Systems and methods for improved sample imaging and analysis

Abstract

Provided herein are methods and systems for improved sample imaging and analysis. An input image of a sample comprising a plurality of elements may be received. The input image may be partitioned into a plurality of classes. Each class of the plurality of classes may be associated with at least one element of the plurality' of elements. Based on the input image and the plurality of classes, a cluster-dominant image may be generated. Based on the cluster-dominant image, a false-color element map and a histogram may be generated. The false-color element map and the histogram may each be indicative of a relative abundance of each element of the plurality of elements within the sample.

Description

SYSTEMS AND METHODS FOR IMPROVED SAMPLE IMAGING AND ANALYSISCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to U.S. Provisional Application No. 63 / 607,616, filed on December 8, 2023, which is incorporated by reference in its entirety herein.BACKGROUND

[0002] Visual analysis of core samples is one of the oldest tools used in the geological analysis of drill hole data. Logs of geological domain, lithology, mineral ogy, and estimates of the concentration of elements or minerals of interest are routinely performed by visual inspection of core at surface. However, visual analysis alone is time-consuming, may be error-prone, etc. These and other considerations are discussed herein.SUMMARY

[0003] It is to be understood that both the following general description and the following detailed description are exemplary and explanatory only and are not restrictive. Provided herein are methods and systems for improved sample imaging and analysis. An input image of a sample comprising a plurality of elements may be received by a computing device. The input image may be partitioned into a plurality of classes, where each class may be associated with at least one element of the plurality of elements. Based on the input image and the plurality of classes, a cluster-dominant image may be generated. Based on the cluster-dominant image, a false-color element map and a histogram may be generated. The false-color element map and the histogram may each be indicative of a relative abundance of each element of the plurality of elements within the sample.

[0004] Additional advantages will be set forth in part in the description which follows or may be learned by practice. The advantages will be realized and attained by means of the elements and combinations particularly pointed out in the appended claims.BRIEF DESCRIPTION OF THE DRAWINGS

[0005] The accompanying drawings, which are incorporated in and constitute a part of the present description serve to explain the principles of the methods and systems described herein:

[0006] FIG. 1 shows an example system for improved sample imaging and analysis.

[0007] FIG. 2 shows an example system.

[0008] FIG. 3 shows an example selection from an example input image.

[0009] FIG. 4 shows an example system for training a machine learning module.

[0010] FIG. 5 shows an example process flowchart illustrating an example training method for generating the machine learning module using the training module.

[0011] FIG. 6 shows example inputs and outputs of a machine learning module.

[0012] FIG. 7 shows example cluster data and false-color mapping.

[0013] FIG. 8 shows example inputs and outputs of a computing device.

[0014] FIG. 9A shows an example representation of an input with cluster labels.

[0015] FIG. 9B shows an example aggregation of clusters.

[0016] FIG. 10 shows example aggregation data.

[0017] FIG. 11 shows a block diagram depicting an environment comprising non-limiting examples of a computing device and a server connected through a network.

[0018] FIG. 12 shows a flowchart for an example method for improved sample imaging and analysis.DETAILED DESCRIPTION

[0019] As used in the specification and the appended claims, the singular forms “a,” “an,” and “the” include plural referents unless the context clearly dictates otherwise. Ranges may be expressed herein as from “about” one particular value, and / or to “about” another particular value. When such a range is expressed, another configuration includes from the one particular value and / or to the other particular value. Similarly, when values are expressed as approximations, by use of the antecedent “about,” it will be understood that the particular value formsanother configuration. It will be further understood that the endpoints of each of the ranges are significant both in relation to the other endpoint, and independently of the other endpoint.

[0020] “Optional’' or “optionally” means that the subsequently described event or circumstance may or may not occur, and that the description includes cases where said event or circumstance occurs and cases where it does not.

[0021] Throughout the description and claims of this specification, the word “comprise” and variations of the word, such as “comprising” and “comprises,” means “including but not limited to,” and is not intended to exclude, for example, other components, integers or steps. “Exemplary” means “an example of’ and is not intended to convey an indication of a preferred or ideal configuration. “Such as” is not used in a restrictive sense, but for explanatory purposes.

[0022] It is understood that when combinations, subsets, interactions, groups, etc. of components are described that, while specific reference of each various individual and collective combinations and permutations of these may not be explicitly described, each is specifically contemplated and described herein. This applies to all parts of this application including, but not limited to, steps in described methods. Thus, if there are a variety of additional steps that may be performed it is understood that each of these additional steps may be performed with any specific configuration or combination of configurations of the described methods.

[0023] As will be appreciated by one skilled in the art, hardware, software, or a combination of software and hardware may be implemented. Furthermore, a computer program product on a computer-readable storage medium (e.g., non- transitory) having processor-executable instructions (e.g., computer software) embodied in the storage medium. Any suitable computer-readable storage medium may be utilized including hard disks. CD-ROMs, optical storage devices, magnetic storage devices, memristors, Non-Volatile Random Access Memory (NVRAM), flash memory, or a combination thereof.

[0024] Throughout this application reference is made to block diagrams and flowcharts. It will be understood that each block of the block diagrams and flowcharts, and combinations of blocks in the block diagrams and flowcharts, respectively, may be implemented by processor-executable instructions. Theseprocessor-executable instructions may be loaded onto a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the processor-executable instructions which execute on the computer or other programmable data processing apparatus create a device for implementing the functions specified in the flowchart block or blocks.

[0025] These processor-executable instructions may also be stored in a computer-readable memory' that may direct a computer or other programmable data processing apparatus to function in a particular manner, such that the processor-executable instructions stored in the computer-readable memory produce an article of manufacture including processor-executable instructions for implementing the function specified in the flowchart block or blocks. The processor-executable instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the processor-executable instructions that execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks.

[0026] Blocks of the block diagrams and flowcharts support combinations of devices for performing the specified functions, combinations of steps for performing the specified functions and program instruction means for performing the specified functions. It will also be understood that each block of the block diagrams and flowcharts, and combinations of blocks in the block diagrams and flowcharts, may be implemented by special purpose hardware-based computer systems that perform the specified functions or steps, or combinations of special purpose hardware and computer instructions.

[0027] The word “sample"’ as used herein may refer to a part, a piece, a chip, a portion, a cutting, a shaving, a mass, a chunk, a combination thereof, and / or the like, of one or more geologic materials. For example, a sample may comprise a part, a piece, a chip, a portion, a cutting, a shaving, a mass, a chunk, a combination thereof, and / or the like, of ore, a rock(s), a mineral(s), an element(s), or any other organic (or inorganic) matter, sample The following list is only meant to be an illustrative listing of possible types / examples of a “sample” asused herein, and this list is not intended to be exhaustive: a "sample" (e.g., a cylindrical section(s) of geologic material(s)); a “drill core" (e.g., long sections of geologic material(s) extracted using a core drill); a “cutting sample’’ (e.g., small pieces of geologic material(s) or sediment type(s); a “grab sample” (e.g., geologic material(s) taken directly from a face(s), an outcrop(s), or a deposit(s)); a “bulk sample” (e.g., a large collection of geologic material(s)); a “composite sample” (e.g.. a combination of two or more geologic materials); a “sludge sample” (e.g.. fine portions of geologic material(s) collected from a fluid(s)); a “soil sample” (e.g., geologic material(s) taken from the surface or a near-surface layer(s)); a “sediment sample” (e g., geologic material(s) collected from a floor / bottom of a body of water); and a “channel sample” (e.g., geologic material(s) collected from a linear trench or channel). Additionally, as used herein, a “sample” may comprise more than one type / example from the aforementioned list.

[0028] The word “element” as used herein may refer to a type, or a portion, of a mineral, a rock, ore, or any other organic (or inorganic) material that may be present within a sample, affixed to a sample, or adjacent a sample. The phrase “plurality of elements” as used herein may refer to a plurality of types of minerals, ore, rocks, or any other organic (or inorganic) material within a sample, affixed to a sample, or adjacent a sample

[0029] FIG. 1 shows an example system 100 for improved sample imaging and analysis. The system 100 may include ajob / excavation site 102 having a computing device(s), such as one or more imaging devices, capable of capturing / generating images of samples. For example, the one or more imaging devices may be configured to capture red-green-blue (RGB) images / data of the samples as further described herein. The computing device(s) at the job / excavation site 102 may provide (e.g.. upload) such images to a server 104 via a network. The network may facilitate communication between each device / entity of the system 100. The network may be an optical fiber network, a coaxial cable network, a hybrid fiber-coaxial network, a wireless network, a satellite system, a direct broadcast system, an Ethernet network, a high-definition multimedia interface network, a Universal Serial Bus (USB) network, or any combination thereof. Data may be sent / received via the network by any device / entity of the system 100 via a variety of transmission paths, includingwireless paths (e.g., satellite paths, Wi-Fi paths, cellular paths, etc.) and terrestrial paths (e.g., wired paths, a direct feed source via a direct line, etc.).

[0030] The server 104 may be a single computing device or a plurality7of computing devices. As shown in FIG. 1, the server may include a storage module 104A and a machine learning module 104B. The storage module 104A may comprise one or more storage repositories that may be local, remote, cloud-based, a combination thereof, and / or the like. The machine learning module 104B, which is discussed further herein, may be configured (e.g., trained) to generate a false- color element map of a sample based on an input image(s) of the sample. The input image(s) may comprise an RGB image(s) of the sample (e.g., a high- resolution line-scan(s) of a core tray(s) with the sample). In some examples, the machine learning module 104B may generate the false-color element map of the sample based on the input image(s) of the sample. Additionally, or in the alternative, the machine learning module 104B may generate the false-color element map of the sample based on the input image(s) of the sample along with additional data related to the sample (e.g., LiDAR-guided x-ray fluorescence (XRF) data). The process for capturing input images and additional data used when generating false-color element maps is discussed further herein.

[0031] Returning to FIG. 1, the system 100 may also include a computing device 106. The computing device 106 may be in communication with the server 104 and / or the computing device(s) at the job / excavation site 102. Analysis of input images of samples may be facilitated using a web-based or locally-installed application, such as a structural logging application (hereinafter an “application”, such as an application associated with the Minalyzer CS™ core scanning system), executing or otherwise controlled by the computing device 106. The computing device 106 may use the application to determine structural data associated using one or more images of each sample.

[0032] In some examples, the system 100 may receive one or more of orientation data, survey data. XRF data, exclusion zone data, a combination thereof, and / or the like, associated with each sample (collectively, “additional data”). The additional data may be provided to the system 100 by the user via the computing device 106, by the server 104, or by a third-party computing device (not shown). In some examples, the additional data may be acquired and / orderived using the one or more imaging devices described herein. The orientation data may be indicative of an orientation, a depth, etc., of samples at an extraction point (e.g., a borehole). The orientation data may be indicative of one or more sine waves, strike angles, dip angles, an azimuth, etc. associated with each sample.

[0033] The XRF data for a sample (e g., LiDAR-guided XRF data) may be at least partially indicative of a plurality of elements that make up the sample. The exclusion zone data may be at least partially indicative of at least one portion of a sample (and / or input image) that is to be excluded from analysis (e.g., due to physical characteristics of the sample, such as fractures, breaks, etc.). The examples above are meant to be exemplary only. The additional data may comprise further information / data related to the samples as well. The structural data, the additional data, the exclusion zone data, etc., for each sample may be stored at the server 104 and / or at the computing device 106.

[0034] FIG. 2 shows an example system 200. The job / excavation site 102 may comprise one or more components of the system 200 in some examples. Additionally, or in the alternative, one or more components of the system 200 may be located elsewhere. As shown in FIG. 2, the system 200 may comprise an imaging apparatus 204, which may comprise an imaging device(s) 204A, such as a high-resolution line scan camera(s). The one or more imaging devices at the job / excavation site 102 described herein may comprise the imaging apparatus 204 and / or the imaging device(s) 204A. It is to be understood that the imaging apparatus 204 and the imaging device(s) 204A shown in FIG. 2 are exemplary only and not meant to be restrictive. For example, the imaging apparatus 204 and the imaging device(s) 204A may be enclosed within, or components of, a core scanning system (e.g., the Minalyzer CS™ core scanning system).

[0035] The imaging device(s) 204A may comprise a series of optical sensors that may capture one or more images of the sample 202. (e.g., high-resolution line scan images) Note that the sample 202 is shown in FIG. 2 as being a split / open sample for exemplary purposes only. The imaging device(s) 204A may capture one or more images of whole samples as well. The imaging device(s) 204A may capture one or more images of the entire sample 202 or only a portion(s) of the sample 202.

[0036] In some examples, as described herein, the imaging device(s) 204A may comprise a high-resolution line scan camera(s), which may be used to capture one or more images of the sample 202. The imaging device(s) 204A may capture the one or more image(s) of the sample 202 one line at a time, which may be as narrow as a few millimeters (e.g., resulting in a linear array of pixels instead of a two-dimensional array that a conventional camera may capture). The sample 202 (and / or the imaging device(s) 204A) may move continuously during the imaging process, and the imaging device(s) 204A may capture one line (e g., row of pixels) at a time. The individual lines captured may then be assembled to generate a complete image, referred to herein as an “RGB image.'’ As noted above and as further described herein, the machine learning module 104B may be configured to generate a false-color element map associated with a sample, such as the sample 202, based on an RGB image(s) (e.g., captured with a high-resolution line scan camera(s)).

[0037] FIG. 3 shows an example selection from an example input image. Specifically, FIG. 3 shows an example image 302 of a plurality of samples (e.g., within a core box). The image 302 may be an RGB image captured using the imaging device(s) 204A as described herein. The image 302 may be depth- registered based on a corresponding depth (or range) of a borehole (or similar) from which the plurality of samples were extracted. Each sample may be separately depth-registered. One or more portions 306 of a sample(s) (e.g., of any desired interval / length) may be extracted from the image 302, such as via processing software and / or the application described herein. As described herein, in some examples, the system 100 may receive exclusion zone data associated with each sample (e.g., as part of the “additional data” described herein). In such examples, the processing software and / or the application that extracts the one or more portions 306 from the image 302 may use corresponding exclusion zone data to determine one or more portions of the image 302 that are not to be extracted (e.g., one or more portions of the corresponding sample(s) that are not to be analyzed). Additionally, or in the alternative, one or more samples, the plurality of samples, within the image 302 may be comprise a physical marking(s) and / or a rendered marking(s) (e.g.. on or within the image 302 itself) to indicate the one or more portions of the image 302 that are not to be extracted (e.g., amanually-marked exclusion zone(s)). The one or more portions 306 that are shown in FIG. 3 are 0.5 feet each, although other lengths / portions may be used as well.

[0038] FIG. 4 shows a system 400 for training a machine learning module 430. The machine learning module 430 may comprise the machine learning module 104B. The machine learning module 430 may be trained by a training module 420 of the system 400 to generate a cluster-dominant image(s) based on one or more input images of a sample. The cluster-dominant image(s), as further described herein, may represent a plurality of segmented classes (e.g., clusters) of pixels within the one or more input images, each class / cluster being associated with one or more particular element(s) present within the sample. The machine learning module 430 may generate a false-color element map for the sample (e.g., based on the cluster-dominant image(s) and corresponding cluster data (e.g., cluster mapping, etc.).

[0039] As further described below, the machine learning module 104B may analyze one or more images of a sample(s), such as the image 302 and / or the one or more portions 306. using a segmentation model and / or algorithm to classify each element of a plurality of elements that may be present therein. For example, one or more images of a sample(s) may be analyzed by the machine learning module 104B to generate the false-color element map, which may be indicative of a plurality of elements therein, and a corresponding histogram indicative of a relative abundance of each element. In some examples, in addition to being based on the one or more images of the sample(s), the false-color element map generated by the machine learning module 104B may also be based on additional data related to the sample, such as LiDAR-guided x-ray fluorescence (XRF) data.

[0040] The training module 420 may use machine learning techniques to train, based on an analysis of one or more training datasets 410, the machine learning module 430. The training dataset 410 may comprise any number of datasets or subsets 410-410N. For example, the training dataset 410 may comprise a first training dataset 410A and a second training dataset 410B. The training module 420 may use a supervised, semi-supervised, or unsupervised training method, or a combination thereof, depending on the training dataset 410. For example, the training dataset 410 may comprise, for each sample, input data. The input datamay comprise at least one input / RGB image of the respective sample. Additionally, or in the alternative, the input data may comprise the at least one input / RGB image of the respective sample as well as the additional data described herein (e.g., XRF data). In examples where the training dataset 410 comprises ground truth data for a respective training sample the training module 420 may use a supervised training method. The ground truth data for a training sample may refer to one or more sources of data that confirm a composition of the training sample. Examples of ground truth data may be expert annotation, such as annotation / identification of classes of elements within the training sample; XRF data associated with the training sample; hyperspectral data (e.g., hyperspectral images and / or related data) associated with the training sample; a false-color element map(s) associated with the training sample (e.g., a previously-generated false-color element map(s) associated with a known level of accuracy). In examples where the training dataset 410 does not include such ground truth data, the training module 420 may use an unsupervised training method. Other examples, such as for semi-supervised training, are possible as well.

[0041] The first training dataset 410A and the second training dataset may each comprise, for each sample used for training, at least one RGB image of the respective sample (referred to as an ‘'original image(s) in FIG. 4); a clusterdominant image(s) associated with the sample; cluster data corresponding to clusters of pixels associated with each element present therein; and (in some examples) the additional data described herein (e.g., LiDAR-guided XRF data). A subset of one or both of the first training dataset 410A or the second training dataset 410B may be randomly assigned to a testing dataset. In some implementations, the assignment to a testing dataset may not be completely random. In this case, one or more criteria may be used during the assignment. In general, any suitable method may be used to assign data to the testing dataset, while ensuring that the distributions of input data are properly assigned for training and testing purposes.

[0042] The machine learning module 430 may comprise a segmentation model (or models). The segmentation model may be trained by applying one or more segmentation algorithms to the training images. In some examples, the RGB / input images may be pre-processed. For example, each image’s brightness across adepicted sample / core may be normalized with respect to brightness of other images used. As another example, each image may be analyzed to determine characteristics of the RGB color space (e.g., color channel values), and HSV (Hue, Saturation, and Value) and / or CIE-LAB color spaces / values may be derived from the image. As noted above with respect to FIG. 3, a selection of one or more portions 306 of a sample(s) (e.g.. of any desired interval / length) may be extracted from the image 302, such as via processing software and / or the application described herein, and the selected one or more portions may be used for training. The segmentation model may comprise a “thresholding” model that, as further described herein, may partition an image of a sample into a number of classes (e.g., using Otsu's method, etc.) based on the particular elements present within the corresponding sample.

[0043] The training module 420 may train the machine learning module 430 by extracting a feature set from the training datasets 410 according to one or more feature selection techniques. For example, the training module 420 may train the machine learning module 430 by extracting a feature set from the training datasets 410 that includes statistically significant features. The training module 420 may extract a feature set from the training datasets 410 in a variety of ways. The training module 420 may perform feature extraction multiple times, each time using a different feature-extraction technique. In an example, the feature sets generated using the different techniques may each be used to generate different machine learning-based models 440A-440N. For example, the feature set with the highest quality metrics may be selected for use in training. The training module 420 may use the feature set(s) to build one or more machine learning-based models 440A-440N, each of which may be the machine learning module 104B or a component / piece thereof.

[0044] The training datasets 410 may be analyzed to determine any dependencies, associations, and / or correlations between determined features in unlabeled input data and the features of labeled input data in the training dataset 410. The identified correlations may have the form of a list of features. The term “feature,” as used herein, may refer to any characteristic of an item of data that may be used to determine whether the item of data falls within one or more specific categories. A feature selection technique may comprise one or morefeature selection rules. The one or more feature selection rules may comprise a feature occurrence rule. The feature occurrence rule may comprise determining which features in the training dataset 410 occur over a threshold number of times and identifying those features that satisfy the threshold as features.

[0045] After the training module 420 has generated a feature set(s), the training module 420 may generate each machine learning-based model 440 based on the feature set(s). The training module 420 may use the feature sets determined or extracted from the training dataset 410 to build the machine learning-based models 440A-440N. In some examples, the machine learning-based models 440A-440N may be combined into a single machine learning-based model 440. Similarly, the machine learning module 430 may represent a single classifier containing a single or a plurality of machine learning-based models 440 and / or multiple classifiers containing a single or a plurality of machine learning-based models 440.

[0046] FIG. 5 shows an example process flowchart illustrating an example training method 500 for generating the machine learning module 430 using the training module 420. The training module 420 can implement supervised, unsupervised, and / or semi-supervised (e.g., reinforcement based) machine learning-based models 440. The method 500 illustrated in FIG. 5 is an example of a supervised learning method; variations of this example of training method are discussed below, however, other training methods can be analogously implemented to train unsupervised and / or semi-supervised machine learning models. The training method 500 may determine (e.g., access, receive, retrieve, etc.) input data at step 510. The input data may comprise at least one RGB image of a respective sample (referred to as an “original image(s) in FIG. 4); a clusterdominant image(s) associated with the sample; cluster data corresponding to clusters of pixels associated with each element present therein; the additional data described herein (e.g., XRF data); a combination thereof, and / or the like.

[0047] The training method 500 may generate, at step 520. a training dataset and a testing data set. The training dataset and the testing data set may be generated by randomly assigning portions of the input data to either the training dataset or the testing data set. In some implementations, the assignment of input data as training or testing data may not be completely random. The training method 500may determine (e.g., extract, select, etc.), at step 530, one or more features. As an example, the training method 500 may determine a set of features from the input data. The training method 500 may train one or more machine learning models using the one or more features at step 540. In one example, the machine learning models may be trained using supervised learning. In another example, other machine learning techniques may be employed, including unsupervised learning and semi-supervised. The machine learning models trained at 540 may be selected based on different criteria depending on the problem to be solved and / or data available in the training dataset. For example, machine learning classifiers can suffer from different degrees of bias. Accordingly, more than one machine learning model can be trained at 540, optimized, improved, and cross-validated at step 550.

[0048] The training method 500 may select one or more of the machine learning models trained at step 550 to build a final model at 560. The final model may be evaluated using the testing data set. The final model may analyze the testing data set and generate testing false-color element maps at step 570. The testing maps may be evaluated at step 580 to determine whether they meet a desired accuracy level compared to ground truth to determine how accurate the testing maps are of the actual, ground truth. Performance of the final model may be evaluated in a number of ways, as can be appreciated by those skilled in the art. When a desired accuracy level is reached, the final model (e.g., the trained machine learning module 430) may be output at step 590. When the desired accuracy level is not reached, then a subsequent iteration of the training method 500 may be performed starting at step 510 with variations such as, for example, considering a larger collection of training data.

[0049] FIG. 6 shows example inputs and outputs of a machine learning module. Specifically, FIG. 6 shows an example input(s) 610 and example output(s) 620 of the machine learning module 430 (e.g., the machine learning module 104B) once it has been trained as described herein. As shown in FIG. 6, the inputs 610 may comprise an input image 612 of a sample, such as the image 302 and / or the one or more portions 306. As shown in FIG. 6, the input image 612 may be depth- registered based on a corresponding depth (or range) of a borehole (or similar) from which the depicted sample(s) was extracted. Additionally, in someexamples, the inputs 610 may comprise additional data 614 associated with the sample, such as LiDAR-guided x-ray fluorescence (XRF) data.

[0050] As described herein, the machine learning module 430 may be trained to generate a cluster-dominant image(s) based on one or more input images of a sample. The cluster-dominant image(s), as described herein, may represent a plurality of classes (e.g., clusters) of pixels within the one or more input images, where each class / cluster is associated with one or more particular element(s) present within the sample.

[0051] Returning to FIG. 6, the outputs 620 may comprise a cluster-dominant image 622. The cluster-dominant image 622 may comprise a modified version of the input image 612 that shows segmented classes (e.g., groups of pixels) within the input image 612. As shown in FIG. 6, the cluster-dominant image 622 may be depth-registered as well (e.g., based on the corresponding depth (or range) of the borehole (or similar) from which the depicted sample(s) was extracted). The machine learning module 430 may generate the cluster-dominant image 622 using a segmentation model. The segmentation model may comprise a “thresholding"’ model that may partition the input image 612 into a number of classes (e.g., using Otsu’s method, etc.). The sample(s) depicted in the input image 612 may comprise a plurality of elements. The plurality of elements may comprise one or more minerals, ore types, rocks, and / or any other organic (or inorganic) material within the sample(s).

[0052] The segmentation model of the machine learning module 430 may partition the input image 612 into the number of classes based on a multi-level thresholding technique. The multi-level thresholding technique may divide the input image 612 into distinct classes that each represent a particular element (e.g., mineral) of the plurality7of elements or a particular combination of multiple elements (e.g., a mixture of two or more minerals; one mineral and one or more non-minerals; etc.). For example, and for purposes of explanation rather than limitation, the input image 612 may depict a sample comprising a lithium deposit(s) interspersed with dirt, mud, ash, clay, other elements and / or minerals, etc. Each of the classes determined using the multi-level thresholding technique may comprise a particular abundance, or concentration, of an element of interest (e.g., lithium, a lithium salt, etc., to use the same example as above). The multi-level thresholding technique may determine (e.g., calculate) multiple thresholds for partitioning the input image 612 into the distinct classes such that intra-class variance is minimized and / or inter-class variance is maximized. The segmentation model of the machine learning module 430 may partition the input image 612 into the distinct classes using the multi-level thresholding technique by iterating through all possible threshold values and determining / calculating the variances for each.

[0053] The input image 612 may be an “RGB image” as described herein, and the segmentation model may partition the input image 612 into the distinct classes using the multi-level thresholding technique by iterating through all possible threshold values and determining / calculating the variances for each based on the corresponding pixel values in the RGB color spaces / channels. Additionally, in some examples, the segmentation model may also use the HSV (Hue, Saturation. Value channels) and CIE-LAB color spaces / channels.

[0054] For example, the segmentation model may convert the input image 612 from the RGB to the HSV color space. The HSV color space may separate the color (e.g.. hue channel) of each pixel of the input image 612 from its intensity (e.g., value). Additionally, or in the alternative, the segmentation model may convert the input image 612 to the CIE-LAB color space by separating pixel values associated with luminance (e.g., L*) from color information (e.g., a* for green-red, and b* for blue-yellow).

[0055] The segmentation model may apply multi-level thresholding to each of the RGB channels by determining thresholds within each channel to segment the input image 612 based on red, green, and blue components thereof (e.g., within each pixel). The segmentation model may apply multi-level thresholding to each channel of the HSV color space by determining thresholds within the hue channel to separate different colors; determining thresholds within the saturation channel to differentiate color intensity’, and / or determining thresholds within the value channel for brightness. The segmentation model may apply multi-level thresholding to each channel in the CIE-LAB color space (the L*, a*, and b* channels), which may allow for segmentation based on both color and luminance differences between pixels.

[0056] The segmentation model of the machine learning module 430 may combine the thresholds in each color space to segment the input image 612 into the distinct classes that each represent a particular element (e.g., mineral) of the plurality of elements or a particular combination of multiple elements. For example, the input image 612 may depict a sample comprising a deposit(s) of an element of interest (e.g., lithium, a lithium salt, etc.) interspersed with dirt, mud, ash, clay, other elements and / or minerals, etc. Thus, each of the distinct classes may comprise a particular abundance, or concentration, of the element of interest (e.g., some classes may comprise only the element of interest, while others may comprise as little as none).

[0057] As shown in FIG. 6, the outputs 620 may comprise a cluster-dominant image 622. The cluster-dominant image 622 may comprise a modified version of the input image 612 that shows the distinct classes more clearly. Additionally, in some examples, the outputs 620 may comprise cluster data 624, which may comprise a cluster mapping indicative of and / or defining each of the distinct classes (e.g., clusters) of pixels within the input image 612. FIG. 7 shows example cluster data and false-color mapping. Specifically, an example cluster mapping 702 is shown in FIG. 7. The cluster mapping 702 shown in FIG. 7 indicates 10 total classes within the input image 612. However, the cluster mapping 702 shown in FIG. 7 is meant to be exemplary only (e.g., more, or less than, 10 total classes is possible depending on the particular input image(s) used).

[0058] The cluster mapping 702 may comprise one color per class (e.g., 10 total colors in the example shown in FIG. 7). The machine learning module 430 may generate the cluster-dominant image 622 (e.g., a modified version of the input image 612 to show the distinct classes more clearly) by converting pixels within each of the classes to the color corresponding to that particular class based on the cluster mapping 702. Expert annotation may be used to associate each of the class colors within the cluster mapping 702 with a known or expected abundance of the particular element of interest. For example, the expert annotation may comprise manually annotating each of the classes within the cluster mapping 702 using the processing software and / or the application described herein (e.g., a structural logging application, etc.).

[0059] Each annotated class within the cluster mapping 702 may be associated with a known or expected abundance of the particular element of interest (e.g., the element of interest, or a particular abundance of it, may be discernable upon visual inspection of each class color). Additionally, or in the alternative, a known or expected abundance of the particular element of interest within each of the classes within the cluster mapping 702 may be derived using XRF data, hyperspectral data, a combination thereof, and / or the like. The class labeled as “10” in the cluster mapping 702 may be representative of a highest abundance of the element of interest (e.g., as high as 100%), while the class labeled as “1” in the cluster mapping 702 may be representative of a lowest abundance of the element of interest (e.g., as low as 0%).

[0060] After the known or expected abundance of the particular element of interest within each of the classes shown in the cluster mapping 702 are labeled (e.g.. via expert annotation ) or determined / derived (e.g.. via XRF data, hyperspectral data, a combination thereof, and / or the like), the cluster mapping 702 may function as a calibration (e.g., a reference, baseline, etc.) for the borehole and / or job / excavati on site 102 associated with the corresponding sample. For example, the particular borehole and / or job / excavation site 102 may be at a first geographic location. And the first geographic location may be associated with one or more particular elements (e.g., such as volcanic ash or certain compositions of clay, etc.) that are typically present within samples extracted therefrom; however, those one or more particular elements may not be the at least one element of interest (e.g.. lithium, a lithium salt, etc.). Therefore, the cluster mapping 702 associated with the first geographic location for the corresponding sample may function as a calibration (e.g., a reference, baseline, etc.) for additional samples / images and corresponding cluster mappings associated with the same borehole and / or job / excavation site 102 at the first geographic location. In such examples, the cluster mapping 702 may be used as a basis for associating a known or expected abundance of the particular element of interest with each class of the cluster mappings for those additional samples / images (e.g., based on the colors used for the classes within the cluster mapping 702 and the cluster mappings for the additional samples / images).

[0061] As described herein, the machine learning module 430 may generate a false-color element map for each sample. For example, returning to FIG. 7, the machine learning module 430 may convert the cluster mapping 702 into a false- color cluster mapping 704. The cluster mapping 702 may use colors present within the input image 612 and / or the cluster-dominant image 622, while the false-color cluster mapping 704 may use colors that are not present within either the input image 612 or the cluster-dominant image 622 (e.g.. for easier interpretation). Based on the false-color cluster mapping 704, the machine learning module 430 may generate cluster labels 706 as shown in FIG. 7. The cluster labels 706 may be used to generate a false-color element map indicative of the plurality of elements depicted in the input image 612. In some examples, the cluster labels 706 may also be used to generate a histogram indicative of a relative abundance of each element of the plurality of elements within the corresponding sample.

[0062] FIG. 8 shows example inputs and outputs of a computing device. Specifically, FIG. 8 shows an example input(s) 810 and example output(s) 820 of a computing device 702 (e.g., the computing device 106, the server 104, etc.). As shown in FIG. 8, the computing device 702 may receive, as the input(s) 810, the cluster-dominant image 622 and the cluster labels 706. The computing device 702 may use the cluster labels 706 to convert the colors shown in the cluster-dominant image 622, resulting in the output(s) 820. The output(s) 820 may comprise a false-color element map 822 indicative of the plurality of elements depicted in the input image 612. The output(s) 820 may further comprise a histogram 824 indicative of a relative abundance of each element of the plurality of elements within the corresponding sample. The histogram 824 may be based on the colors corresponding to the cluster labels 706 and the false-color element map 822. The output(s) 820 may further comprise aggregation data 826, which may be indicative of the relative abundance of each element of the plurality of elements within the corresponding sample.

[0063] FIG. 9A shows an example representation of an input with cluster labels and FIG. 9B shows an example aggregation of clusters. Specifically, FIG. 9A shows a larger version of the false-color element map 822 and FIG. 9B shows a larger version of the histogram 824. As shown in FIG. 9B, the histogram 824 maycomprise a column of data indicating a relative abundance of a particular element within the sample. In the example shown in FIG. 9B, the histogram 824 indicates a relative abundance 902 of an element between a first depth 904A and a second depth 904B (e.g., based on the aggregation data 826). Though the histogram 824 in FIG. 9B only shows one column (e.g., for the element of interest described above), it is to be understood that a column may be shown for each element of the plurality of elements. FIG. 10 shows example aggregation data. Specifically. FIG. 10 shows an example data table 1002 based on the aggregation data 826. The data table 1002 may indicate a relative abundance of each element within each cluster of the sample for a plurality of depth ranges (first column). Though data table 1002 in FIG. 10 only shows data for one element (e.g., for the element of interest described above), it is to be understood that both the data table 1002 and the aggregation data 826 may include data for each of the elements.

[0064] As discussed herein, the present methods and systems may be computer- implemented. FIG. 11 shows a block diagram depicting an environment 1100 comprising non-limiting examples of a computing device 1101 and a server 1102 connected through a network 1104. As an example, the server 104 and / or the computing device 106 of the system 100 may be a computing device 1101 and / or a server 1102 as described herein with respect to FIG. 11. In an aspect, some or all steps of any described method may be performed on a computing device as described herein. The computing device 1101 can comprise one or multiple computers configured to store one or more of the training module 1120, training data 1110, and the like. The server 1102 can comprise one or multiple computers configured to store sample data 1124. Multiple servers 1102 can communicate with the computing device 1101 via the network 1104.

[0065] The computing device 1101 and the server 1102 can be a digital computer that, in terms of hardware architecture, generally includes a processor 1108, memory system 1110, input / output (I / O) interfaces 1112, and network interfaces 1114. These components (1108. 1110, 1112, and 1114) are communicatively coupled via a local interface 1116. The local interface 1116 can be, for example, but not limited to, one or more buses or other wired or wireless connections, as is known in the art. The local interface 1116 can have additional elements, which are omitted for simplicity, such as controllers, buffers (caches),drivers, repeaters, and receivers, to enable communications. Further, the local interface may include address, control, and / or data connections to enable appropriate communications among the aforementioned components.

[0066] The processor 1108 can be a hardware device for executing software, particularly that stored in memory system 1110. The processor 1108 can be any custom made or commercially available processor, a central processing unit (CPU), an auxiliary processor among several processors associated with the computing device 1101 and the server 1102, a semiconductor-based microprocessor (in the form of a microchip or chip set), or generally any device for executing software instructions. When the computing device 1101 and / or the server 1102 is in operation, the processor 1108 can be configured to execute software stored within the memory system 1110, to communicate data to and from the memory system 1110. and to generally control operations of the computing device 1101 and the server 1102 pursuant to the software.

[0067] The I / O interfaces 1112 can be used to receive user input from, and / or for providing system output to, one or more devices or components. User input can be provided via, for example, a keyboard and / or a mouse. System output can be provided via a display device and a printer (not shown). I / O interfaces 1112 can include, for example, a serial port, a parallel port, a Small Computer System Interface (SCSI), an infrared (IR) interface, a radio frequency (RF) interface, and / or a universal serial bus (USB) interface.

[0068] The network interface 1114 can be used to transmit and receive from the computing device 1101 and / or the server 1102 on the network 1104. The network interface 1114 may include, for example, an Ethernet Adaptor, a Token Ring Adaptor, a wireless network adapter (e.g., WiFi, cellular, satellite), or any other suitable network interface device. The network interface 1114 may include address, control, and / or data connections to enable appropriate communications on the network 1104.

[0069] The memory system 1110 can include any one or combination of volatile memory7elements (e.g., random access memory (RAM, such as DRAM, SRAM, SDRAM, etc.)) and nonvolatile memory elements (e.g., ROM, hard drive, tape, CDROM, DVDROM. etc.). Moreover, the memory system 1110 may incorporate electronic, magnetic, optical, and / or other ty pes of storage media. Note that thememory7system 1110 can have a distributed architecture, where various components are situated remote from one another, but can be accessed by the processor 1108.

[0070] The software in memory system 1110 may include one or more software programs, each of which comprises an ordered listing of executable instructions for implementing logical functions. In the example of FIG. 11, the software in the memory system 1110 of the computing device 1101 can comprise the training module 420 (or subcomponents thereof), the training dataset 410A, the training dataset 410B, and a suitable operating system (O / S) 1118. In the example of FIG. 11. the software in the memory system 1110 of the server 1102 can comprise, the sample data 1124, and a suitable operating system (O / S) 1118. The operating system 1118 essentially controls the execution of other computer programs and provides scheduling, input-output control, file and data management, memory' management, and communication control and related services.

[0071] The environment 1100 may be cloud-based (e.g., disparately-located client devices with respect to servers). For example, the environment 1100 may further comprise a computing device 1103. The computing device 1103 may be a computing device and / or system, such as the server 104 and / or the computing device 106 of the system 100. The computing device 1103 may use a model(s) stored in a Machine Learning (ML) module 1103A to generate the false-color element maps described herein. The computing device 1103 may include a display 1103B for presentation of a user interface.

[0072] For purposes of illustration, application programs and other executable program components such as the operating system 1118 are illustrated herein as discrete blocks, although it is recognized that such programs and components can reside at various times in different storage components of the computing device 1101 and / or the server 1102. An implementation of the training module 420 can be stored on or transmitted across some form of computer readable media.

[0073] Any of the disclosed methods can be performed by computer readable instructions embodied on computer readable media. Computer readable media can be any available media that can be accessed by a computer. By way of example and not meant to be limiting, computer readable media can comprise “computer storage media” and “communications media.” “Computer storage media” cancomprise volatile and non-volatile, removable and non-removable media implemented in any methods or technology for storage of information such as computer readable instructions, data structures, program modules, or other data. Exemplar^7computer storage media can comprise RAM, ROM, EEPROM, flash memory7or other memory7technology, CD-ROM, digital versatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium which can be used to store the desired information and which can be accessed by a computer.

[0074] FIG. 12 shows a flowchart for an example method 1200 for improved sample imaging and analysis. At step 1210, a computing device (e.g.. comprising and / or controlling the machine learning module 104B or 430) may receive an input image of a sample, such as the image 302 and / or the one or more portions 306. The input image may be depth-registered based on a corresponding depth (or range) of a borehole (or similar) from which the depicted sample(s) was extracted. Additionally, in some examples, the computing device may receive additional data associated with the sample, such as LiDAR-guided x-ray fluorescence (XRF) data.

[0075] At step 1220, the computing device may partition the input image into a plurality of classes based on a multi-level thresholding technique. The multi-level thresholding technique may divide the input image into distinct classes that each represent a particular element (e.g., mineral) of a plurality of elements or a particular combination of multiple elements. Each of the classes determined using the multi-level thresholding technique may comprise a particular abundance, or concentration, of an element of interest. The multi-level thresholding technique may determine (e.g., calculate) multiple thresholds for partitioning the input image into the plurality of classes such that intra-class variance is minimized and / or inter-class variance is maximized.

[0076] At step 1230. the computing device may generate a cluster-dominant image(s) based on the input image and the plurality of classes. The clusterdominant image may comprise a modified version of the input image that shows the plurality of classes (e.g., segmented classes / groups of pixels) within the input image. The computing device may also generate cluster data, which may comprise a cluster mapping indicative of and / or defining each of the classes (e.g.,clusters) of pixels within the input image. The computing device may generate the cluster-dominant image by converting pixels within each of the classes to the color corresponding to that particular class based on the cluster mapping.

[0077] At step 1240, the computing device may generate a false-color element map. For example, the computing device may convert the cluster mapping into a false-color cluster mapping. Based on the false-color cluster mapping, the computing device may generate cluster labels, which may be used to generate the false-color element map. The false-color element map may be indicative of the plurality of elements depicted in the input image.

[0078] At step 1250, the computing device may generate a histogram indicative of a relative abundance of each element of the plurality of elements within the corresponding sample. For example, the computing device may generate the histogram based on the cluster labels and the false-color element map.

[0079] While specific configurations have been described, it is not intended that the scope be limited to the particular configurations set forth, as the configurations herein are intended in all respects to be possible configurations rather than restrictive. Unless otherwise expressly stated, it is in no way intended that any method set forth herein be construed as requiring that its steps be performed in a specific order. Accordingly, where a method claim does not actually recite an order to be followed by its steps or it is not otherwise specifically stated in the claims or descriptions that the steps are to be limited to a specific order, it is in no way intended that an order be inferred, in any respect. This holds for any possible non-express basis for interpretation, including: matters of logic with respect to arrangement of steps or operational flow; plain meaning derived from grammatical organization or punctuation; the number or type of configurations described in the specification.

[0080] It will be apparent to those skilled in the art that various modifications and variations may be made without departing from the scope or spirit. Other configurations will be apparent to those skilled in the art from consideration of the specification and practice described herein. It is intended that the specification and described configurations be considered as exemplary only, with a true scope and spirit being indicated by the following claims.

Claims

CLAIMS1. A method comprising: receiving, by a computing device, an input image of a sample comprising a plurality of elements, wherein the input image is depth-registered based on a borehole from which the sample was extracted; partitioning, based on a multi-level thresholding technique, the input image into a plurality of classes, wherein each class of the plurality of classes is associated with at least one element of the plurality of elements; generating, based on the input image and the plurality of classes, a clusterdominant image; generating, based on the cluster-dominant image, a false-color element map; and generating, based on the false-color element map, a histogram, wherein the histogram is indicative of a relative abundance of each element of the plurality' of elements within the sample.

2. The method of claim 1. wherein the input image comprises a red-green-blue (RGB) two-dimensional image of the sample.

3. The method of claim 1, wherein receiving the input image further comprises receiving LiDAR-guided x-ray fluorescence (XRF) data associated with the sample.

4. The method of claim 1, wherein each class of the plurality' of classes is associated with an abundance of an element of interest.

5. The method of claim 1, wherein the cluster-dominant image comprises a modified version of the input image.

6. The method of claim 1, further comprising generating, by the computing device, a cluster mapping indicative of each class of the plurality of classes.

7. The method of claim 6, wherein the sample is a first sample, wherein the cluster mapping is indicative of a relative abundance of each element of the plurality of elements within each class of the plurality of classes . wherein theplurality of elements comprises an element of interest, and wherein the method further comprises: receiving a second input image of a second sample, wherein the second input image is associated with the borehole; partitioning, based on the cluster mapping associated with the first sample, the second input image into the plurality of classes; and generating, based on the second input image and the plurality of classes, at least one of: a second cluster-dominant image, a second false-color element map, or a second histogram associated with the second sample.

8. The method of claim 1, wherein the false-color element map is indicative of the plurality of elements.

9. The method of claim 1. wherein generating the false-color element map comprises generating, by the computing device, a cluster mapping and cluster labels.

10. An apparatus comprising: one or more processors; and memory storing processor-executable instructions that, when executed by the one or more processors, cause the apparatus to: receive an input image of a sample comprising a plurality7of elements, wherein the input image is depth-registered based on a borehole from which the sample was extracted; partition, based on a multi-level thresholding technique, the input image into a plurality7of classes, wherein each class of the plurality7of classes is associated with at least one element of the plurality' of elements; generate, based on the input image and the plurality of classes, a clusterdominant image; generate, based on the cluster-dominant image, a false-color element map; and generate, based on the false-color element map, a histogram, wherein the histogram is indicative of a relative abundance of each element of the plurality of elements within the sample.

11. The apparatus of claim 10, wherein the input image comprises a red-green- blue (RGB) two-dimensional image of the sample.

12. The apparatus of claim 10, wherein the processor-executable instructions that cause the apparatus to receive the input image further cause the apparatus to receive LiDAR-guided x-ray fluorescence (XRF) data associated with the sample.

13. The apparatus of claim 10, wherein each class of the plurality of classes is associated with an abundance of an element of interest.

14. The apparatus of claim 10, wherein the cluster-dominant image comprises a modified version of the input image.

15. The apparatus of claim 10, wherein the processor-executable instructions further cause the apparatus to generate a cluster mapping indicative of each class of the plurality of classes.

16. The apparatus of claim 15, wherein the sample is a first sample, wherein the cluster mapping is indicative of a relative abundance of each element of the plurality of elements within each class of the plurality of classes , wherein the plurality of elements comprises an element of interest, and wherein the processorexecutable instructions further cause the apparatus to: receive a second input image of a second sample, wherein the second input image is associated with the borehole; partition, based on the cluster mapping associated with the first sample, the second input image into the plurality of classes; and generate, based on the second input image and the plurality of classes, at least one of: a second cluster-dominant image, a second false-color element map. or a second histogram associated with the second sample.

17. The apparatus of claim 10, wherein the false-color element map is indicative of the plurality of elements.

18. The apparatus of claim 10, wherein the processor-executable instructions that cause the apparatus to generate the false-color element map further cause the apparatus to generate a cluster mapping and cluster labels.

19. One or more non-transitory, computer-readable media storing processorexecutable instructions that, when executed by one or more processors, cause the one or more processors to: receive an input image of a sample comprising a plurality' of elements, wherein the input image is depth-registered based on a borehole from which the sample was extracted; partition, based on a multi-level thresholding technique, the input image into a plurality of classes, wherein each class of the plurality of classes is associated with at least one element of the plurality of elements; generate, based on the input image and the plurality' of classes, a clusterdominant image; generate, based on the cluster-dominant image, a false-color element map; and generate, based on the false-color element map, a histogram, wherein the histogram is indicative of a relative abundance of each element of the plurality of elements within the sample.

20. The one or more non-transitory, computer-readable media of claim 19, wherein the input image comprises a red-green-blue (RGB) two-dimensional image of the sample.

21. The one or more non-transitory', computer-readable media of claim 19, wherein the processor-executable instructions that cause the one or more processors to receive the input image further cause the one or more processors to receive LiDAR- guided x-ray fluorescence (XRF) data associated yvith the sample.

22. The one or more non-transitory, computer-readable media of claim 19, wherein each class of the plurality of classes is associated with an abundance of an element of interest.

23. The one or more non-transitory, computer-readable media of claim 19, wherein the cluster-dominant image comprises a modified version of the input image.

24. The one or more non- transitory, computer-readable media of claim 19, wherein the processor-executable instructions further cause the one or more processors to generate a cluster mapping indicative of each class of the plurality of classes.

25. The one or more non- transitory, computer-readable media of claim 24, wherein the sample is a first sample, wherein the cluster mapping is indicative of a relative abundance of each element of the plurality of elements within each class of the plurality of classes , wherein the plurality of elements comprises an element of interest, and wherein the processor-executable instructions further cause the one or more processors to: receive a second input image of a second sample, wherein the second input image is associated with the borehole; partition, based on the cluster mapping associated with the first sample, the second input image into the plurality of classes; and generate, based on the second input image and the plurality of classes, at least one of: a second cluster-dominant image, a second false-color element map, or a second histogram associated with the second sample.

26. The one or more non-transitory, computer-readable media of claim 19, wherein the false-color element map is indicative of the plurality of elements.

27. The one or more non-transitory, computer-readable media of claim 19, wherein the processor-executable instructions that cause the one or more processors to generate the false-color element map further cause the one or more processors to generate a cluster mapping and cluster labels.

Images