Automatic terrain-based detection of well pads and their surroundings

Abstract

Aspects of the invention include detecting a first well pad at the first location based at least in part on a first dataset including spectral data describing gas emissions from the first location using a first machine learning model, detecting an environmental event within a threshold distance of the well pad, and determining the likelihood of damage to the first well pad from the environmental event.

Description

[Technical field]

[0001] The present invention relates generally to computer-based detection systems, and more particularly to automated terrain-based detection of a well pad and its surroundings under operating conditions. [Background technology]

[0002] Pad drilling is a drilling technique in which multiple wellbore are drilled from a single compact piece of land called a well pad. A wellbore is a hole that defines an oil well. A well pad is a piece of land where a drilling rig is cleared to drill the wellbore. A well pad is typically 4 to 5 square acres and can contain any number of oil wells. Pad drilling uses vertical drilling along with horizontal drilling. The drilling rig drills a wellbore vertically into the earth, and once the wellbore reaches a certain depth, the drilling rig drills horizontally in any given direction. Once the drilling rig has completed drilling the well, the oil rig is deployed to extract any oil from the ground. The oil extraction process requires specialized equipment, including the oil rig, protective barriers, and storage tanks. Summary of the Invention

[0003] Embodiments of the present invention are directed to automated terrain-based detection of well pads. A non-limiting exemplary computer-implemented method includes detecting a first well pad at a first location based at least in part on a first data set including spectral data describing gas emissions from the first location using a first machine learning model; detecting an environmental event within a threshold distance of the well pad; and determining a likelihood of damage to the first well pad from the environmental event.

[0004] According to one embodiment, the environmental event is a fire, a landslide, or a flood.

[0005] According to one embodiment, the computer-implemented method includes predicting a suitable location for a second well pad at the second location based at least in part on a similarity of the topography of the second location and the first location. The computer-implemented method further includes predicting a likelihood of damage due to an environmental event at the second location.

[0006] According to one embodiment, the computer-implemented method further includes extracting features from the first data set using a second machine learning model to determine a topography of the location.

[0007] According to one embodiment, the computer-implemented method further includes encoding the data to delineate a boundary of a potential well pad based at least in part on the topography.

[0008] According to one embodiment, a first machine learning model determines whether a defined bounded potential well pad is a gas emission source.

[0009] According to one embodiment, the computer-implemented method further includes determining an identity, composition, and volume of the gas emission, and comparing the emission to regulatory requirements.

[0010] Other embodiments of the present invention implement features of the above methods in computer systems and computer program products.

[0011] Additional technical features and advantages are realized through the techniques of the present invention. Embodiments and aspects of the present invention are described in detail herein and are considered a part of the claimed subject matter. For a better understanding, reference should be made to the detailed description and drawings.

[0012] The particulars of the exclusive rights set forth herein are particularly pointed out and distinctly claimed in the claims at the conclusion of this specification. The foregoing and other features and advantages of embodiments of the present invention will become apparent from the following detailed description taken in conjunction with the accompanying drawings. [Brief description of the drawings]

[0013] [Figure 1] FIG. 1 is a block diagram of components of a well pad detection system in accordance with one or more embodiments of the present invention. [Figure 2A] 1 is an image from an aerial source showing multiple potential well pads. [Figure 2B] 1 is a magnified image showing a single well pad including well pad related features. [Figure 3A] 1 is a map showing well pad boundaries according to one or more embodiments of the present invention. [Figure 3B] FIG. 2 illustrates a well pad boundary without a visible light image in accordance with one or more embodiments of the present invention. [Figure 4] 1 is a flow chart illustrating a process for well pad detection in accordance with one or more embodiments of the present invention. [Diagram 5] FIG. 1 illustrates a cloud computing environment in accordance with one or more embodiments of the present invention. [Figure 6] FIG. 2 illustrates abstraction model layers in accordance with one or more embodiments of the present invention. [Figure 7] FIG. 1 is a block diagram of a computer system for use in implementing one or more embodiments of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] The drawings shown herein are exemplary. There may be many variations to the drawings or the operations depicted in the drawings without departing from the spirit of the invention. For example, actions may be performed in a different order, or actions may be added, deleted, or modified. Additionally, the term "coupled" and variations thereof describe having a communication path between two elements and do not imply a direct connection between the elements without an intervening element / connection between the elements. All of these variations are considered part of this specification.

[0015] One or more embodiments of the present invention provide systems and methods for using machine learning techniques on spatially related imagery to identify physical structures, detect well pads, and assess flood, wildfire, and landslide risk to those well pads.

[0016] Commercial companies and regulatory agencies need accurate information about the number and location of well pads to determine discharge levels and accurately assess risk. Current regulations require well pad operators to cease operations under certain environmental events, such as floods, landslides, and wildfires, in the area surrounding the well pad. In the event of a flood, landslide, or wildfire, regulatory agencies must verify that the operator has ceased operations near the well pad.

[0017] It is therefore desirable to have the capability to remotely detect adverse environmental events near well pads and to calculate the distance between the epicenter of such environmental events and any nearby well pads. It is also desirable to calculate the speed at which an environmental event approaches any well pad and to calculate the time of arrival to provide advance warning to well pad operators. Traditional computer-based detection systems rely on photographic images consisting of red, green, and blue (RGB) bands for classification purposes by machine learning algorithms. However, these machine learning algorithms are often trained on outdated and inaccurate training data. Information is often provided in different formats depending on the source, which can result in high error margins.

[0018] One or more embodiments of the present invention address one or more of the above deficiencies by providing a method and system that uses spectral image data to detect a combination of topography, oil extraction equipment, and gas emissions to identify well pads. Spectral information extracted from satellites can also be used to detect infrared or high temperature signatures that can indicate wildfires, floods, and landslides. The methods and systems described herein further generate real-time and current training data for training machine learning units used to identify well pads and any nearby environmental conditions.

[0019] Turning now to Figure 1, a well pad detection system 100 is generally shown in accordance with one or more embodiments of the present invention. The well pad detection system 100 includes a data pre-processing unit 102, a terrain detection unit 104, and a well pad detection unit 106. The well pad detection system 100 is also operable to exchange data with at least one sensor 108 and at least one database 110 via a communication network 112. The sensor 108 may be mounted on a satellite, aircraft, balloon, unmanned aerial vehicle (UAV), or other aerial source. The database 110 may be any storage device that includes data describing a parcel of land.

[0020] The data pre-processing unit 102 receives data describing a parcel of land and prepares the data for analysis by the terrain detection unit 104 and the well pad detection unit 106. The data pre-processing unit 102 is operable to receive data from multiple sources and organize the received data regardless of the source. The data may include satellite imagery, spectral imagery, digital maps, or other data describing the parcel of land. The data may be organized based on one or more of, for example, location, time of capture, changes detected based on a reference image, proximity to an environmental event, level of oil production, level of detected gas, or other suitable criteria. The data pre-processing unit 102 includes a normalization unit 114 for normalizing the data from the different sources (i.e., the different sensors 108 and the database 110). The data pre-processing unit 102 further includes a feature extraction unit 116 for extracting distinctive features from the data, such as terrain-related features, well pad-related features, and environmental event-related features. For example, the topography of a well pad generally includes a flat rectangular portion of land, little or no vegetation, and distribution of drilling equipment. The feature extraction unit 116 can extract features indicative of these characteristics from data including spectral signatures and / or satellite imagery. The feature extraction unit 116 can also extract features describing the surrounding area, such as the type of vegetation and natural structures including mountains, caves, and bodies of water.

[0021] In some cases, one or more of these features may be associated with areas that do not include well pads. Thus, to distinguish between terrains that may include well pads and terrains that do not include well pads, the feature extraction unit 116 may also identify features that may be associated with other terrains. The feature extraction unit 116 may further code the data with metadata to indicate whether the feature is indicative of terrain with or without a well pad. The feature extraction unit 116 may also extract features that are indicative of terrains that are similar to well pad terrains but do not include well pads. For example, some agricultural environments include one or more features associated with well pad terrains but do not include any well pads. Thus, the feature extraction unit 116 may extract features associated with agricultural crops or farm-related structures. The terrain detection unit 104 and the well pad detection unit 106 may use this information to distinguish between well pad terrains and other terrains.

[0022] The data pre-processing unit 102 further includes a data augmentation unit 118. Training machine learning units typically require large data sets. The data augmentation unit 118 creates new data points / images by manipulating the original data. For example, the data augmentation unit 118 can receive image data and create multiple alterations of the image to increase the size of the training data. The alterations can include generating a mirror image, rotating the image, changing the color, cropping the image, scaling the image, or other alterations. The altered images are then appended to the original data set to create a larger data set for training the terrain detection unit 104 and the well pad detection unit 106. The data augmentation unit 118 can use different augmentation techniques, such as geometric transformation, color space augmentation, kernel filters, blended images, random elimination, feature space augmentation, adversarial training, generative adversarial networks (GANs), neural style transfer, and meta-learning.

[0023] The terrain detection unit 104 can receive data from the data pre-processing unit 102 and analyze the data for terrain features indicative of a well pad. The terrain feature extraction unit 120 detects features indicative of the terrain associated with a potential well pad. The features indicative of the terrain are analyzed by the terrain neural network model 122 to characterize the terrain (i.e., flat or rolling, vegetated or non-vegetated, boundaries, natural features, etc.). The terrain neural network model 122 can further track changes to the terrain from historical snapshots or images of the terrain. The snapshots are considered to be the "baseline" and the terrain neural network model 122 can identify the space and time in which such changes occurred from the baseline. These changes can be new structures, new equipment, changes in land use, changes in vegetation, vegetation types, and water body levels. By detecting the changes, the terrain detection unit 104 can determine whether any changes are indicative of a well pad. For example, road structures can indicate that a road is being built into a well pad.

[0024] When classifying the terrain of an unanalyzed parcel of land, the terrain feature extraction unit 120 receives data from the data pre-processing unit 102 and the geospatial-temporal data platform 124. The geospatial-temporal data platform 124 includes a repository of data, including maps, satellite images, weather drone data, and sensor-based device data. Data from both the terrain feature extraction unit 120 and the geospatial-temporal data platform 124 can relate to a particular parcel of land. The data received from the geospatial-temporal data platform 124 is used to augment the data received from the data pre-processing unit 102. The terrain feature extraction unit 120 processes the augmented data and extracts any terrain-related features.

[0025] The terrain neural network model unit 122 can receive the extracted features from the terrain feature extraction unit 120, process the information, and classify the land's topography. The topography can be, for example, a topography describing the land's elevation, vegetation, and man-made features (roads, buildings, infrastructure). The topography can also include the proximity of a location on the land to any other features such as rivers, roads, mountains, and buildings. The terrain neural network model unit 122 can analyze data, for example, raster or bitmap images, that store information about the topography to determine whether a location has a topography that is likely associated with a well pad.

[0026] Based on the data, the terrain neural network model unit 122 can determine characteristics of the parcel of land, such as slope, presence of gas, and type of vegetation. For example, the slope of the terrain can be determined by noting the difference in the values ​​of two adjacent pixels in the image data. If the difference is zero, the terrain is flat. However, if the difference is a positive number, the slope slopes toward the first pixel used to calculate the difference, and a negative slope indicates that the terrain may slope toward the second pixel. By examining the spectral image data, the terrain neural network model unit 122 can identify gases and create a density map of gas concentrations across the parcel of land. The terrain neural network model unit 122 can also detect vegetation by comparing the spectral characteristics from the image to a vegetation index (e.g., the Normalized Difference Vegetation Index). When sunlight hits different vegetation, the vegetation absorbs certain wavelengths and reflects other wavelengths. The cellular structure of the vegetation has evolved to reflect near-infrared (NIR) wavelength light, while chlorophyll or plant pigments absorb visible wavelengths of light. Thus, if the terrain feature extraction unit 120 determines that a portion of an image displays a higher percentage of infrared light than visible light, it may classify the portion as having dense vegetation versus sparse vegetation. The terrain neural network model unit 122 may also identify environmental events such as floods, landslides, or wildfires. For example, the terrain neural network model unit 122 may receive time series data and identify that the location of an identified fire or water is changing and moving in a certain direction. In other cases, the terrain neural network model unit 122 may analyze the time series data and identify that the location of an object (i.e., a rock, tree, or building) is moving in the same direction and identify that a landslide is occurring.

[0027] The terrain neural network model unit 122 can identify a portion of land as a potential well pad location based on characteristics of the terrain. The terrain neural network model unit 122 identifies based in part on the terrain having characteristics of a pre-identified well pad terrain. For example, the pre-identified well pad terrain includes a portion of land characterized as a generally vegetation-free geometric shape, the size of the geometric shape being within a threshold of an average geometric shape area around the vegetation-free well pad, and the terrain neural network model unit 122 can identify the terrain as a potential well pad location. After identifying the data describing the potential well pad location, the terrain neural network model unit 122 can code the data to segment the potential well pad location. The boundary can be in the form of a machine-readable code. As seen in Figures 3A and 3B, the code can also instruct the well pad detection system 100 to augment any image of the land with an example boundary that defines a potential well pad. The terrain neural network model unit 122 can then label the segmented images to identify each potential well pad location within the region.

[0028] In addition to defining boundaries for potential well pads, the terrain neural network model unit 122 may classify objects in the terrain surrounding each potential well pad. For example, the terrain neural network model unit 122 may classify the density or type of vegetation surrounding each potential well pad using the spectral imagery, data from the geospatial-temporal data platform, and vegetation indices. The terrain neural network model unit 122 may also identify other objects near the potential well pads. For example, the terrain neural network model unit 122 may use data provided by the geospatial-temporal data platform 124 to identify and determine proximity to other locations such as mountains, water, or towns. The terrain neural network model unit 122 may further determine the likelihood that the well pad is in proximity to each identified vegetation, object, and / or has a particular terrain profile. In this analysis, the distance between the potential well pad location and the object is calculated based on the line of sight. These calculations are performed to identify the closest potential well pads within an area and detect differences or similarities between them.

[0029] For example, there is often a minimum distance that must be maintained between two well pads, determined by the horizontal extent of the pipes used to extract oil from the ground. From the geospatial records, the direction and length of any pipes are combined with the distance between the potential well pad location and the pre-identified well pad to determine if the potential well pad location is outside of the minimum distance from the pre-identified well pad.

[0030] The terrain neural network model unit 122 may analyze the area surrounding the potential well pad location to detect any environmental events, such as tornadoes, hurricanes, fires, or floods that may damage the well pad. If an environmental event is detected, the terrain neural network model unit 122 further analyzes the time series data to determine a speed at which the environmental event approaches the potential well pad location. The terrain detection unit 104 is further operable to determine a threshold safety distance between the potential well pad location and the environmental event.

[0031] The terrain neural network model unit 122 may also predict whether a portion of land without a well pad will be suitable for a well pad. For example, if a determined well pad has a higher than threshold likelihood of having one or more particular characteristics (i.e., flat topography, proximity to a particular type of vegetation, and proximity to other identified well pads), the terrain neural network model unit 122 may identify land without a well pad but having the characteristics as a potential location for a well pad. The terrain neural network model unit 122 may further rank the potential locations based on their characteristics. For example, each terrain-related feature may be provided with a score representing the likelihood that it is associated with a well pad. Based on this score, the terrain neural network model unit 122 may rank one location as being more likely for a well pad than another location. The terrain neural network model unit 122 may assign each feature as indicating either a suitable location or an unsuitable location for a well pad and assign a likelihood score. Each likelihood score is a representation of how likely it is that a feature is associated with the topography of the well pad. Each likelihood score may be further weighted based on whether the terrain neural network model unit 122 has identified a combination of two or more features that makes the location more or less suitable for the well pad.

[0032] The well pad detection unit 106 receives data from the terrain detection unit 104 to determine whether a potential well pad location includes an actual well pad. The well pad detection unit 106 includes multiple machine learning models for multi-scale object detection. The well pad detection unit 106 can select a machine learning model based on the identified features of the terrain. For example, the well pad detection unit 106 can use a region convolutional neural network (R-CNN) model unit 126. The R-CNN model unit 126 receives data from the terrain detection unit and determines whether the terrain includes a well pad. In some embodiments of the invention, the R-CNN 126 first generates bounding boxes in the image received from the terrain detection unit and runs a classifier on each bounding box. Upon classifying different objects, the R-CNN model unit 126 removes duplicate classifications and rescores the boxes based on other detected objects in the image. In other embodiments of the invention, the well pad detection unit 106 can also use a You Only Look Twice (YOLT) model unit 128. The YOLT model unit 128 helps locate small dense objects in the image by using a sliding window, which increases the sampling rate of the image. Furthermore, the YOLT neural network architecture can be distinguished from other neural networks by including a dense final convolutional layer. In yet another embodiment of the present invention, the well pad detection unit 106 can also use a single shot detector (SSD) model unit 130. The SSD model unit 130 includes a backbone model and an SSD head. The backbone model is an image classification neural network used to extract features. The SSD head includes one or more convolutional layers that are added to the end of the backbone model neural network.

[0033] The well pad detection unit 106 selects an appropriate neural network model based on the topographical features determined by the terrain detection unit 104. The well pad detection unit 106 analyzes the potential well pad location using the appropriate neural network model to identify the presence of an actual well pad. The well pad detection unit 106 can compare any detected gas to gas associated with a well pad. Gas has unique spectral characteristics and reflects electromagnetic waves at specific frequencies and wavelengths. For example, methane or carbon dioxide are gases that are commonly emitted from well pads. Thus, the presence of gas can indicate that the potential well pad location does, in fact, include an actual well pad.

[0034] The well pad detection unit 106 may account for gas movement due to wind, temperature gradients, or other phenomena. The well pad detection unit 106 may apply fluid engineering models to the gas data to compensate for wind turbulence, which may cause plumes to expand spatially and also dilute concentrations. The distribution of these plumes may be expanded across the image and communicated to a neural network in terms of extent, shape, and expected concentration. Based on application of one or more of the models, the well pad detection unit 106 may then identify gas sources to aid in determining whether a parcel of land contains a well pad.

[0035] The well pad detection unit 106 can further use a suitable neural network model to compare the machinery and equipment detected at the location with the machinery associated with the well pad. For example, the well pad detection unit 106 can detect the presence of piping, oil rigs, gas container trucks, and other equipment and machinery to determine whether the location includes a well pad. In some cases, if there is enough gas in the air, the spectral characteristics of the gas can overwhelm the visible light data that indicates physical structures below the gas. Thus, the well pad detection unit 106 can use a suitable neural network model to analyze the time series images of the potential location to determine whether any physical structures are present on the potential well pad.

[0036] The well pad detection unit 106 may also determine the likelihood of an environmental event affecting a well pad. For example, the well pad detection unit 106 may analyze the topographical data using a suitable neural network model to predict the likelihood that a location will be affected by an environmental event. For example, how a location may be affected by a landslide. The well pad detection unit 106 may also analyze the speed of a wildfire to determine whether the well pad is at risk of a wildfire. The well pad detection unit 106 may further be operable to use natural language processing techniques to analyze historical data, such as weather forecasts, electronic news articles, and other information to optimize the prediction. The well pad detection unit 106 may use a suitable neural network model and apply the data to predict the likelihood that a well pad will be damaged by an environmental event and also whether a location without a well pad will be affected by an environmental event. In this sense, the well pad detection unit 106 may help determine the suitability of a piece of land to be considered a well pad location.

[0037] In response to determining that the location includes an actual well pad, the well pad detection unit 106 may also alert the well pad operator in response to the environmental event data. Based on the rate of the environmental event and the minimum safety threshold distance, the well pad detection unit 106 may transmit an alert to the well pad operator that includes a description of the environmental event, the rate of the environmental event, and the maximum time available to cease operation of the well pad before the environmental event exceeds the minimum safety threshold distance. The minimum safety threshold distance may vary based on the environmental event and the topography.

[0038] The well pad detection unit 106 may further use a suitable neural network model to determine whether the well pad is in compliance with existing regulations, including any shutdown orders from a regulatory agency. The well pad detection unit 106 receives as data the identity and composition of any gas emissions from the well pad and compares this value to thresholds established via regulations to determine whether the well pad is in compliance with applicable emission standards. This includes detecting the presence and level of gas emissions after a shutdown order to determine whether the well pad will shut down in response to the order.

[0039] The sensor 108 may be any sensor operable to collect data for spectroscopy. Spectroscopy involves surveying and measuring data from any portion of the electromagnetic spectrum. The sensor 108 may include, for example, a remote imaging sensor such as a Moderate Resolution Imaging Spectroradiometer (MODIS) or a Visible Infrared Imaging Radiometer Suite (VIIRS), or a Light Detection and Ranging System (LIDAR) acquired at various wavelengths capable of capturing spectral data. The sensor 108 may further include a gas imager sensitive to infrared spectral bands to detect chemical plumes that may be produced by oil and gas operations. Spectral bands are various ranges of electromagnetic signals along the electromagnetic spectrum. Examples of chemical plumes include methane, CH4, a gas that leaks from a well pad; sulfur dioxide, SO2, a natural by-product of oil and gas; and nitrogen dioxide, NO2, a gas that is released from pumps and compressors on the well pad. The spectral bands may be used to detect the spectral signatures of the respective gases. The spectral range of the sensor can be configured to include the spectral signature of any chemical species of interest.

[0040] The well pad detection system 100 receives and analyzes data from either the sensor 108 or the database 110 to identify well pads. The data can be recovered in real time as the sensor 108 passes over a piece of land, or the data can be pre-stored in the database 110. The well pad detection system 100 is also operable to receive data from third parties, such as a company operating a well pad, a city taxing authority, a regulatory agency, or an insurance company.

[0041] The communications network 112 may include the Internet, fiber optics, microwave, xDSL (Digital Subscriber Line), wireless local area network (WLAN) technology, satellite, wireless cellular technology, Bluetooth technology, or any other suitable communications technology, or combinations thereof.

[0042] The phrases "neural network" and "machine learning" broadly describe the ability of electronic systems to learn from data. A machine learning system, engine, or module may include machine learning algorithms that can be trained to learn functional relationships between currently unknown inputs and outputs, such as in an external cloud environment (e.g., cloud computing environment 50). In one or more embodiments, the machine learning function can be implemented using a topographical neural network model unit 122 that has the ability to be trained to perform currently unknown functions. In machine learning and cognitive science, neural networks are a family of statistical learning models inspired by the biological neural networks of animals, particularly the brain. Neural networks can be used to estimate or approximate systems and functions that depend on a large number of inputs.

[0043] The topographical neural network model unit 122 can be embodied as a so-called "neuromorphic" system of interconnected processor elements that act as simulated "neurons" and exchange "messages" between each other in the form of electronic signals. Similar to the so-called "plasticity" of synaptic neurotransmitter connections that carry messages between biological neurons, the connections in the topographical neural network model unit 122 that carry electronic messages between simulated neurons are provided with numerical weights that correspond to the strength or weakness of a given connection. During training, the weights can be adjusted and tuned based on experience that allows the topographical neural network model unit 122 to adapt to the inputs and to learn. After being weighted and transformed by functions determined by the designer of the network, the activation of these input neurons is passed on to other downstream neurons, often called "hidden" neurons. This process is repeated until an output neuron is activated. The activated output neuron identifies which letter has been read.

[0044] Figure 2A is an image showing the distribution of well pads 200, 202 throughout the area. Figure 2B is a close-up image of a single well pad 204. As shown in Figure 2B, the well pad 204 has a generally rectangular shape and is in close proximity to a roadway 206. The well pad 204 has a generally flat terrain with little to no vegetation and is surrounded by a protective fence 208. A storage tank 212 is located on top of the well pad and an extraction point 214 is designated at the center of the well pad 204. Contamination sources can occur on or between any features with a probability distribution throughout the well pad.

[0045] FIG. 3A illustrates a boundary definition map 300 generated from data collected by the sensor 108. The boundary definition map 300 can include one or more aerial images, such as satellite images or autonomous aircraft images. In some embodiments, the sensor 108 includes a light detection and ranging (LIDAR) system that measures distances to targets (e.g., land) by illuminating the target with a laser light. The boundary definition map 300 can be generated from one or more satellite images of a particular geographic area over a predetermined time span. Each detected well pad 302 includes a boundary defined on the boundary definition map 300. FIG. 3B illustrates well pad boundary definitions from the boundary definition map 300 without a surface image. Each boundary definition has a polygon and shows the perimeter of a well pad or the perimeter of multiple adjacent well pads.

[0046] Referring to FIG. 4, a flow diagram of a process for well pad detection according to one or more embodiments of the present invention is shown. In block 402, the well pad detection system receives data regarding a parcel of land. The land may include one or more well pads. The data may be received from one or more sources, including air-based sources or databases. In block 404, the well pad detection system selects any geospatial data from the data of the analyzed parcel of land. Geospatial data refers to any data regarding the characteristics of the parcel of land. This includes data describing the surface profile of the land, any structures, pre-identified well pads, equipment, or other man-made objects on the land. Geospatial data also includes any data describing the surrounding topography as well, including proximity to mountains, valleys, bodies of water, and other natural locations. In block 406, the well pad detection system applies a machine learning model to the data to determine topography-related characteristics of the land. Before providing the data to the machine learning model, the well pad detection system pre-processes the data in block 408, if not pre-processed. The well pad detection system further annotates the data with boundaries that identify each portion of the parcel of land as a potential well pad. At block 410, the well pad detection system applies a terrain-specific machine learning model to the data to classify each potential well pad as including or not including a well pad. At block 412, the well pad detection provides any additional pre-processing to the data. The terrain-specific machine learning model analyzes the data to determine if any gas associated with the well pad is discharging from any of the boundary areas. The terrain-specific machine learning model also analyzes the data to determine if any machinery or equipment associated with the well pad is found in any of the boundary areas. At block 414, the well pad detection system annotates the data as to whether the boundary area includes or does not include a well pad. If any gas is identified, the well pad detection system determines the volume and discharge rate of the gas based on the time series data.The well pad detection system then compares the gas identity and gas volume to any applicable emission standards or shutdown commands.

[0047] The well pad detection system is further operable to assist in any mitigation of the volume of gas discharged within the threshold or above the emission standards. The well pad detection system also analyzes the composition of the gas above the well pad and calculates the mixing ratio between the different gases, where methane is discharged from leaks while nitrogen dioxide can be discharged by the operation of the equipment. Based on the comparison, the well pad detection system can generate a report disclosing the gas emissions and the applicable emission standards. The well pad detection system can further remotely establish communication to instruct an entity to control either the oil rig by either slowing down the oil rig, speeding up the oil rig, or disabling the oil based on the gas emissions and any applicable emission standards. The well pad detection system can further remotely establish communication with another aircraft or satellite to verify its findings or change its course to collect subsequent data to determine if the gas emissions have dropped below the emission standards or in accordance with a shutdown command.

[0048] Although this disclosure includes detailed descriptions of cloud computing, it will be understood that implementation of the teachings described herein is not limited to a cloud computing environment. Rather, embodiments of the invention may be implemented in conjunction with any other type of computing environment now known or later developed.

[0049] Cloud computing is a model of service delivery for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, network bandwidth, servers, processing, memory, storage, applications, virtual machines, and services) that can be instantly provisioned and released with minimal administrative effort or interaction with a service provider. The cloud model can include at least five characteristics, at least three service models, and at least four deployment models.

[0050] The characteristics are as follows:

[0051] On-Demand Self-Service: Cloud consumers can unilaterally provision computing capacity, such as server time and network storage, automatically as needed, without the need for human interaction with the service provider.

[0052] Pervasive Network Access: Capabilities are available over the network and accessed through standard mechanisms that facilitate use by thin or thick client platforms (e.g., cell phones, laptops, and PDAs) in heterogeneous environments.

[0053] Resource Pooling: Provider computing resources are pooled with different physical and virtual resources dynamically allocated and reallocated according to demand to serve multiple consumers using a multi-tenant model. Consumers generally have no control or knowledge over the exact location of the provided resources, although there is an impression of location independence in that it is possible to specify a location (e.g., country, state, or data center) at a high level of abstraction.

[0054] Rapid Elasticity: Capacity can be rapidly and elastically provisioned, sometimes automatically, to instantly scale out, and rapidly de-provisioned to instantly scale in. To the consumer, the capacity available for provisioning often appears unlimited and can be purchased in any quantity at any time.

[0055] Measured Services: Cloud systems automatically control and optimize resource usage by leveraging metering capabilities at some level of abstraction appropriate to the type of service (e.g., storage, processing, bandwidth, and active user accounts). Resource utilization is monitorable, controlled, and reportable, providing transparency to both providers and consumers of the services utilized.

[0056] The service model is as follows:

[0057] Software as a Service (SaaS): The ability offered to a consumer is to use a provider's applications running on a cloud infrastructure. The applications are accessible from a variety of client devices through thin-client interfaces such as web browsers (e.g., web-based email). The consumer does not manage or control the underlying cloud infrastructure, including the network, servers, operating systems, storage, or even individual application capabilities, with the possible exception of limited user-specific application configuration settings.

[0058] Platform as a Service (PaaS): The capability offered to a consumer is to deploy consumer-created or acquired applications, written using programming languages ​​and tools supported by the provider, on a cloud infrastructure. The consumer does not manage or control the underlying cloud infrastructure, including networks, servers, operating systems, or storage, but does have control over the deployed applications and possible application host environment configuration.

[0059] Infrastructure as a Service (IaaS): The capability provided to the consumer is to provide processing, storage, network, and other basic computing resources on which the consumer can deploy and run optional software, which may include operating systems and applications. The consumer does not manage or control the underlying cloud infrastructure, but does have control over the operating systems, storage, deployed applications, and possibly limited control over the selection of networking components (e.g., host firewalls).

[0060] The deployment model is as follows:

[0061] Private Cloud: The cloud infrastructure is operated solely for the organization. It can be managed by the organization or a third party and can reside on-premise or off-premise.

[0062] Community Cloud: The cloud infrastructure is shared by several organizations to support a specific community with shared concerns (e.g., mission, security requirements, policies, and compliance concerns). It can be managed by the organization or a third party and can reside on-premise or off-premise.

[0063] Public Cloud: Cloud infrastructure is made available to the general public or large business groups and is owned by an organization that sells cloud services.

[0064] Hybrid Cloud: The cloud infrastructure is still a unique entity, but it is a composite of two or more clouds (private, community, or public) tied together by standardized or proprietary technologies that enable data and application portability (e.g., cloud bursting for load balancing between clouds).

[0065] Cloud computing environments are stateless, loosely coupled, modular, and service-oriented, with a focus on semantic interoperability. At the heart of cloud computing is an infrastructure that includes a network of interconnected nodes.

[0066] Referring now to FIG. 5, an exemplary cloud computing environment 50 is shown. As shown in the figure, the cloud computing environment 50 includes one or more cloud computing nodes 10 with which a cloud consumer can use a local computing device, such as, for example, a personal digital assistant (PDA) or mobile phone 54A, a desktop computer 54B, a laptop computer 54C, or an automotive computer system 54N, or a combination thereof. The nodes 10 can communicate with each other. They can be physically or virtually grouped in one or more networks, such as a private, community, public, or hybrid cloud, or a combination thereof, as previously described (not shown). This enables the cloud computing environment 50 to provide infrastructure, platform, and / or software as a service, without the cloud consumer having to maintain resources on a local computing device. It will be understood that the types of computing devices 54A-N shown in FIG. 5 are merely exemplary, and that the computing node 10 and cloud computing environment 50 are capable of communicating with any type of computerized device via any type of network and / or network addressable connections (e.g., using a web browser).

[0067] Referring now to Figure 6, a set of functional abstraction layers provided by cloud computing environment 50 (Figure 5) is shown. It should be understood in advance that the components, layers, and functions shown in Figure 6 are for illustrative purposes only, and embodiments of the present invention are not limited thereto. As shown in the figure, the following layers and corresponding functions are provided:

[0068] Hardware and software layer 60 includes hardware and software components. Examples of hardware components include mainframes 61, RISC (reduced instruction set computing) architecture based servers 62, servers 63, blade servers 64, storage devices 65, and networks and networking components 66. In some embodiments, software components include network application server software 67 and database software 68.

[0069] The virtualization layer 70 provides an abstraction layer over which virtual entities such as virtual servers 71 , virtual storage 72 , virtual networks including virtual private networks 73 , virtual applications and operating systems 74 , and virtual clients 75 can be provided.

[0070] In one example, the management layer 80 may provide the functions described below. Resource provisioning 81 provides dynamic procurement of computing and other resources utilized to execute tasks within the cloud computing environment. Metering and pricing 82 provides cost tracking as resources are utilized within the cloud computing environment and billing or invoicing for the consumption of these resources. In one example, these resources may include application software licenses. Security provides identity verification for cloud consumers and tasks, and protection of data and other resources. User portal 83 provides access to the cloud computing environment for consumers and system administrators. Service level management 84 provides cloud computing resource allocation and management to meet requested service levels. Service level agreement (SLA) planning and achievement 85 provides pre-negotiation and procurement of cloud computing resources in anticipation of future requirements according to SLAs.

[0071] The workload tier 90 provides examples of functionality available to a cloud computing environment. Examples of workloads and functionality that can be provided from this tier include mapping and navigation 91, software development and lifecycle management 92, virtual classroom education distribution 93, data analytics processing 94, transaction processing 95, and well pad classification from imagery 96.

[0072] It will be appreciated that the present disclosure can be implemented with any other type of computing environment now known or later developed. For example, FIG. 7 illustrates a block diagram of a processing system 700 for implementing the techniques described herein. In an example, the processing system 700 has one or more central processing units (processors) 721a, 721b, 721c, etc. (collectively or generically referred to as processors 721 and / or processing devices). In an embodiment of the present disclosure, each processor 721 can include a reduced instruction set computer (RISC) microprocessor. The processors 721 are coupled to a system memory (e.g., random access memory (RAM) 724) and various other components via a system bus 733. A read only memory (ROM) 722 is coupled to the system bus 733 and can include a basic input / output system (BIOS) that controls certain basic functions of the processing system 700.

[0073] Also shown are an input / output (I / O) adapter 727 and a network adapter 726 coupled to the system bus 733. The I / O adapter 727 can be a small computer system interface (SCSI) adapter that communicates with a hard disk 723 and / or a storage device 725, or any other similar components. The I / O adapter 727, the hard disk 723, and the storage device 725 are collectively referred to herein as mass storage 734. An operating system 740 for execution on the processing system 700 can be stored in the mass storage 734. The network adapter 726 interconnects the system bus 733 with an external network 736, allowing the processing system 700 to communicate with other such systems.

[0074] A display (e.g., a display monitor) 735 is connected to the system bus 733 by a display adapter 732, which may include a graphics adapter for improving performance of graphics-rich applications and a video controller. In one embodiment of the present disclosure, adapters 726, 727, and / or 732 may be connected to one or more I / O buses that are connected to the system bus 733 through an intermediate bus bridge (not shown). I / O buses suitable for connecting peripheral devices such as hard disk controllers, network adapters, and graphics adapters typically include common protocols such as Peripheral Component Interconnect (PCI). Additional input / output devices are shown connected to the system bus 733 through a user interface adapter 728 and a display adapter 732. Input devices 729 (e.g., keyboard, microphone, touch screen, etc.), input pointer 730 (e.g., mouse, track pad, touch screen, etc.), and / or speakers 731 can be interconnected to the system bus 733 via a user interface adapter 728, which may include, for example, a super I / O chip that integrates multiple device adapters into a single integrated circuit.

[0075] In some embodiments of the present disclosure, the processing system 700 includes a graphics processing unit 737. The graphics processing unit 737 is a specialized electronic circuit designed to manipulate and modify memory to accelerate the creation of images in a frame buffer intended for output to a display. In general, the graphics processing unit 737 is very efficient at operating in computer graphics and image processing, and has a highly parallel structure that makes it more effective than a general purpose CPU for algorithms in which the processing of large blocks of data is performed in parallel.

[0076] Thus, as configured herein, processing system 700 includes processing capability in the form of processor 721, storage capability including system memory (e.g., RAM 724) and mass storage 734, input means such as keyboard 729 and mouse 730, and output capability including speaker 731 and display 735. In some embodiments of the present disclosure, a portion of the system memory (e.g., RAM 724) and mass storage 734 collectively store an operating system 740 for coordinating the functionality of the various components illustrated within processing system 700.

[0077] Various embodiments of the present invention are described herein with reference to the associated drawings. Alternative embodiments of the present invention may be devised without departing from the scope of the present invention. Various connections and relationships (e.g., above, below, adjacent, etc.) are described between elements in the following description and drawings. These connections and / or relationships may be direct or indirect unless otherwise specified, and the present invention is not intended to be limited in this respect. Thus, coupling of entities may refer to either direct or indirect coupling, and relationships between entities may be direct or indirect. Additionally, various tasks and processing steps described herein may be combined into a more comprehensive procedure or process having additional steps or functions not described in detail herein.

[0078] One or more of the methods described herein may be implemented using any one or combination of technologies, such as discrete logic circuitry having logic gates for implementing logical functions on data signals, application specific integrated circuits (ASICs) having appropriate combinatorial logic gates, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), or the like, each of which is well known in the art.

[0079] For the sake of brevity, conventional techniques for making and using aspects of the present invention may or may not be described in detail herein. In particular, various aspects of computing systems and specific computer programs for implementing various technical features described herein are well known. Thus, for the sake of brevity, many conventional implementation details are merely briefly mentioned herein or omitted altogether without providing details of well-known systems and / or processes.

[0080] In some embodiments, various functions or operations may be performed at a given location, or in conjunction with the operation of one or more devices or systems, or both. In some embodiments, a portion of a given function or operation may be performed at a first device or location, and the remaining portion of the function or operation may be performed at one or more additional devices or locations.

[0081] The terms used herein are for the purpose of describing particular embodiments only and are not intended to be limiting. As used herein, the singular forms "a", "an" and "the" are intended to include the plural unless the context clearly indicates otherwise. It will be further understood that the term "comprises" or "comprising" as used herein specifies the presence of stated features, integers, steps, operations, elements, or components, or combinations thereof, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or groups thereof, or combinations thereof.

[0082] The corresponding structures, materials, acts, and equivalents of all means or steps and functional elements in the following claims are intended to include any structures, materials, or acts for performing functions in combination with other claim elements specifically claimed. The present disclosure has been presented for purposes of illustration and description, but is not intended to be exhaustive or limited to the disclosed form. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the present disclosure. The embodiments have been selected and described in order to best explain the principles and practical applications of the present disclosure, and to enable others skilled in the art to understand the disclosure of the various embodiments with various modifications as suitable for the particular use contemplated.

[0083] The drawings shown herein are exemplary. Many variations to the drawings or steps (or operations) described are possible without departing from the spirit of the disclosure. For example, actions can be performed in a different order, or actions can be added, deleted, or modified. Also, the term "coupled" describes having a signal path between two elements, and does not imply a direct connection between the elements with no intervening elements / connections between them. All of these variations are considered to be part of the disclosure.

[0084] The following definitions and abbreviations are to be used for the interpretation of the claims and the specification. As used herein, "comprises," "comprising," "includes," "including," "has," "having," "contains," or "containing," or any other variation thereof, is intended to cover a non-exclusive inclusion. For example, a composition, mixture, process, method, article, or device that comprises a list of elements is not necessarily limited to only those elements, but can include other elements that are not expressly listed or that are inherent to such composition, mixture, process, method, article, or device.

[0085] Additionally, the term "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or design described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments or designs. The terms "at least one" and "one or more" are understood to include any integer greater than or equal to one, i.e., 1, 2, 3, 4, etc. The term "multiple" is understood to include any integer greater than or equal to two, i.e., 2, 3, 4, 5, etc. The term "connected" can include both indirect and direct "connections."

[0086] The terms "about," "approximately," and "approximately," and variations thereof, are intended to include the degree of error associated with the measurement of a particular quantity based on the equipment available at the time of filing this application. For example, "about" can include a range of ±8%, or 5%, or 2% of a given value.

[0087] The present invention may be a system, method, or computer program product, or combination thereof, at any possible level of technical detail integration. The computer program product may include a computer-readable storage medium having computer-readable program instructions for causing a processor to implement aspects of the present invention.

[0088] A computer readable storage medium may be a tangible device capable of holding and storing instructions for use by an instruction execution device. A computer readable storage medium may be, for example, but not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of computer readable storage media includes portable computer diskettes, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disk read only memory (CD-ROM), digital versatile disk (DVD), memory sticks, floppy disks, mechanical coding devices such as punch cards or ridge structures in grooves with instructions recorded thereon, and any suitable combination of the foregoing. Computer-readable storage media, as used herein, should not be construed as being, itself, a ephemeral signal, such as an electric wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., light pulses passed through a fiber optic cable), or an electrical signal transmitted through a wire.

[0089] The computer readable program instructions described herein can be downloaded from a computer readable storage medium to each computing / processing device or to an external computer or external storage device over a network, such as the Internet, a local area network, a wide area network, or a wireless network, or a combination thereof. The network can comprise copper transmission cables, optical transmission fiber, wireless transmission, routers, firewalls, switches, gateway computers, or edge servers, or a combination thereof. A network adapter card or network interface in each computing / processing device receives the computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium in the respective computing / processing device.

[0090] The computer readable program instructions for carrying out the operations of the present invention may be either assembler instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, configuration data for integrated circuit elements, or source or object code written in any combination of one or more programming languages, including object oriented programming languages ​​such as Smalltalk®, C++, and procedural programming languages ​​such as the “C” programming language or similar programming languages. The computer readable program instructions may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter scenario, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet Service Provider). In some embodiments, electronic circuitry, including, for example, a programmable logic circuitry, a field programmable gate array (FPGA), or a programmable logic array (PLA), can execute computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry to perform aspects of the present invention.

[0091] Aspects of the present invention are described herein with reference to flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer readable program instructions.

[0092] Such computer readable program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, where the instructions, executed by the processor of the computer or other programmable data processing apparatus, produce means for performing the functions / operations specified in one or more blocks of the flowcharts and / or block diagrams. Such computer readable program instructions may also be stored in a computer readable storage medium having instructions stored thereon to direct a computer, programmable data processing apparatus, or other device, or combination thereof, to function in a particular manner to perform aspects of the functions / operations specified in one or more blocks of the flowcharts and / or block diagrams.

[0093] The computer-readable program instructions may be loaded into a computer, other programmable data processing apparatus, or other device to perform the functions / operations specified in one or more blocks of the flowcharts and / or block diagrams, and may cause a series of operational steps to be performed on the computer, other programmable apparatus, or other device to create a computer-implemented process.

[0094] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in the flowcharts or block diagrams may represent a module, segment, or part of instructions, comprising one or more executable instructions for implementing a specified logical function. In some alternative implementations, the functions shown in the blocks may be performed out of the order shown in the figures. For example, two blocks shown in sequence may in fact be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order depending on the functionality involved. It will also be noted that each block of the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented by a special-purpose hardware-based system that performs the specified functions or operations or performs a combination of special-purpose hardware and computer instructions.

[0095] The description of various embodiments of the present invention has been presented for illustrative purposes, but is not intended to be exhaustive or limited to the disclosed embodiments. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the described embodiments. The terms used in this specification have been selected to best explain the principles of the embodiments, practical applications, or technical improvements over the art found in the market, or to enable others skilled in the art to understand the embodiments described herein.

Claims

1. 1. A computer-implemented method comprising: identifying potential well pad locations based on terrain characteristics; using a first machine learning model to detect an identity, composition, and volume of the gas emission based at least in part on a first dataset including spectral data describing a gas emission from a first one of the potential well pad locations, thereby identifying a presence of a first well pad at the first location; using a second machine learning model to detect an environmental event that may affect the first well pad, the environmental event being selected from the group consisting of a fire, a landslide, and a flood, within a threshold distance of the first well pad based on the terrain characteristics, the spectral data, and the historical data of the first location, and to predict the likelihood of damage to the first well pad from the environmental event; 4. A computer-implemented method comprising:

2. providing a likelihood score for a suitable location for a second well pad at the potential well pad location based at least in part on a similarity of the topographic characteristics of the first location to a second well pad at the potential well pad location; using the second machine learning model to predict a likelihood of damage due to the environmental event at the second location based on characteristics of the terrain and historical data of the second location; The computer-implemented method of claim 1 , further comprising:

3. 2. The computer-implemented method of claim 1 , wherein the first data set includes data related to gas movement, and further comprising determining, by the first machine learning model or a combination of the first machine learning model with other models, whether the first well pad is a source of the gas emission.

4. A computer program product which, when run on a computer, causes the computer to carry out the method according to any one of claims 1 to 3.

5. A memory having computer readable instructions; and One or more processors for executing the computer-readable instructions A system comprising: The computer readable instructions include instructions for controlling the one or more processors to perform a method according to any one of claims 1 to 3.

Images