Atomistic-based sensor fusion inference

Abstract

The present disclosure provides a system for atomistic-based sensor fusion inference. In operation, based on sensor data (such as hyperspectral data), the systems and methods disclosed herein may be used for more advanced and accurate methods for geological sample analysis and prediction of atomic properties. For example, as disclosed, the system may provide an atomistic-based sensor fusion inference. A geological sample may be received, and a neural network prediction may be applied comprising at least one of an electronic structure or molecular dynamics. The prediction may be used to create an atomistic-based inference of a distribution of underlying geophysical resources located proximate to the sample based on the machine-learned data set of the neural network and the sensor data. This system provides a comprehensive approach to geophysical exploration and resource extraction.

Description

ATOMISTIC-BASED SENSOR FUSION INFERENCECROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This Patent Application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 546,904, entitled “ATOMISTIC-BASED SENSOR FUSION INFERENCE,” filed 11 / 1 / 2023, and U.S. Provisional Patent Application No. 63 / 696,260, entitled “ATOMISTIC-BASED SENSOR FUSION INFERENCE,” filed 9 / 18 / 2024, both of which are assigned to the assignee hereof; the disclosures of all prior Applications are considered part of and are incorporated by reference in this Patent Application.FIELD OF THE INVENTION

[0002] The present invention relates to geophysical exploration technologies, and more particularly to an atomistic -based sensor fusion inference system for analyzing geological samples and predicting their atomic properties.BACKGROUND

[0003] Geophysical exploration, particularly in the context of mineral and resource extraction, has long been a challenging and complex field. Traditional methods of geological sample analysis often involve labor-intensive processes, such as physical collection of samples, laboratory testing, and manual interpretation of results. These methods can be time-consuming, costly, and subject to human error. Furthermore, they often provide limited information about the atomic and molecular properties of the geological samples, which are of paramount interest in resource exploration and extraction.

[0004] In recent years, advancements in sensor technology and artificial intelligence have opened up new possibilities for geophysical exploration. Sensors can now collect a vast array of data from geological samples, including physical and chemical properties. However, the sheer volume and complexity of this data present new challenges. Extracting meaningful insights from the data requires sophisticated analysis techniques, capable of handling the high-dimensional, multi-modal nature of the data.

[0005] Moreover, the prediction of atomic properties and molecular dynamics of geological samples based on sensor data is a complex task. Traditional methods often rely on simplistic models and assumptions, which may not accurately reflect the complex realities of geological structures. This can lead to inaccurate predictions and suboptimal decision-making in resource exploration and extraction. Therefore, there is a pressing demand for more advanced and accurate methods for geological sample analysis and prediction of atomic properties.

[0006] As such, there is thus a need for addressing these and / or other issues associated with the prior art.SUMMARY

[0007] In some aspects, the techniques described herein relate to a system for atomistic-based sensor fusion inference, including: a remote sensor platform configured to obtain sensor data; a processor configured to retrieve a machine-learned data set from a geophysics library; and the remote sensor platform configured to create an atomistic-based inference of a distribution of underlying geophysical resources located proximate to a geological sample based on the machine-learned data set and the sensor data.

[0008] In some aspects, the techniques described herein relate to a system for atomistic-based sensor fusion inference, including: a sensor configured to receive a geological sample and output an analysis of the geological sample; a neural network configured to predict at least one of an electronic structure, chemical complexity, molecular dynamics, or distribution of minerals of the geological sample based on the analysis; a processor configured to: derive at least one of physical properties or chemical properties of the geological sample, create a machine-learned data set based on the geological sample, and save the machine-learned data set to a geophysics library.

[0009] In some aspects, the techniques described herein relate to a system for atomistic-based sensor fusion inference, including: a sensor configured to receive a geological sample; a neural network configured to: receive, at at least one computing device, one or more datasets corresponding to possible materials based on the geological sample, create, using at least two machine learning models, a new dataset for the possible materials, wherein the at least two machine learning models are trained based on the one or more datasets, to model properties of the possible materials, and output, using the at least one computing device, a prediction including the possible materials; a remote sensor platform configured to create an atomistic-based inference of a distribution of underlying geophysical resources located proximate to the geological sample based on the prediction.

[0010] In some aspects, the techniques described herein relate to a system, further including a spectral reflectance graph used by the neural network in predicting the electronic structure or molecular dynamics of the geological sample.

[0011] In some aspects, the techniques described herein relate to a system, wherein the atomistic-based inference is created using a weighted-average of the machine-learned data set and the sensor data.

[0012] In some aspects, the techniques described herein relate to a system, wherein the atomistic-based inference is used to guide further sample collection and analysis.

[0013] In some aspects, the techniques described herein relate to a system, wherein the atomistic-based inference is used to predict the distribution of underlying minerals for a given surface area represented by the geological sample.

[0014] In some aspects, the techniques described herein relate to a system, wherein the atomistic-based inference provides insights to extract meaningful information from a set of sensors combined in a sensor-fusion manner.

[0015] In some aspects, the techniques described herein relate to a system, wherein the geological sample is a soil sample obtained from a prospective mining site.

[0016] In some aspects, the techniques described herein relate to a system, wherein the geological sample is obtained from a drone mediated magnetometry study.

[0017] In some aspects, the techniques described herein relate to a system, wherein the geological sample is obtained from a geophysical survey technique.

[0018] In some aspects, the techniques described herein relate to a system, wherein the geological sample is obtained from a hyperspectral sensing conducted remotely.

[0019] In some aspects, the techniques described herein relate to a system, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in a natural gas volume.

[0020] In some aspects, the techniques described herein relate to a system, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in a soil sample.

[0021] In some aspects, the techniques described herein relate to a system, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in an oil deposit.

[0022] In some aspects, the techniques described herein relate to a system, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in the geological sample.

[0023] In some aspects, the techniques described herein relate to a system, wherein the machine-learned data set is used to create geophysical priors for reverse engineering hyperspectral signatures.

[0024] In some aspects, the techniques described herein relate to a system, wherein the neural network applies at least one of density functional theory or molecular dynamics to create the prediction.

[0025] In some aspects, the techniques described herein relate to a system, wherein the neural network is further configured to apply density functional theory in predicting the electronic structure or molecular dynamics of the geological sample.

[0026] In some aspects, the techniques described herein relate to a system, wherein the neural network is further configured to simulate the geological sample at the atomic scale over time and across different temperatures.

[0027] In some aspects, the techniques described herein relate to a system, wherein the neural network is further configured to simulate the geological sample under various environmental conditions.

[0028] In some aspects, the techniques described herein relate to a system, wherein the processor is further configured to derive the physical properties or chemical properties of the geological sample based on the analysis from the sensor.

[0029] In some aspects, the techniques described herein relate to a system, wherein the remote sensor platform includes two or more remote sensors.

[0030] In some aspects, the techniques described herein relate to a system, wherein the remote sensor platform is configured to obtain sensor data based on a weighted-average of the two or more remote sensors.

[0031] In some aspects, the techniques described herein relate to a system, wherein the remote sensor platform is configured to obtain sensor data based on hyperspectral geophysical data derived from two or more sensors detecting physical material.

[0032] In some aspects, the techniques described herein relate to a system, wherein the remote sensor platform is configured to obtain sensor data based on metrological data derived from two or more sensors detecting physical material.

[0033] In some aspects, the techniques described herein relate to a system, wherein the sample is based on hyperspectral geophysical data derived from two or more sensors detecting physical material.

[0034] In some aspects, the techniques described herein relate to a system, wherein the sample is based on metrological data derived from two or more sensors detecting physical material.

[0035] In some aspects, the techniques described herein relate to a system, wherein the sensor data is based on a weighted- average of the two or more remote sensor platforms.

[0036] In some aspects, the techniques described herein relate to a system, wherein the sensor is configured to receive the geological sample from a soil sample.

[0037] In some aspects, the techniques described herein relate to a system, wherein the sensor is further configured to perform x-ray crystallography on the soil sample.BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 depicts a system for providing an atomistic -based sensor fusion inference, according to an embodiment.

[0039] Figure 2 illustrates a flowchart process for atomistic-based sensor fusion inference, according to aspects of the present disclosure.

[0040] Figure 3 shows a block diagram of a computer system architecture, in accordance with example embodiments.

[0041] Figure 4 presents a flowchart process for analyzing satellite data, according to an embodiment.

[0042] Figure 5 depicts a block diagram of a calibration preprocessing system, according to aspects of the present disclosure.

[0043] Figure 6 illustrates a flowchart system for processing hyperspectral imaging data, in accordance with example embodiments.

[0044] Figure 7 shows a block diagram of an atomistic-based sensor fusion inference system, according to an embodiment.

[0045] Figure 8 presents a topographic map illustrating various elevation levels, according to aspects of the present disclosure.

[0046] Figure 9 depicts a topographic map with contour lines representing elevation changes in a geological landscape, in accordance with example embodiments.DETAILED DESCRIPTION

[0047] The following description sets forth exemplary aspects of the present disclosure. It should be recognized, however, that such description is not intended as a limitation on the scope of the present disclosure. Rather, the description also encompasses combinations and modifications to those exemplary aspects described herein.

[0048] A number of implementations have been described. Nevertheless, it will be understood that various modifications may be made without departing from the spirit and scope of the disclosure. Accordingly, other implementations are within the scope of the following claims.

[0049] Hyperspectral data, collected by satellite systems, offers a powerful tool for detailed and precise Earth observation. Unlike traditional imaging systems that capture data in a few broad spectral bands, hyperspectral sensors acquire data across hundreds of narrow, contiguous spectral bands, providing a unique spectral fingerprint for each pixel in an image. This rich spectral information enables the identification and analysis of a wide range of materials and phenomena, from distinguishing between different types of vegetation and monitoring crop health to detecting mineral compositions and assessing water quality. By leveraging hyperspectral data, satellite systems can deliver enhanced environmental monitoring, natural resource management, and disaster response capabilities.

[0050] The integration of artificial intelligence (Al) with hyperspectral data promises to significantly enhance the capabilities and applications of satellite-based Earth observation. Al algorithms, particularly those involving machine learning and deep learning, can efficiently process and analyze the vast amounts of data generated by hyperspectral sensors, uncovering patterns and insights that would be difficult or impossible to detect manually. These advanced techniques can improve the accuracy and speed of data interpretation, enabling more precise identification of materials, better prediction of environmental changes, and more effective monitoring of agricultural, geological, and ecological systems.

[0051] The present disclosure provides a framework for receiving data (including imaging data, hyperspectral data, etc.), analyzing such data using a machine learningsystem, and providing an inference of distribution of geophysical resources. Of particular significance, the inference may include an atomistic -based inference.

[0052] The present disclosure provides a system for atomistic-based sensor fusion inference, which is particularly applicable in the field of geophysical exploration and resource extraction. This system leverages sensor data and machine learning techniques to predict atomic properties and molecular dynamics of geological samples. The system includes a sensor configured to receive a geological sample and output an analysis of the sample. A neural network may be utilized to predict the electronic structure or molecular dynamics of the geological sample based on the analysis. Artificial intelligence models may work in conjunction with the neural network for training and prediction purposes.

[0053] It is to be understood that within the context of the present disclosure, the use of the term “geological samples” refers to a geological sample from an area of interest, a region of interest, and / or a target mineralization.

[0054] The system further includes a processor configured to derive physical or chemical properties of the geological sample, create a machine-learned data set based on the sample, and save the data set to a geophysics library. The system also employs remote sensor platforms to obtain sensor data and create an atomistic-based inference of the distribution of underlying geophysical resources proximate to the geological sample. This inference is based on the machine-learned data set and the sensor data. The disclosed system thus provides a technologically advanced solution for predicting and analyzing the distribution of geophysical resources, thereby enhancing the efficiency and accuracy of resource extraction processes.

[0055] The present disclosure relates to a system for atomistic -based sensor fusion inference, which is particularly applicable in the field of geophysical exploration and resource extraction. This system leverages sensor data (including hyperspectral data) and machine learning techniques to predict atomic properties and molecular dynamics of geological samples. The system includes a sensor configured to receive a geological sample and output an analysis of the sample. A neural network is utilized to predict the electronic structure or molecular dynamics of the geological sample based on the analysis. The system further includes a processor configured to derive physical or chemical properties of thegeological sample, create a machine-learned data set based on the sample, and save the data set to a geophysics library.

[0056] The system also employs remote sensor platforms to obtain sensor data and create an atomistic-based inference of the distribution of underlying geophysical resources proximate to the geological sample. This inference is based on the machine-learned data set and the sensor data. The disclosed system thus provides a technologically advanced solution for predicting and analyzing the distribution of geophysical resources, thereby enhancing the efficiency and accuracy of resource extraction processes.Definitions and Use of Figures

[0057] Some of the terms used in this description are defined below for easy reference. The presented terms and their respective definitions are not rigidly restricted to these definitions — a term may be further defined by the term’s use within this disclosure. The term “exemplary” is used herein to mean serving as an example, instance, or illustration. Any aspect or design described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects or designs. Rather, use of the word exemplary is intended to present concepts in a concrete fashion. As used in this application and the appended claims, the term “or” is intended to mean an inclusive “or” rather than an exclusive “or”. That is, unless specified otherwise, or is clear from the context, “X employs A or B” is intended to mean any of the natural inclusive permutations. That is, if X employs A, X employs B, or X employs both A and B, then “X employs A or B” is satisfied under any of the foregoing instances. As used herein, at least one of A or B means at least one of A, or at least one of B, or at least one of both A and B. In other words, this phrase is disjunctive. The articles “a” and “an” as used in this application and the appended claims should generally be construed to mean “one or more” unless specified otherwise or is clear from the context to be directed to a singular form.

[0058] Various embodiments are described herein with reference to the figures. It should be noted that the figures are not necessarily drawn to scale, and that elements of similar structures or functions are sometimes represented by like reference characters throughout the figures. It should also be noted that the figures arc only intended to facilitate the description of the disclosed embodiments — they are not representative of an exhaustivetreatment of all possible embodiments, and they are not intended to impute any limitation as to the scope of the claims. In addition, an illustrated embodiment need not portray all aspects or advantages of usage in any particular environment.

[0059] An aspect or an advantage described in conjunction with a particular embodiment is not necessarily limited to that embodiment and can be practiced in any other embodiments even if not so illustrated. References throughout this specification to “some embodiments” or “other embodiments” refer to a particular feature, structure, material or characteristic described in connection with the embodiments as being included in at least one embodiment. Thus, the appearance of the phrases “in some embodiments” or “in other embodiments” in various places throughout this specification are not necessarily referring to the same embodiment or embodiments. The disclosed embodiments are not intended to be limiting of the claims.Descriptions of Exemplary Embodiments

[0060] Figure 1 depicts a system 100 for providing an atomistic -based sensor fusion inference, according to an embodiment. As shown, the system 100 depicts a flowchart system, beginning with satellite data 102, where data is gathered from multiple satellites.

[0061] It is to be appreciated that the satellite data 102 may span a diverse range of data, essential for numerous applications across various fields. For example, optical imaging data, including visible light, may provide clear and intuitive visuals of Earth's surface. Multispectral imaging expands on this by collecting data in a few broad spectral bands, often encompassing visible and near- infrared regions. Hyperspectral imaging goes even further, acquiring data across hundreds of narrow spectral bands to offer detailed spectral information for each pixel, enabling precise identification of materials and monitoring of environmental conditions. Additionally, thermal infrared data measures emitted thermal radiation to determine surface temperatures, which is crucial for tracking volcanic activity, detecting urban heat islands, and assessing land surface temperature variations. Radar data, particularly Synthetic Aperture Radar (SAR), provides high- resolution images regardless of weather conditions or time of day, useful for mapping topography, monitoring deforestation, and detecting changes in surface structures. In viewof such diverse types of data, it is to be appreciated that any, all, or combinations thereof may be collected by the satellites via the satellite data 102.

[0062] In one particular embodiment, the satellite data 102 may include hyperspectral data 104. It is to be appreciated that although the context of the system 100 includes hyperspectral data 104, other types of data may be used as input to the further analysis (such as by machine learning systems). Thus, the disclosure herein should not be limited, in any manner, to only the analysis of hyperspectral data.

[0063] The hyperspectral data 104 may undergo machine learning analysis 106, where advanced algorithms may be used to analyze the hyperspectral data 104 to extract meaningful insights. It is to be appreciated that a variety of Al models may be used by the machine learning analysis 106 to interpret and analyze the hyperspectral data 104.

[0064] For example, analyzing hyperspectral data may involve utilizing various Al models to handle its high dimensionality and complexity effectively, including Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM) networks, autoencoders, Support Vector Machines (SVMs), Random Forests, K-Nearest Neighbors (KNN), Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), Generative Adversarial Networks (GANs), and / or Graph Neural Networks (GNNs). In some instances, an Al model may be trained using historical satellite data, which in turn, may be used to improve the inference provided.

[0065] The outcome of this analysis may be used to create an atomistic-based inference of the distribution of geophysical resources 108. Such an inference may provide a detailed understanding of the underlying geophysical resources, which is valuable for applications such as resource exploration and environmental monitoring.

[0066] Within the context of the present description, an atomistic-based inference may include making predictions and drawing conclusions about the macroscopic properties and behaviors of materials or biological systems based on detailed atomic-level information. Such an approach may rely on computational methods (such as Molecular Dynamics (MD) and Quantum Mechanics (QM) simulations).

[0067] Molecular Dynamics may encompass classical mechanics to simulate the physical movements (i.e. dynamic behavior) of atoms and molecules over time. Quantum Mechanics seeks to understand the quantum behavior of particles (such as via theSchrodinger equation). Such models may he used to describe the potential energy of a system based on atomic positions and potential energy surfaces (which may map the potential energy with respect to atomic configurations). As such, these methods and models may be used to predict atomic-scale properties such as structural stability, reaction mechanisms, and material properties.

[0068] In various embodiments, the system may include a remote sensor platform configured to obtain sensor data (such as via the satellite(s)). The remote sensor platform may comprise two or more remote sensors, each capable of collecting different types of data. For example, one sensor may be designed to collect hyperspectral geophysical data, while another sensor may be designed to collect metrological data. Hyperspectral geophysical data may include data derived from two or more sensors detecting physical material, such as the presence and distribution of specific minerals or other geophysical resources. Metrological data, on the other hand, may include data derived from two or more sensors detecting physical material, such as temperature, pressure, humidity, or other environmental conditions.

[0069] In various embodiments, the remote sensor platform may be configured to obtain sensor data based on a weighted- average of the two or more remote sensors. Such a weighted-average approach may allow the system to balance the contributions of each sensor, for example, taking into account factors such as the reliability, accuracy, or relevance of the data collected by each sensor. For example, if one sensor is known to provide more accurate data for a particular type of geological sample, the system may assign a higher weight to the data collected by that sensor.

[0070] In various embodiments, the system may include a processor configured to create an atomistic-based inference of a distribution of underlying geophysical resources located proximate to the geological sample. This inference may be based on the machine- learned data set and / or the sensor data. The atomistic -based inference may provide a detailed and accurate prediction of the distribution of underlying geophysical resources, enhancing the efficiency and accuracy of resource extraction processes. The atomisticbased inference may be used to guide further sample collection and analysis, predict the distribution of underlying minerals for a given surface area represented by the geologicalsample, and / or extract meaningful information from a set of sensors combined in a sensorfusion manner.

[0071] Further, it is to be appreciated that the hyperspectral data 104 may relate to a variety of fields and applicability. For example, in agriculture, hyperspectral imaging can be used to monitor crop health, identify diseases, and optimize irrigation by analyzing the reflectance spectra of plants. In environmental monitoring, it can help detect pollutants, map land cover changes, and assess ecosystem health. In medicine, it has applications in tissue analysis, cancer detection, and drug development. In geology, it has applications in mining, mineral economics, metallurgy, extraction of minerals, and environmental engineering. In health, it has applications to medical diagnostics, cancer detection, non- invasive monitoring, pathology, and telemedicine / remote diagnostics. It is to be appreciated that the examples provided herein are not to be construed as limiting in any way, and are intended merely as possible applications of the disclosure herein.

[0072] As such, data from sensing devices (such as hyperspectral, but not limited solely thereto) may be used in combination with a machine learning analysis 106 for purposes of resulting in atomic -based inferences. Such inferences, within one particular context, may relate to distribution of geophysical resources (such as the atomistic-based inference of the distribution of geophysical resources 108). However, as has been discussed, atomic-based inferences may expand beyond merely geophysical resources. Therefore, sensor data may be used by a machine learning analysis to create atomic-based inferences.

[0073] In other embodiments, hyperspectral imaging may be used to draw atomistic conclusion based on the ability to cross-reference the spectral response function (how light reflects off a molecule) from the sensor with atomistic simulation in to determine specific materials. In general, atomic -based inferences may be applied to any property or characteristic that can be both attained from sensor data and deduced from atomic conditions. For example, atomic-based inferences may be applied with respect to magnetic properties, both from an aerial survey and from hand-held ground measurements. These magnetics properties can be calculated through atomistic simulation through investigating exchange interactions and spin-orbit coupling. Electronic properties (such as conductivity) may also be applied to the atomic-based inferences. A variety of sensor methods (such as but not limited to induced polarization or magnetotellurics) may be used to provideatomistic inferences that are associated with atomic outputs (such as band structure or density of states, etc.). It is to be appreciated that although magnetic properties and electronic properties are detailed herein as possible examples associated with atomic-based inference data, other properties may in a similar manner be extracted and used.

[0074] As one example, an image may be captured by a first camera sensor of a human body scan. The image may include hyperspectral data. The image data may be analyzed by an artificial intelligence system to infer atomic-level inferences, including hot spots of potential cancer growth and / or abnormalities.

[0075] As another example, an image may be captured by a first camera sensor of a satellite. The image may include hyperspectral data. The image data may be analyzed by an artificial intelligence system to infer atomic-level inferences, including the ability to precisely identify and map minerals (based on each of their unique spectral signatures).

[0076] More illustrative information will now be set forth regarding various optional architectures and uses in which the foregoing method may or may not be implemented, per the desires of the user. It should be strongly noted that the following information is set forth for illustrative purposes and should not be construed as limiting in any manner. Any of the following features may be optionally incorporated with or without the exclusion of other features described.

[0077] Figure 2 illustrates a flowchart process 200 for atomistic -based sensor fusion inference, according to aspects of the present disclosure As an option, the flowchart process 200 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequent figure(s) and / or description thereof. Of course, however, the flowchart process 200 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0078] As shown, the flowchart process 200 illustrates the process of obtaining the geological sample, the application of the neural network for prediction, and the specific configurations and roles of the remote sensors in data collection. The geological sample is obtained from a hyperspectral sensing conducted remotely.

[0079] In practice, geological samples are received and analyzed (step 202). A neural network predicts the electronic structure and molecular dynamics (step 204) of the samples after crystallographic and other chemical and physical information is provided as outputfrom analysis (e.g., the analysis of step 202). Further characteristics of the sample are derived (step 206). At least some of the derived characteristics may be used to configure two or more remote sensor platforms (e.g., UAV1 remote sensors of step 208B and UAV2 remote sensors of step 208 A).

[0080] Inferencing may be performed by accessing a machine-learned geophysics library. This machine-learned geophysics library may be constructed by predicting properties such as the spectral reflectance graph of the different minerals and materials found in the geological samples. This spectral reflectance graph may be predicted using the aforementioned neural networks which may apply density functional theory and molecular dynamics to predict physical and chemical properties that different constituents in the soil samples include. These pre-computed properties, calibrated with real world validation and characterization act as “Physics Priors,” or physical property signatures that correlate to the different materials found in the geological sample.

[0081] After sensor data from UAV1 in step 208B and UAV2 in step 208A are obtained, in step 210, a neural network (e.g. an “expert network”) may be used to combine (1) the information derived from the machine learned physics libraries and (2) the fusion of the two (or more) different remote sensors to make a weighted-average, atomistic -based inference on what the distribution of underlying minerals must be for a given surface area represented by the geological sample obtained. This weighted- average approach allows the system to balance the contributions of each sensor, taking into account factors such as the reliability, accuracy, or relevance of the data collected by each sensor. Learnings from the inferencing influence further sample collection and analysis. For example, per step 212, a sample may be drilled, and more samples may be taken for crystallographic information.

[0082] In various embodiments, the neural network may be configured to predict at least one of an electronic structure or molecular dynamics of the geological sample based on the analysis of the flowchart process 200. The neural network applies at least one of density functional theory or molecular dynamics to create the prediction. By applying this theory (or any other deemed applicable), the neural network may accurately predict the electronic structure of the geological sample, providing valuable insights into the sample's atomic and molecular properties.

[0083] In various embodiments, the geological sample received by the sensor is a soil sample obtained from a prospective mining site. In other cases, the geological sample may be obtained from a drone mediated magnetometry study or a geophysical survey technique (including but not limited to seismic, magnetic, gravity, electrical, electromagnetic, and radiometric surveys).

[0084] In various embodiments, the neural network may be further configured to simulate the geological sample at the atomic scale over time and across different temperatures. This simulation may provide detailed insights into the atomic and molecular dynamics of the geological sample, enhancing the accuracy of the predictions made by the neural network. In addition, the neural network may also be configured to simulate the geological sample under various environmental conditions. This feature allows the system to predict how the geological sample might behave under different environmental conditions, providing valuable information for resource extraction processes.

[0085] In various embodiments, the system may be further configured to derive the physical properties or chemical properties of the geological sample based on the analysis from the sensor.

[0086] In various embodiments, conventional techniques for analysis of soil samples suffer from an extreme lack of data about the soil. The present disclosure therefore overcomes the deficiencies of conventional soil analysis techniques.

[0087] Aspects of the present disclosure solve problems associated with using computer systems for lack of data from conventional sensing techniques. These problems are unique to, and may have been created by, various computer-implemented sensing techniques as applied in the context of geophysical exploration. Some embodiments are directed to approaches for application of x-ray crystallography and other characterization techniques over a soil sample to construct useful information for construction of a family of neural networks that work together to predict chemical and physical properties based on the characterization results of the soil sample. The accompanying figures and discussions herein present example environments, systems, methods, and computer program products for autonomous atomistic -based sensor fusion inference.

[0088] In the context of geophysical surveys such as conducting a drone mediated magnetometry study of a prospective mining site, or hyperspectral sensing conductedremotely either via drone or satellite, there exists a gap in the data generated by these sensors and the ability to fully extract the useful, actionable information from the data and apply it for geological resource extraction. The present disclosure includes a method of automatic geological resource extraction where a human in the loop operates a series of remote sensors that result in a collection of passive and inductive test data. The results of these passive and inductive tests are the result of a macroscopic collection of atoms of varying chemical composition, crystal structure, phase, orientation and spatial distribution. The collection of atoms in this case can include any part of a geological resource that is being prospected, for example the ionic conductivity of a soil sample can be derived and measured. Information such as this, x-ray crystallography and other characterization techniques applied to the soil sample allow us to construct useful information for a family of neural networks which work together to predict chemical and physical properties based on the characterization results of the soil sample.

[0089] In various embodiments, the methods disclosed herein may apply to any and all geophysical survey techniques. Properties such as the local density of states, charge density, dielectric function, electrical and ionic conductivity are predicted after an x-ray crystallographic representation of the geological samples being surveyed are passed to this family of neural networks. This enables the creation of geophysical priors that are useful in reverse engineering hyperspectral signatures when processing data from a hyperspectral sensing source, simulated or actualized. Another family of neural networks simulates the samples at the atomic scale over time and across different temperatures as well as under other environmental conditions. This creates a comprehensive atomic scale, large number of atoms digital replica that provides insights to extract meaningful information from a set of sensors that are combined together in a sensor-fusion manner.

[0090] Techniques for atomistic-based sensor fusion inference are disclosed. In use, a method for atomistic-based sensor fusion inference includes receiving a soil sample, outputting an analysis, based on one or more sensors, of the soil sample, and predicting, using the analysis and a neural network, at least one of an electronic structure or molecular dynamics of the soil sample. Additionally, at least one of physical properties or chemical properties of the soil sample is derived, a machine-learned data set is created based on the soil sample, and the machine-learned data is saved set to a geophysics library. Further,sensor data from two or more remote sensor platforms is obtained, and an atomistic-based inference is created, using the neural network, of a distribution of underlying minerals located proximate to the soil sample based on the machine-learned data set and the sensor data.

[0091] This summary is provided to introduce a selection of concepts that are further described elsewhere in the written description and in the figures. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to limit the scope of the claimed subject matter. Moreover, the individual embodiments of this disclosure each have several innovative aspects, no single one of which is solely responsible for any particular desirable attribute or end result.

[0092] The present disclosure describes techniques used in systems, methods, and computer program products for atomistic-based sensor fusion inference, which techniques advance the relevant technologies to address technological issues with legacy approaches. More specifically, the present disclosure describes techniques used in systems, methods, and in computer program products for autonomous atomistic-based sensor fusion inference. Certain embodiments are directed to technological solutions for application of x-ray crystallography and other characterization techniques over a soil sample to construct useful information for construction of a family of neural networks that work together to predict chemical and physical properties based on the characterization results of the soil sample.

[0093] The disclosed embodiments modify and improve beyond legacy approaches. In particular, the herein-disclosed techniques provide technical solutions that address the technical problems attendant to lack of data from conventional sensing techniques. Such technical solutions involve specific implementations (e.g., data organization, data communication paths, module-to-module interrelationships, etc.) that relate to the software arts for improving computer functionality. Various applications of the herein-disclosed improvements in computer functionality serve to reduce demand for computer memory, reduce demand for computer processing power, reduce network bandwidth usage, and reduce demand for intercomponent communication. For example, when performing computer operations that address the various technical problems underlying lack of data from conventional sensing techniques both memory usage and CPU cycles demanded aresignificantly reduced as compared to the memory usage and CPU cycles that would be needed but for practice of the herein-disclosed techniques for application of x-ray crystallography and other characterization techniques over a soil sample to construct useful information for construction of a family of neural networks that work together to predict chemical and physical properties based on the characterization results of the soil sample. Strictly as one case, the data structures as disclosed herein and their use serve to reduce both memory usage and CPU cycles as compared to alternative approaches. Moreover, information that is received during operation of the embodiments is transformed by the processes that store data into and retrieve data from the aforementioned data structures.

[0094] The ordered combination of steps of the embodiments serve in the context of practical applications that perform steps for application of x-ray crystallography and other characterization techniques over a soil sample to construct useful information for construction of a family of neural networks that work together to predict chemical and physical properties based on the characterization results of the soil sample. As such, the disclosed techniques overcome long-standing yet heretofore unsolved technological problems associated with lack of data from conventional sensing techniques that arise in the realm of computer systems.

[0095] Many of the herein-disclosed embodiments for application of x-ray crystallography and other characterization techniques over a soil sample to construct useful information for construction of a family of neural networks that work together to predict chemical and physical properties based on the characterization results of the soil sample are technological solutions pertaining to technological problems that arise in the hardware and software arts that underlie geophysical exploration. Aspects of the present disclosure achieve performance and other improvements in peripheral technical fields including, but not limited to, material property analysis and neural networks.

[0096] Some embodiments include a sequence of instructions that are stored on a non- transitory computer readable medium. Such a sequence of instructions, when stored in memory and executed by one or more processors, causes the one or more processors to perform a set of acts for application of x-ray crystallography and other characterization techniques over a soil sample to construct useful information for construction of a familyof neural networks that work together to predict chemical and physical properties based on the characterization results of the soil sample.

[0097] Some embodiments include the aforementioned sequence of instructions that are stored in a memory, which memory is interfaced to one or more processors such that the one or more processors can execute the sequence of instructions to cause the one or more processors to implement acts for application of x-ray crystallography and other characterization techniques over a soil sample to construct useful information for construction of a family of neural networks that work together to predict chemical and physical properties based on the characterization results of the soil sample.

[0098] In various embodiments, any combinations of any of the above can be organized to perform any variation of acts for autonomous atomistic-based sensor fusion inference, and many such combinations of aspects of the above elements are contemplated.

[0099] Figure 3 shows a block diagram 300 of a computer system architecture, in accordance with example embodiments. As an option, the block diagram 300 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequent figure(s) and / or description thereof. Of course, however, the block diagram 300 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0100] As shown, the block diagram 300 is of an instance of a computer system suitable for implementing embodiments of the present disclosure. The computer system includes a bus 306 or other communication mechanism for communicating information. The bus interconnects subsystems and devices such as a CPU, or a multi-core CPU (such as data processor 307), a system memory (such as main memory 308, or an area of random access memory (RAM)), a non-volatile storage device or non-volatile storage area (such as readonly memory or ROM 309), an internal storage device 310 or external storage device 313 (such as magnetic or optical), a data interface 333, a communications interface 314 (such as PHY, MAC, Ethernet interface, modem, etc.). The aforementioned components are shown within processing element partition 301, however other partitions are possible. The shown computer system further includes a display 311 (such as CRT or LCD), various input devices 312 (such as keyboard, cursor control), and an external data repository 331.All or portions of computer system maybe subsumed into another device such as a smart phone or a pad or tablet.

[0101] In some implementations, the computer system performs specific operations by data processor 307 executing one or more sequences of one or more program instructions contained in a memory. Such instructions (such as program instructions 3021, program instructions 3022, program instructions 3023, etc.) may be contained in or may be read into a storage location or memory from any computer readable / u sable medium such as a static storage device or a disk drive. The sequences may be organized to be accessed by one or more processing entities configured to execute a single process or configured to execute multiple concurrent processes to perform work. A processing entity may be hardwarebased (such as involving one or more cores) or software-based, and / or may be formed using a combination of hardware and software that implements logic, and / or may carry out computations and / or processing steps using one or more processes and / or one or more tasks and / or one or more threads or any combination thereof.

[0102] In addition, or in the alternative, computer system of the block diagram 300 performs specific networking operations using one or more instances of communications interface 314. Instances of communications interface 314 may include one or more networking ports that are configurable (such as pertaining to speed, protocol, physical layer characteristics, media access characteristics, etc.) and any particular’ instance of communications interface 314 or port thereto may be configured differently from any other particular instance. Portions of a communication protocol may be carried out in whole or in part by any instance of communications interface 314, and data (such as packets, data structures, bit fields, etc.) may be positioned in storage locations within communications interface 314, or within system memory, and such data may be accessed (such as using random access addressing, or using direct memory access DMA, etc.) by devices such as data processor 307.

[0103] Communications link 315 may be configured to transmit (such as send, receive, signal, etc.) any types of communications packets (such as communications packet 3381, communications packet 338N) including any organization of data items. The data items may include a payload data area 337, a destination address 336 (such as a destination IP address), a source address 335 (such as a source IP address), and may include variousencodings or formatting of bit fields to populate packet characteristics 334. In some cases, the packet characteristics include a version identifier, a packet or payload length, a traffic class, a flow label, etc. In some cases, payload data area 337 includes a data structure that is encoded and / or formatted to fit into byte or word boundaries of the packet.

[0104] In some embodiments, hard-wired circuitry may be used in place of or in combination with software instructions to implement aspects of the disclosure. Thus, embodiments of the disclosure are not limited to any specific combination of hardware circuitry and / or software. In embodiments, the term “logic” shall mean any combination of software or hardware that is used to implement all or part of the disclosure.

[0105] The term “computer readable medium” or “computer usable medium” as used herein refers to any medium that participates in providing instructions to data processor 307 for execution. Such a medium may take many forms including, but not limited to, nonvolatile media and volatile media. Non-volatile media includes, for example, optical or magnetic disks such as disk drives or tape drives. Volatile media includes dynamic memory such as RAM.

[0106] Common forms of computer readable media include, for example, floppy disk, flexible disk, hard disk, magnetic tape, or any other magnetic medium; CD-ROM or any other optical medium; punch cards, paper tape, or any other physical medium with patterns of holes; RAM, PROM, EPROM, FLASH-EPROM, or any other memory chip or cartridge, or any other non-transitory computer readable medium. Such data may be stored, for example, in any form of external data repository 331, which in turn may be formatted into any one or more storage areas, and which may include parameterized storage 339 accessible by a key (such as filename, table name, block address, offset address, etc.).

[0107] Execution of the sequences of instructions to practice certain embodiments of the disclosure are performed by a single instance of computer system of the block diagram300. According to certain embodiments of the disclosure, two or more instances of computer system coupled by a communications link 315 (such as LAN, PSTN, or wireless network) may perform the sequence of instructions required to practice embodiments of the disclosure using two or more instances of components of computer system.

[0108] Computer system of the block diagram 300 may transmit and receive messages such as data and / or instructions organized into a data structure (such as communicationspackets). The data structure may include program instructions (such as application code 303), communicated through communications link 315 and communications interface 314. Received program instructions may be executed by data processor 307 as it is received and / or stored in the shown storage device or in or upon any other non-volatile storage for later execution. The computer system may communicate through a data interface 333 to a database 332 on an external data repository 331. Data items in a database may be accessed using a primary key (such as a relational database primary key).

[0109] Processing element partition 301 is merely one sample partition. Other partitions may include multiple data processors, and / or multiple communications interfaces, and / or multiple storage devices, etc. within a partition. For example, a partition may bound a multi-core processor (such as possibly including embedded or co-located memory), or a partition may bound a computing cluster having plurality of computing elements, any of which computing elements are connected directly or indirectly to a communications link. A first partition may be configured to communicate to a second partition. A particular first partition and particular second partition may be congruent (such as in a processing element array) or may be different (such as including disjoint sets of components).

[0110] A module as used herein may be implemented using any mix of any portions of the system memory and any extent of hard-wired circuitry including hard-wired circuitry embodied as a data processor 307. Some embodiments include one or more special-purpose hardware components (such as power control, logic, sensors, transducers, etc.).

[0111] Various implementations of the database 332 include storage media organized to hold a series of records or files such that individual records or files are accessed using a name or key (such as a primary key or a combination of keys and / or query clauses). Such files or records may be organized into one or more data structures (such as data structures used to implement or facilitate aspects of autonomous atomistic-based sensor fusion inference). Such files or records may be brought into and / or stored in volatile or nonvolatile memory.

[0112] Figure 4 presents a flowchart process 400 for analyzing satellite data, according to an embodiment. As an option, the flowchart process 400 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequentfigure(s) and / or description thereof. Of course, however, the flowchart process 400 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0113] As shown, the flowchart process 400 illustrates the steps involved in analyzing satellite data 402. The satellite data 402 undergoes calibration preprocessing 404, which involves correcting the data (such as but not limited to sensor errors, atmospheric effects, and other distortions). This calibration preprocessing 404 step ensures that the data is accurate and reliable, providing a solid foundation for further analysis.

[0114] Following the calibration preprocessing 404, a decision is made based on the surface reflectance within range decision 406. This decision involves determining whether the surface reflectance of the geological sample, as measured by the satellite data 402, is within an acceptable predetermined range (such as between 0-1). If the surface reflectance is within the acceptable range, the data is stored in the sensor database 408, which may serve as a repository for the sensor data. If the surface reflectance is not within an acceptable range, the data is returned to calibration preprocessing 404, or otherwise discarded.

[0115] Concurrently, data is also stored in the spectral database 410. The spectral database 410 stores spectral data related to the geological sample, such as the spectral reflectance graph of the sample. This spectral data provides detailed information about the sample's spectral properties, enhancing the understanding of the sample's composition and potential uses.

[0116] Further, it is to be appreciated that the spectral data of the spectral database 410 may include measurements of the intensity of light across various wavelengths (such as that which is gathered from different materials or substances). Further, the spectral database 410 may include spectra from a wide range of sources, including natural and synthetic compounds, minerals, astronomical objects, and biological samples.

[0117] The processing of hyperspectral imaging data 412 utilizes information from both the sensor database 408 and the spectral database 410. Hyperspectral imaging involves the collection and processing of information across the electromagnetic spectrum, providing detailed spectral information about the geological sample. The processing of hyperspectral imaging data 412 involves analyzing the hyperspectral imaging data toextract meaningful insights, such as the presence and distribution of specific minerals or other geophysical resources.

[0118] The processed data is then used to identify geophysical indicators 414. These geophysical indicators may indicate the presence and distribution of specific minerals or other geophysical resources. The identification of these geophysical indicators enhances the understanding of the geological sample, which may be of particular value for resource extraction processes.

[0119] Finally, the geophysical indicators 414 are used for categorical prediction for mineral deposits 416. This prediction involves predicting the presence and distribution of mineral deposits based on the geophysical indicators 414. The categorical prediction for mineral deposits 416 provides a detailed and accurate prediction of the distribution of mineral deposits, enhancing the efficiency and accuracy of resource extraction processes.

[0120] In some instances, the data from the sensor database 408 may be passed directly for providing the categorical prediction for mineral deposits 416 (not shown). For example, the spectral database 410 may be configured to use known information regarding the expected values of the spectral response curve for a known material. The sensor data of sensor database 408 may be associated with specific minerals. For example, the sensor data may be obtained by utilizing computational methodologies to benchmark and match the measured spectral curves to the expected for a known mineral. As such, in some instances, data from the sensor database 408 may be used to me categorical predictions for mineral deposits 416 (although additional processing and / or preconfigured threshold conditions would have to be satisfied in such an instance).

[0121] Figure 5 depicts a block diagram 500 of a calibration preprocessing system, according to aspects of the present disclosure. As an option, the block diagram 500 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequent figure(s) and / or description thereof. Of course, however, the block diagram 500 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0122] As shown, the calibration preprocessing system 500 is a subset of the calibration preprocessing 404 of Figure 4. In various embodiments, the calibration preprocessingsystem may include more or less than components provided within the context of the calibration preprocessing system 500.

[0123] The radiometric correction 502 component may be responsible for correcting sensor errors that may occur during the data collection process. These errors may be due to factors such as sensor noise, sensor drift, or other sensor-related issues. By correcting these errors, the radiometric correction 502 ensures that the sensor data is accurate and reliable.

[0124] The atmospheric correction 504 component may be responsible for correcting distortions in the sensor data caused by the atmosphere. When sensor data is collected remotely, such as from a satellite, the data may be distorted by atmospheric effects such as scattering, absorption, and refraction. The atmospheric correction 504 corrects these distortions, ensuring that the sensor data accurately represents the geological sample.

[0125] The signal-to-noise ratio improvement 506 component may be responsible for improving the signal-to-noise ratio of the sensor data. By improving the signal-to-noise ratio, the signal-to-noise ratio improvement 506 enhances the quality of the sensor data, making it easier to extract meaningful information from the data.

[0126] The cloud detection 508 component may be responsible for detecting clouds in the sensor data. Clouds can obscure the geological sample, making it difficult to accurately analyze the sample. By detecting clouds, the cloud detection 508 can flag areas of the sensor data that may be unreliable due to cloud cover.

[0127] In some cases, the calibration preprocessing system 500 may include additional components not shown in FIG. 5. For example, the system may include a component for correcting geometric distortions in the sensor data, a component for correcting sensor drift, a component for correcting other types of errors or distortions in the sensor data, etc. The specific components included in the calibration preprocessing system 500 may vary depending on the specific requirements of the atomistic-based sensor fusion inference system.

[0128] Figure 6 illustrates a flowchart system 600 for processing hyperspectral imaging data, in accordance with example embodiments. As an option, the flowchart system 600 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequent figure(s) and / or description thereof. Of course,however, the flowchart system 600 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0129] As shown, the flowchart system 600 includes a variety of subset items of Figure 4, including the sensor database 408 and the spectral database 410, and their interaction with the processing of hyperspectral imaging data 412, and of the interaction between the processing of hyperspectral imaging data 412 and the geophysical indicators 414, resulting in the categorical prediction for mineral deposits 416.

[0130] As such, the sensor database 408 and the spectral database 410 may provide data for the processing of hyperspectral imaging data 412. The sensor database 408 may store data collected from various sensors, such as hyperspectral sensors, magnetometers, or other geophysical sensors. The spectral database 410, on the other hand, may store spectral data related to the geological sample, such as the spectral reflectance graph of the sample.

[0131] The processing of hyperspectral imaging data 412 involves several steps, each of which contributes to the extraction of meaningful insights from the hyperspectral imaging data. The first step is the minimum noise function transformation 602, which is used to reduce the noise in the hyperspectral data. This transformation separates the data into components that contain the majority of the signal (useful information) and those that contain mostly noise. The minimum noise function transformation may assist with clarifying the true spectral signals.

[0132] The next step is the automatic target generation process (ATGP) 604, which is a method used to identify potential endmembers from the hyperspectral data automatically. Such a process may select pixels that are likely to represent pure materials based on their spectral uniqueness and distinctiveness. In one embodiment, ATGP 604 may be used to provide an initial guess or determination. ATGP may use a deterministic approach to select pixels that are spectrally distinct and likely to be pure. It may involve a simple computational process where pixels are chosen based on their geometric relationships in spectral space, making it relatively fast and efficient.

[0133] Following the ATGP 604, the N-FINDR process 606 may be applied. This algorithm refines the initial guesses provided by ATGP 604. It may iteratively search forthe set of endmembers that maximize the volume enclosed by them in spectral space, identifying the true spectral signatures of the materials present in the image. As such, N- FINDR may improve the accuracy of the material quantification.

[0134] Based on the N-FINDR 606 results, the fully constrained least squares (FCLS) 608 may then applied to provide a per-pixel quantification of each end member's presence by solving an inverse problem that fits the observed spectrum to the end member signatures under non-negativity and full sum-to-one constraints. In one embodiment, the FCLS 608 may be used to determine amount (abundance) of a mineral.

[0135] Additionally, based on the N-FINDR 606 results, the spectral angle mapper (SAM) 610 may be applied. In one embodiment, SAM may be used to compare the angle between the spectra of the pixels and the identified endmembers (e.g. smaller angles may indicate a closer match to the reference spectrum). SAM 610 may be used to classify materials in each pixel based on their spectral similarity to the extracted end members.

[0136] The outputs from the FCLS 608 and SAM 610 contribute to the creation of the abundance map 614 and the angle map 612, respectively. The abundance map 614 serves a role in quantifying the proportion of various materials within each pixel of a hyperspectral image. The angle map 612 focuses on the similarity between the spectral signature of each pixel and predetermined reference spectra.

[0137] In various embodiments, the abundance map 614 may be used to quantify the proportion of various materials within each pixel of a hyperspectral image. By enforcing non-negativity and sum-to-one constraints, the FCLS 608 algorithm ensures that the computed proportions are realistic. This abundance map 614 may be useful in applications such as resource management, enabling an understanding of the distribution of minerals or vegetation, and in environmental monitoring, where detecting changes in land use or contamination levels over time is vital. The abundance map 614 provides detailed, quantitative insights into the volume of the detected end member.

[0138] In various embodiments, the angle map 612 may focus on the similarity between the spectral signature of each pixel and predetermined reference spectra. The angle map 612 uses the spectral angle — a metric that quantifies the cosine of the angle between two spectra — to determine similarity. Pixels with smaller angles are considered more similar to the reference, making this angle map 612 extremely valuable for identifyingspecific materials or objects within the scene based on their spectral characteristics. Additionally, the angle map 612 may be applicable for mineralogy, where pinpointing specific mineral deposits based on their spectral signatures is needed or crop monitoring in agriculture, where identifying plant types and their health is crucial.

[0139] Further, in one embodiment, an output of the sensor database 408 may be provided to a SpectralFormer module which may be used to effectively handle the sequential nature of spectral data. As such, the SpectralFormer may allow it to model both global and local spectral information accurately. One purpose of the SpectralFormer may be to overcome limitations in existing architectures which may not capture long-term dependencies in spectral data or struggle with the sequential attributes of such data. By incorporating transformers, the SpectralFormer can capture intricate spectral details necessary for precise material identification and classification both spectrally and spatially. An output of the SpectralFormer may be therefore used directly for the categorical prediction for mineral deposits 416.

[0140] Based on such analysis, a categorical prediction for mineral deposits 416 may be more accurately provided.

[0141] Figure 7 shows a block diagram 700 of an atomistic-based sensor fusion inference system, according to an embodiment. As an option, the block diagram 700 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequent figure(s) and / or description thereof. Of course, however, the block diagram 700 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0142] As shown, the block diagram 700 includes a focus on details associated with the spectral database 410 of Figure 4.

[0143] The system 700 includes a spectral library 702, an XRD data source 704, and DFT generated spectral response functions 706. These components are interconnected and provide data to the spectral database 410.

[0144] In various embodiments, the spectral library 702 may include data, including spectral data, from lab measured spectral response functions from any mineral (corresponding with the obtained geological sample).. This spectral data may include spectral reflectance graphs of minerals and materials found in different types of geologicalsamples, such as natural gas volumes, soil samples, oil deposits, and other geological samples. The spectral reflectance graph may provide detailed information about the spectral properties of the minerals and materials in the geological sample.

[0145] The XRD data source 704 may provide X-ray diffraction (XRD) data related to the geological sample, which may be used to determine the atomic and molecular structure of a crystal. Such XRD data may provide detailed information about the atomic and molecular structure of the geological sample.

[0146] The DFT generated spectral response functions 706 may be used to predict the spectral response of the geological sample based on its atomic and molecular structure. It is to be appreciated that the density functional theory (DFT) generated spectral response functions may be used to investigate and / or model the electronic structure of atoms, molecules, condensed phases, etc.

[0147] The spectral database 410 may store the machine-learned data set, which includes the spectral reflectance graph of the minerals and materials found in the geological sample. This machine-learned data set is used to create geophysical priors for reverse engineering hyperspectral signatures. These geophysical priors may provide a basis for predicting the distribution of underlying geophysical resources based on the geological sample and sensor data.

[0148] As shown, each of the spectral library 702, the XRD 704, and the DFT generated spectral response functions 706 may be provided as inputs to the spectral database 410. Additionally, the XRD 704 may also be provided as an input to the DFT generated spectral response functions 706 such that the DFT models may be more accurately applied to the relevant hyperspectral data.

[0149] In some cases, the machine-learned data set may include a spectral reflectance graph of minerals and materials found in a natural gas volume. In other cases, the machine- learned data set may include a spectral reflectance graph of minerals and materials found in a soil sample. In yet other cases, the machine-learned data set may include a spectral reflectance graph of minerals and materials found in an oil deposit. In still other cases, the machine-learned data set may include a spectral reflectance graph of minerals and materials found in the geological sample. As such, these variations of the machine-learned data set may provide a comprehensive view of the spectral properties of the geological sample.

[0150] Figure 8 presents a topographic map 800 illustrating various elevation levels, according to aspects of the present disclosure. As an option, the topographic map 800 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequent figure(s) and / or description thereof. Of course, however, the topographic map 800 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0151] As shown, a topographic map 802 is depicted, illustrating various elevation levels and the use of contour lines and shaded areas to indicate changes in terrain and regions of interest.

[0152] Within the topographic map 802 is provided an area of interest 804. For example, the area of interest 804 may be used in conjunction with the atomistic -based sensor fusion inference system 100 to guide the collection of geological samples. For example, the system 100 may be configured to indicate the area of interest 804.

[0153] It is to be appreciated that the topographic map 802 may represent a user interface display corresponding to the categorical prediction for mineral deposits 416. For example, after using the flowchart process 400, the topographic map 802 may display results corresponding with the predictions provided.

[0154] In other embodiments, the system 100 may be configured to indicate areas of high probability of a predetermined mineral for collection. As such, the topographic map 802 may thus provide a valuable tool for planning and guiding the collection of geological samples.

[0155] Figure 9 depicts a topographic map 900 with contour lines representing elevation changes in a geological landscape, in accordance with example embodiments. As an option, the topographic map 900 may be implemented in the context of any one or more of the embodiments set forth in any previous and / or subsequent figure(s) and / or description thereof. Of course, however, the topographic map 900 may be implemented in the context of any desired environment. Further, the aforementioned definitions may equally apply to the description below.

[0156] As shown, a topographic map 902 is depicted, illustrating various elevation levels and the use of contour lines and shaded areas to indicate changes in terrain and regions of interest.

[0157] Within the topographic map 902 is provided an area of interest 904. For example, the area of interest 904 may be used in conjunction with the atomistic -based sensor fusion inference system 100 to guide the collection of geological samples.

[0158] In a manner similar to Figure 8, the topographic map 900 may be used as a user interface display corresponding to the categorical prediction for mineral deposits 416. For example, after using the flowchart process 400, the topographic map 902 may display results corresponding with the predictions provided.

[0159] In various embodiments, the topographic map 902 may provide detailed visualization of the terrain's structure, which can inform decisions about where to collect geological samples or where to focus resource extraction efforts. Additionally, the system 100 may be configured to collect samples from regions of the topographic map 902 with specific elevation levels or terrain characteristics. The system 100 may also use the topographic map 902 to predict the distribution of underlying geophysical resources for a given surface area represented by the geological sample.

[0160] In other cases, the topographic map may be used to guide further sample collection and analysis. For instance, if the atomistic-based inference predicts a high concentration of a particular’ mineral in a specific area of the map, additional samples may be collected from that area for further analysis.

[0161] In comparing the topographic map 802 with the topographic map 902, the topographic map 802 shows a top-down perspective, whereas the topographic map 902 shows a side perspective. In this manner, the topographic map 802 may be used to indicate a potential area of interest (such as longitude and latitude), and the topographic map 902 may provide further details (such as point of depth).

[0162] It is to be appreciated that the topographic map 802 and the topographic map 902 may be combined into a single user interface, including a 3D map construction that can be viewed from multiple angles.

[0163] Further, in other embodiments, the flowchart process 400 may be used to not only provide predictions of mineral deposits, but may also be used, in combination with an artificial intelligence system, to train a neural network to also predict potential areas of mineral deposits. In this manner, the system may be used either based on the satellite data402 and / or trained data (via an artificial intelligence system) to provide potential predictions.

[0164] In various embodiments, the atomistic-based sensor fusion inference system 100 includes a sensor configured to receive a geological sample. The geological sample may be obtained through various methods, each providing a different type of data about the sample. For instance, the geological sample may be a soil sample obtained from a prospective mining site. In this case, the sensor may be configured to receive the soil sample and output an analysis of the sample. The analysis may include data about the sample's physical and chemical properties, such as its mineral composition, density, moisture content, and other relevant properties.

[0165] In other cases, the geological sample may be obtained from a drone mediated magnetometry study. In a drone mediated magnetometry study, a drone equipped with a magnetometer may be used to collect magnetic data over a large area, and the sensor in the atomistic-based sensor fusion inference system 100 may be configured to receive this magnetic data and output an analysis of the geological sample based on the data. The analysis may include data about the sample's magnetic properties, such as its magnetic susceptibility, remanent magnetization, and other relevant properties.

[0166] In yet other cases, the geological sample may be obtained from a geophysical survey technique, and the sensor in the atomistic-based sensor fusion inference system 100 may be configured to receive data from a geophysical survey and output an analysis of the geological sample based on the data. The analysis may include data about the sample's physical properties, such as its resistivity, conductivity, gravity, and other relevant properties.

[0167] In still other cases, the geological sample may be obtained from a hyperspectral sensing conducted remotely, and the sensor in the atomistic-based sensor fusion inference system 100 may be configured to receive hyperspectral data and output an analysis of the geological sample based on the data. The analysis may include data about the sample's spectral properties, such as its reflectance, absorbance, and other relevant properties.

[0168] In other embodiments, materials science and quantum computing are two distinct fields that have seen substantial advancements in recent years. Materials science involves the study and design of new materials, with a focus on discovering or predictingthe properties of these materials based on their atomic-scale structure. Quantum computing, on the other hand, leverages the principles of quantum mechanics to perform computational tasks more efficiently than classical computers.

[0169] In the realm of materials science, a common approach to predicting the properties of a material involves the use of Density Functional Theory (DFT). DFT is a computational quantum mechanical modelling method used to investigate the electronic structure of many-body systems, particularly atoms, molecules, and the condensed phases. It approximates the properties of a many-electron system in terms of functionals of the electron density. DFT simplifies the many-body Schrodinger equation by replacing the wavefunction of a many-body system with the electron density as the basic quantity.

[0170] On the other hand, quantum computing utilizes quantum bits, or qubits, which can exist in multiple states at once, thanks to the principle of superposition. This allows quantum computers to process a vast number of possibilities simultaneously. Quantum computing also leverages another quantum phenomenon known as entanglement, which allows qubits that are entangled to be linked together in such a way that the state of one qubit can directly influence the state of another, no matter how far apart they are.

[0171] Hamiltonians play a central role in both quantum computing and materials science. In quantum computing, a Hamiltonian is a mathematical operator corresponding to the total energy of the system. It is used to solve the Schrodinger equation, which describes how the quantum state of a quantum system changes over time. In materials science, a Hamiltonian can be used to describe the total energy of a material system, including both kinetic and potential energy.

[0172] Machine learning, a subset of artificial intelligence, involves the use of algorithms and statistical models to perform tasks without explicit instructions, relying on patterns and inference instead. It is used in a variety of computing tasks where designing and programming explicit algorithms is unfeasible. Machine learning models can be trained on a dataset and then used to predict or classify new data. Machine learning has been applied in various fields, including computer vision, speech recognition, natural language processing, and more recently, materials science and quantum computing.

[0173] Additionally, generating novel materials that fulfill specific needs involves extremely time-consuming efforts that often involve running a number of experiments thatare designed based on expert knowledge, and then refining those experiments and adjusting those experimental parameters in order to approach a more optimal solution and outcome. For example, those researching in this space struggle with the vastness and complexity of the materials design space (due often to a multitude of variables influencing material properties). Additionally, materials science often requires expertise in various fields, making collaborative efforts essential for breakthroughs. Further issues are exacerbated through traditional trial-and-error approaches. Such approaches are often very resourceintensive and costly, impacting the feasibility of potential discoveries.

[0174] The outcome of such approach is a desired physical material that retains the properties of interest, but which must undergo many cycles of experimentation and evaluation to determine the resulting material’s level of compliance and utility. As such, previous efforts to perform advanced materials research and development are typically slow and lead to protracted development cycles. Such efforts are typically siloed across organizations and infrastructures, which can increase friction and fail to capitalize on possible synergies. In addition, such efforts inherently bring with them greater data generation demands and higher computational costs. Further, such efforts are limited in the complexity and scale of systems that they can address.

[0175] Based on such known issues, many are turning to artificial intelligence (Al) for purposes of speeding up the research process for novel material generation, as well to drive down developmental costs. However, generating novel materials using Al faces several challenges that impact the effectiveness of the process, including issues of finding high quality (and availability) of materials data, issues with understanding the underlying principles governing materials behavior, and / or issues with high-cost barriers to involve an Al system.

[0176] In view of such considerations, the present disclosure introduces a generative atomistic design workflow for producing novel materials, including inorganic solid-state materials. The workflow integrates Density Functional Theory (DFT), Classical Machine Learning (CML), and / or Quantum Probabilistic Machine Learning (QPML). Further, the disclosure herein details a unique formulation that accelerates the research and optimization of known materials as well as the generative discovery of novel functional materials. By leveraging DFT and publicly available data, CML and QPML algorithmsmay be trained to predict material properties and dynamics. These computational models inform and arc corroborated by experimental fabrication, synthesis, and characterization experiments.

[0177] Furthermore, the system is able to leverage uncertainty-driven active learning cycles to reduce the data requirements for training these models. Such a workflow decreases data generation requirements and extends the simulation capacity to accommodate larger systems and enables generative design of new material configurations with desired properties.

[0178] As such, the present disclosure represents a significant advancement in the field of novel material generation by effectively addressing and resolving known issues. Through innovative approaches and methodologies, the disclosure herein has successfully overcome challenges and limitations that have previously impeded progress in the art (noted hereinabove). These novel solutions contribute to a more comprehensive understanding material generation, and in particular, integration of Al systems for purposes of generating new materials.

[0179] As such, by harnessing the power of Al systems in materials discovery, almost all markets can benefit from optimized properties and functionalities tailored to specific needs. For example, in manufacturing, new materials can provide enhanced efficiency, durability, and cost-effectiveness, driving innovation in product design and production processes. The healthcare industry may benefit from personalized and precisely engineered materials for medical devices and drug delivery systems. The energy sector may allow for improved performance and efficiency for renewable energy technologies and energy storage. The foregoing examples are not to be limiting in any manner and are intended as representations of the vastness of applicability of the new generated materials. Therefore, the transformative impact of these new materials, generated by the generative atomistic design workflow disclosed herein, extends beyond specific industries (and in fact may apply to any and all industries), fostering a new era of material technological advancement, creation, personalization, and sustainability.

[0180] In some aspects, the hyperspectral data discussed herein may work in combination with other sensed data, including multispectral, light detection and ranging (LiDAR), thermal infrared, synthetic aperture radar (SAR), digital elevation models(DEM), optical imagery, climate and / or meteorological, soil and / or geophysical, oceanographic, ground-based measurements, radio detection and ranging (RADAR), etc.

[0181] For example, it is recognized that radar may utilize radio waves to detect structural and topographic information (such as surface roughness, elevation, and moisture content). The hyperspectral data disclosed herein may include detailed spectral information across a wide range of wavelengths, allowing for the identification of specific materials, vegetation types, or surface compositions based on their unique spectral signatures. The hyperspectral data may operate and work in combination with the radar data such that radar data, comprising structural and textural differences, can be combined with the hyperspectral data, comprising identification of materials or surface compositions, resulting in a combined data set that distinguishes between different land cover types or vegetation species, even those that are spectrally similar.

[0182] In one embodiment, the combination of the hyperspectral data and the radar data may be of particular use for agriculture, where radar might be used to assess crop structure and soil moisture, and the hyperspectral data identifies crop health and species. Similarly, in disaster monitoring, radar data can track flood extent through cloud cover, and the hyperspectral data may assess vegetation stress or chemical contamination, providing a comprehensive assessment of the affected areas. Thus, the integration of radar and hyperspectral data may provide for more accurate and detailed environmental monitoring and resource management.

[0183] It is to be appreciated the hyperspectral data may be combined (as discussed above) with other forms of data. For example, when the hyperspectral data is integrated with multispectral data, the hyperspectral data may specify land cover classification and vegetation monitoring, and the multispectral data may offer broader spatial coverage. When the hyperspectral data is combined with LiDAR, the resulting data may provide a more comprehensive analyses of forest canopy health, urban infrastructure, and terrain features. When the hyperspectral data is combined with thermal infrared data, the resulting data may be particularly beneficial for drought stress detection and urban heat island studies. When the hyperspectral data is combined with Synthetic Aperture Radar (SAR), the resulting data may include structural and moisture-related insights in combination with the hyperspectral data comprising reflectance data, improve assessments of soil moisture,vegetation health, and land surface changes. When the hyperspectral data is combined with the digital elevation models (DEMs), the combined data may include greater information on topographic context. As such, the hyperspectral data can work in combination with a variety of other sensed data to provide a more comprehensive and accurate data tool for environmental monitoring, agriculture, urban planning, resource management, etc.

[0184] The use of the terms "a" and "an" and "the" and similar referents in the context of describing the subject matter (particularly in the context of the following claims) are to be construed to cover both the singular and the plural, unless otherwise indicated herein or clearly contradicted by context. Recitation of ranges of values herein are merely intended to serve as a shorthand method of referring individually to each separate value falling within the range, unless otherwise indicated herein, and each separate value is incorporated into the specification as if it were individually recited herein. Furthermore, the foregoing description is for the purpose of illustration only, and not for the purpose of limitation, as the scope of protection sought is defined by the claims as set forth hereinafter together with any equivalents thereof entitled to. The use of any and all examples, or exemplary language (e.g., "such as") provided herein, is intended merely to better illustrate the subject matter and does not pose a limitation on the scope of the subject matter unless otherwise claimed. The use of the term “based on” and other like phrases indicating a condition for bringing about a result, both in the claims and in the written description, is not intended to foreclose any other conditions that bring about that result. No language in the specification should be construed as indicating any non-claimed element as essential to the practice of the invention as claimed.

[0185] The embodiments described herein included the one or more modes known to the inventor for carrying out the claimed subject matter. Of course, variations of those embodiments will become apparent to those of ordinary skill in the art upon reading the foregoing description. The inventor expects skilled artisans to employ such variations as appropriate, and the inventor intends for the claimed subject matter to be practiced otherwise than as specifically described herein. Accordingly, this claimed subject matter includes all modifications and equivalents of the subject matter recited in the claims appended hereto as permitted by applicable law. Moreover, any combination of the above-described elements in all possible variations thereof is encompassed unless otherwise indicated herein or otherwise clearly contradicted by context.

Claims

CLAIMSWhat is claimed is:

1. A system for atomistic-based sensor fusion inference, comprising: a sensor configured to receive a geological sample and output an analysis of the geological sample; a neural network configured to predict at least one of an electronic structure, chemical complexity, molecular dynamics, or distribution of minerals of the geological sample based on the analysis; and a processor configured to: derive at least one of physical properties or chemical properties of the geological sample, create a machine-learned data set based on the geological sample, and save the machine-learned data set to a geophysics library.

2. The system of claim 1, wherein the sensor is configured to receive the geological sample from a soil sample.

3. The system of claim 2, wherein the sensor is further configured to perform x-ray crystallography on the soil sample.

4. The system of claim 1, wherein the neural network is further configured to apply density functional theory in predicting the electronic structure or molecular dynamics of the geological sample.

5. The system of claim 1, wherein the neural network is further configured to simulate the geological sample at the atomic scale over time and across different temperatures.

6. The system of claim 1, wherein the neural network is further configured to simulate the geological sample under various environmental conditions.

7. The system of claim 1 , wherein the processor is further configured to derive the physical properties or chemical properties of the geological sample based on the analysis from the sensor.

8. The system of claim 1, wherein the prediction is created using a weighted-average of the machine-learned data set and the analysis.

9. The system of claim 1, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in a soil sample.

10. The system of claim 1, wherein the processor is further configured to derive the physical properties and chemical properties of the geological sample based on the analysis from the sensor.

11. The system of claim 1, wherein the geological sample is obtained from a hyperspectral sensing conducted remotely.

12. The system of claim 1, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in a natural gas volume.

13. The system of claim 1, further comprising remote sensor platform comprising the sensor, wherein the remote sensor platform is configured to create an atomistic-based inference of a distribution of underlying geophysical resources located proximate to the geological sample based on the machine-learned data set and the analysis.

14. The system of claim 13, wherein the atomistic -based inference is used to predict the distribution of underlying minerals for a given surface area represented by the geological sample.

15. The system of claim 1, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in the geological sample.

16. The system of claim 1, wherein the machine-learned data set includes a spectral reflectance graph of minerals and materials found in an oil deposit.

17. The system of claim 1 , wherein the neural network applies at least one of density functional theory or molecular dynamics to create the prediction.

18. The system of claim 1, wherein the machine-learned data set is used to create geophysical priors for reverse engineering hyperspectral signatures.

19. A system for atomistic -based sensor fusion inference, comprising: a remote sensor platform configured to obtain sensor data; a processor configured to retrieve a machine-learned data set from a geophysics library; and the remote sensor platform configured to create an atomistic-based inference of a distribution of underlying geophysical resources located proximate to a geological sample based on the machine-learned data set and the sensor data.

20. A system for atomistic -based sensor fusion inference, comprising: a sensor configured to receive a geological sample; a neural network configured to: receive, at at least one computing device, one or more datasets corresponding to possible materials based on the geological sample, create, using at least two machine learning models, a new dataset for the possible materials, wherein the at least two machine learning models are trained based on the one or more datasets, to model properties of the possible materials, and output, using the at least one computing device, a prediction comprising the possible materials; and a remote sensor platform configured to create an atomistic-based inference of a distribution of underlying geophysical resources located proximate to the geological sample based on the prediction.

Images