sample image are standardized parameters. ### Claim 16 (depends on 15) The method of claim 15 wherein the standardized parameters include settings of the imaging device operation, including one or more of resolution, brightness, type of support on which the sample of drill cuttings rest during the imaging, and distance between the support and the imaging device. ### Claim 17 An apparatus comprising: a processing system comprising a processor and a memory storing an executable computer program code that, when executed by the processor: ascribes one of a plurality of predetermined sub-classes to each of a plurality of pixels of an image of an unknown rock sample (“unknown rock sample image”) retrieved from a geological formation, wherein: the ascription utilizes a deep learning model that is stored in the memory and that was trained with an annotated training dataset; and the annotated training dataset comprises: a plurality of multi-pixel images of known rock samples (“known rock sample images”); and for each known rock sample image, which of the sub-classes corresponds to at least a subset of pixels of that image; for each pixel of the unknown rock sample image having an ascribed sub-class, derives which one of a plurality of predetermined meta-classes is associated with that pixel based on the sub-class ascribed to that pixel, wherein the meta-classes are each representative of a different one of a plurality of predetermined rock types; and predicts at least one property of the geological formation utilizing the ascription-derived meta-classes, wherein the at least one property includes: which one or more of the rock types are present in the geologic formation; and a quantity of each of the one or more rock types predicted to be present in the geologic formation. ### Claim 18 (depends on 17) The apparatus of claim 17 further comprising an imaging device for obtaining the unknown rock sample image. ### Claim 19 A non-transitory, tangible, computer-readable medium having stored thereon computer-readable instructions executable by a processing system for: ascribing one of a plurality of predetermined sub-classes to each of a plurality of pixels of an image of an unknown rock sample (“unknown rock sample image”) retrieved from a geological formation, wherein: the ascription utilizes a deep learning model that is trained with an annotated training dataset; and the annotated training dataset comprises: a plurality of multi-pixel images of known rock samples (“known rock sample images”); and for each known rock sample image, which of the sub-classes corresponds to at least a subset of pixels of that image; for each pixel of the unknown rock sample image having an ascribed sub-class, deriving which one of a plurality of predetermined meta-classes is associated with that pixel based on the sub-class ascribed to that pixel, wherein the meta-classes are each representative of a different one of a plurality of predetermined rock types; and predicting at least one property of the geological formation utilizing the ascription-derived meta-classes, wherein predicting the at least one property includes: predicting which one or more of the rock types are present in the geologic formation; and predicting a quantity of each of the one or more rock types predicted to be present in the geologic formation. ### Claim 20 A method comprising: ascribing one of a plurality of predetermined sub-classes to each of a plurality of pixels of an image of an unknown rock sample (“unknown rock sample image”) retrieved from a geological formation, wherein: the ascription utilizes a deep learning model that is trained with an annotated training dataset; and the annotated training dataset comprises: a plurality of multi-pixel images of known rock samples (“known rock sample images”), wherein the known rock sample comprises drill cuttings; and for each known rock sample image, which of the sub-classes corresponds to at least a subset of pixels of that image; for each pixel of the unknown rock sample image having an ascribed sub-class, deriving which one of a plurality of predetermined meta-classes is associated with that pixel based on the sub-class ascribed to that pixel, wherein the meta-classes are each representative of a different one of a plurality of predetermined rock types; and predicting at least one property of the geological formation utilizing the ascription-derived meta-classes, wherein predicting the at least one property includes predicting which one or more of the rock types are present in the geologic formation; wherein: the unknown rock sample is one of a plurality of unknown rock samples each retrieved from a different depth in the geological formation and depicted in a corresponding one of a plurality of unknown rock sample images; and the method further comprises: performing the sub-class ascription, the meta-class derivation, and the at least one property prediction utilizing each of the plurality of unknown rock sample images; and generating a lithology profile of the geologic formation versus depth based on the property predictions.