Methods and systems for locating a borehole collar

Abstract

A method for determining a ground plane on which a collar of a borehole is located. One or more processors are configured to receive imaging data of a region surrounding the collar, where the imaging data is generated by scanning the region with one or more sensors. The imaging data is processed by the one or more processors to generate an estimate of a pre-drilled ground plane of the collar. An indication of a reduced level of the collar is determined by the one or more processors from the estimate of the pre-drilled ground plane.

Description

"Methods and systems for locating a borehole collar" Technical Field

[0001] The present invention relates to logging of a borehole, and specifically to determining a location of a collar of the borehole to improve post-drilling logging data, such as for example measured depth values, generated for the borehole. Background

[0002] The term “borehole” is used to collectively refer to any of the various types of holes that may be drilled into a ground surface. Boreholes are created by a drilling process generally performed by a drill rig, for example in order to perform resource extraction or geotechnical investigation or assessment of an environmental site, such as a mine site, for example to enable the collection of soil samples, water samples or rock cores, or to install monitoring wells or piezometers.

[0003] Borehole measurement (or “logging”) systems obtain measurements of the geological parameters of the interior of a borehole via the use of a logging device containing sensors, probes and / or other instrumentation components that are configured to obtain estimates of particular parameters. Depth measurement of a borehole is a particularly important function of a borehole measurement system, in order for example, to determine the depth of the hole, or indications of the depths at which particular features of the borehole occur (e.g., the occurrence of mineral deposits).

[0004] For example, by logging many holes it is possible to build up a model of one or more geological formations or bodies beneath the surface in terms of their properties over a range of depths. Specifically, this data enables a picture of the geological formations to be created in the form of a geological block model. The geological block model provides utility for assessing the surface and each individual borehole within, such as by assisting with increased efficiencies in planning and operating a mine site.

[0005] Measurement data, including depth measurement values, may be obtained during the drilling of the borehole. This data may include directional-drilling measurements, e.g., for decision support for the wellbore path (generally referred to as “measurement while drilling” (MWD) data), and data related to the geological formations penetrated while drilling (generally referred to separately as “logging while drilling” (LWD) data). The various data that are obtained during the drilling of the borehole are referred to collectively as “hole drilling data” herein.

[0006] In addition, or as an alternative to relying on hole drilling data to profile the borehole, it is often desirable to obtain measurements from the borehole in its post- drilled state. This is referred to as a post-drilling measurement, or “hole logging”, and typically involves the insertion of a measurement device at least partially into the hole, and the transmission of the obtained data to instruments on or above the surface. The measurements obtained by the hole logging process provide additional information of the hole and its surrounding strata, that are unable to be measured at the time of drilling. Summary

[0007] There is provided a method for determining a ground plane on which a collar of a borehole is located, the method comprising: receiving imaging data of a region surrounding the collar, the imaging data generated by scanning the region with one or more sensors; processing the imaging data to generate an estimate of a pre-drilled ground plane of the collar; and determining an indication of a reduced level of the collar from the estimate of the pre-drilled ground plane.

[0008] In some embodiments, the one or more sensors comprises at least one of: one or more LIDAR devices; one or more stereo cameras; and one or more radar devices.

[0009] In some embodiments, processing the imaging data comprises: detecting, from the imaging data, one or more features of the region surrounding the collar; andprocessing the one or more features to determine a continuous pre-drilled ground plane segment of the collar.

[0010] In some embodiments, the one or more features of the region surrounding the collar comprise at least: a tailings feature indicating a drill tailings pile at least partially surrounding the collar; and a ground plane feature indicating one or more portions of a post-drilled ground plane associated with the drill tailings pile.

[0011] In some embodiments, the one or more features are detected, from the imaging data, using a machine learning model.

[0012] In some embodiments, the continuous pre-drilled ground plane segment is determined by performing semantic segmentation on the imaging data based on the one or more features of the region surrounding the collar.

[0013] In some embodiments, performing semantic segmentation on the imaging data comprises: extending one of the one or more portions of the post-drilled ground plane to approximate a continuous surface beneath the drill tailings pile; and determining the continuous pre-drilled ground plane segment from the approximated continuous surface.

[0014] In some embodiments, processing the imaging data comprises: (i) generating local ground plane data representing the continuous pre-drilled ground plane segment in a local reference frame of the one or more sensors; and (ii) transforming the local ground plane data to represent the continuous pre-drilled ground plane segment in a global reference frame.

[0015] In some embodiments, transforming the local ground plane data comprises: determining a position and an orientation of the one or more sensors in the global reference frame; and applying a transformation to positions and orientations of the local ground plane data, based on the position and an orientation of the one or more sensors in the global reference frame.

[0016] In some embodiments, the position and the orientation of the one or more sensors is determined using a localization and pose system associated with the one or more sensors, and a spatial arrangement of the one or more sensors relative to the localization and pose system.

[0017] In some embodiments, the local ground plane data comprises respective data points each having: a 3D cartesian coordinate vector; a roll offset; and a pitch offset.

[0018] In some embodiments, the method further comprises determining a position of the collar in the global reference frame from the estimated pre-drilled ground plane.

[0019] In some embodiments, the position of the collar is determined by calculating an intersection point between a line vector projected from the one or more sensors and the estimated pre drilled ground plane.

[0020] In some embodiments, the line vector is projected from a reference point of the one or more sensors towards a drill hole feature visually indicating a centre point of the collar, wherein the drill hole feature is determined from the imaging data.

[0021] In some embodiments, the method further comprises: receiving or determining a probe reference location of a measurement probe to be deployed into the borehole via the collar; and using the determined reduced level of the collar and the probe reference location to determine or adjust one or more depth values of measurement data generated by the measurement probe during its deployment.

[0022] In some embodiments, the imaging data is generated by scanning with the one or more sensors: the region surrounding the collar; and one or more components of a probe deployment system configured to deploy the measurement probe into the borehole via the collar, and wherein determining the probe reference location of the measurement probe comprises: processing the imaging data to detect the probe deployment system; and determining the probe reference location from the detection of the probe deployment system.

[0023] In some embodiments, determining the probe reference location further comprises: generating, from the imaging data, detection data identifying one or more relevant components of the probe deployment system; processing the detection data and the imaging data to determine local position data representing a local position of the probe reference location from the identification of the one or more relevant components of the probe deployment system; and transforming the local position data to determine a global position of the probe reference location as a reference position for measurements obtained from deploying the measurement probe into the borehole.

[0024] There is also provided a device for determining a ground plane on which a collar of a borehole is located, the device comprising: a collar detection system having one or more sensors configured to scan a region surrounding the collar from a reference position and orientation; and a controller having one or more processors in communication with the collar detection system, the one or more processors configured to determine the ground plane of the collar by performing any of the methods described herein.

[0025] There is also provided an apparatus for determining a ground plane on which a collar of a borehole is located, the apparatus comprising: one or more sensors configured to scan a region surrounding the collar from a collar detection reference position and orientation; and one or more processors configured to: receive imaging data of the region surrounding the collar, the imaging data generated by the scanning of the region by the one or more sensors; process the imaging data to generate an estimate of a pre-drilled ground plane of the collar; and determine an indication of a reduced level of the collar from the estimate of the pre-drilled ground plane.

[0026] In some embodiments, the one or more processors are further configured to: receive geological data generated by a measurement probe in response to deployment of the measurement probe into the borehole via the collar; and use the determined reduced level of the collar and a deployment reference location of the measurement probe to determine or adjust one or more depth values of the geological data.

[0027] In some embodiments, the apparatus further comprises a support member configured to hold the one or more sensors in an adjustable position and orientation around the region surrounding the collar of the borehole.

[0028] In some embodiments, the support member has an internal area configured to at least partially house the measurement probe, and an external surface configured to mount the one or more sensors.

[0029] In some embodiments, the one or more sensors comprises at least one of: one or more LIDAR devices; one or more stereo cameras; and one or more radar devices.

[0030] In some embodiments, processing the imaging data comprises: detecting, from the imaging data, one or more features of the region surrounding the collar; and processing the one or more features to determine a continuous pre-drilled ground plane segment of the collar.

[0031] In some embodiments, the one or more features of the region surrounding the collar comprise at least: a tailings feature indicating a drill tailings pile at least partially surrounding the collar; and a ground plane feature indicating one or more portions of a post-drilled ground plane associated with the drill tailings pile.

[0032] In some embodiments, the one or more features are detected, from the imaging data, using a machine learning model.

[0033] In some embodiments, the continuous pre-drilled ground plane segment is determined by performing semantic segmentation on the imaging data based on the one or more features of the region surrounding the collar.

[0034] In some embodiments, performing semantic segmentation on the imaging data comprises: extending one of the one or more portions of the post-drilled ground plane to approximate a continuous surface beneath the drill tailings pile; and determining thecontinuous pre-drilled ground plane segment from the approximated continuous surface.

[0035] In some embodiments, processing the imaging data comprises: (i) generating local ground plane data representing the continuous pre-drilled ground plane segment in a local reference frame of the one or more sensors; and (ii) transforming the local ground plane data to represent the continuous pre-drilled ground plane segment in a global reference frame.

[0036] In some embodiments, transforming the local ground plane data comprises: determining a position and an orientation of the one or more sensors in the global reference frame; and applying a transformation to positions and orientations of the local ground plane data, based on the position and orientation of the one or more sensors in the global reference frame.

[0037] In some embodiments, the position and the orientation of the one or more sensors is determined using a localization and pose system associated with the one or more sensors, and a spatial arrangement of the one or more sensors relative to the localization and pose system.

[0038] In some embodiments, the local ground plane data comprises respective data points each having: a 3D cartesian coordinate vector; a roll offset; and a pitch offset.

[0039] In some embodiments, the one or more processors are further configured to determine a position of the collar in the global reference frame from the estimated pre drilled ground plane.

[0040] In some embodiments, the position of the collar is determined by calculating an intersection point between a line vector projected from the one or more sensors and the estimated pre drilled ground plane.

[0041] In some embodiments, the line vector is projected from a reference point of the one or more sensors towards a drill hole feature visually indicating a centre point of the collar, wherein the drill hole feature is determined from the imaging data.

[0042] In some embodiments, the one or more processors are further configured to: receive or determine a probe reference location of a measurement probe deployed into the borehole via the collar; and use the determined reduced level of the collar and the probe reference location to determine or adjust one or more depth values of measurement data generated by the measurement probe during its deployment.

[0043] In some embodiments, the imaging data is generated by scanning with the one or more sensors: the region surrounding the collar; and one or more components of a probe deployment system configured to deploy the measurement probe into the borehole via the collar, and wherein the one or more processors are further configured to determine the probe reference location by: processing the imaging data to detect the probe deployment system; and determining the probe reference location from the detection of the probe deployment system.

[0044] In some embodiments, the one or more processors are further configured to determine the probe reference location by: generating, from the imaging data, detection data identifying one or more relevant components of the probe deployment system; processing the detection data and the imaging data to determine local position data representing a local position of the probe reference location from the identification of the one or more relevant components of the probe deployment system; and transforming the local position data to determine a global position of the probe reference location as a reference position for measurements obtained from deploying the measurement probe into the borehole. Brief Description of Drawings

[0045] Some embodiments are described herein below with reference to the accompanying drawings, wherein:

[0046] Figure 1a is a schematic diagram of a prior art apparatus configured to determine a collar location in a first deployment;

[0047] Figure 1b is a schematic diagram of the prior art apparatus of Fig.1a configured to determine a collar location in a second deployment;

[0048] Figure 2a is a block diagram of a collar locating apparatus, according to some embodiments;

[0049] Figure 2b is a schematic diagram of a first exemplary configuration of a support member of the collar locating apparatus of Fig.2a;

[0050] Figure 2c is a schematic diagram of a second exemplary configuration of a support member of the collar locating apparatus of Fig.2a;

[0051] Figure 2d is a schematic diagram of a third exemplary configuration of a support member of the collar locating apparatus of Fig.2a;

[0052] Figure 2e is a schematic diagram of a first exemplary collar locating apparatus with a support member used to locate the collar of a borehole, according to some embodiments;

[0053] Figure 2f is a schematic diagram of a second exemplary collar locating apparatus with a support member used to locate the collar of a borehole, according to some embodiments;

[0054] Figure 2g is a schematic diagram of a third exemplary collar locating apparatus with a support member used to locate the collar of a borehole, according to some embodiments;

[0055] Figure 3a is a flow diagram of a method for locating the collar of a borehole using a collar locating apparatus, according to some embodiments;

[0056] Figure 3b is a schematic diagram of performing a scanning operation with the first exemplary collar locating apparatus of Fig.2e, according to some embodiments;

[0057] Figure 3c is a schematic diagram of performing a scanning operation with the second exemplary collar locating apparatus of Fig.2f, where a deployment vehicle is in a first position on a surface, according to some embodiments;

[0058] Figure 3d is a schematic diagram of performing a scanning operation with the second exemplary collar locating apparatus of Fig.2f, where a deployment vehicle is in a second position on a surface, according to some embodiments;

[0059] Figure 3e is a schematic diagram of performing a scanning operation with the second exemplary collar locating apparatus of Fig.2f, where a deployment vehicle is in a third position on a surface, according to some embodiments;

[0060] Figure 3f is a schematic diagram of performing a scanning operation with the second exemplary collar locating apparatus of Fig.2f, where a deployment vehicle is in a fourth position on a surface, according to some embodiments;

[0061] Figure 4a is a flow diagram of a method for processing imaging data based on the determination of continuous segments of a ground plane of the collar of the borehole, according to some embodiments;

[0062] Figure 4b is a block diagram of an example detection model implemented as a neural network, according to some embodiments;

[0063] Figure 4c is a flow diagram of a method to perform semantic segmentation on imaging data of a region near to, around, and / or surrounding a collar of a borehole based on one or more features, according to some embodiments;

[0064] Figure 4d is an illustration of classified images of a region near to, around, and / or surrounding the collar of the borehole as generated by the semantic segmentation method of Fig.4c, according to some embodiments;

[0065] Figure 4e is a schematic diagram of an example model of features extracted from images of a region near to, around, and / or surrounding the collar, according to some embodiments;

[0066] Figure 5a is a flow diagram of a method for converting a set of ground plane portions, as detected from imaging data of the region near to, around, and / or surrounding the collar, to a continuous pre-drilled ground plane segment, according to some embodiments;

[0067] Figure 5b is a schematic diagram of the extension of a ground plane portion to generate an approximated continuous surface, according to some embodiments;

[0068] Figure 6a is a schematic diagram of a transformation of a local ground plane estimate of a collar to a global ground plane estimate of a collar, according to some embodiments;

[0069] Figure 6b is a schematic diagram of the determination of a reduced level of a collar using a global ground plane, according to some embodiments;

[0070] Figure 7a is a schematic diagram of a measurement probe configured to measure a borehole in response to the use of a collar locating apparatus to locate the collar, according to some embodiments;

[0071] Figure 7b is a flow diagram of a method for determining or correcting depth data values generated during measurement of a borehole with the measurement probe depicted in Fig.7a;

[0072] Figure 8a is a block diagram of the collar locating apparatus of Fig.2a configured to determine a probe reference location, according to some embodiments;

[0073] Figure 8b is a flow diagram of a method for locating the probe reference location using a collar locating apparatus, according to some embodiments;

[0074] Figure 8c is a picture of an example apparatus for locating the probe reference location, according to some embodiments;

[0075] Figure 8d is a schematic diagram of performing a scanning operation with the example apparatus of Fig.8c, where a deployment vehicle is in a first position on a surface, according to some embodiments; and

[0076] Figure 8e is a schematic diagram of performing a scanning operation with the example apparatus of Fig.8c, where a deployment vehicle is in a second position on a surface, according to some embodiments. Description of Embodiments

[0077] Obtaining accurate depth values in post-drilling measurements is of critical importance to logging a borehole. This is complicated by the presence of relative misalignment between the positions of the collar and the logging device at the commencement of the logging activity which results in depth errors occurring in the logging data generated by the measurement device.

[0078] Previous approaches have addressed this issue by using hole drilling data (i.e., MWD or LWD data) to improve the accuracy of depth values for geological measurements obtained from the measurement device post-drilling. For example, International Patent Publication No. WO2023060313A1, which is incorporated herein by reference, describes various devices, apparatus, and methods for correcting depth measurements against an offset in the reference position of the measurement device from which measurements are taken, relative to the position of the collar of the borehole on the drilling surface.

[0079] The correction and / or compensation of errors in the depth values of post- drilling logging data relies on knowledge of the collar position (or location) as obtained from the MWD or other hole drilling data such as the pre-drilling data. However, the MWD or other hole drilling data often provides a collar location that is inaccurate. Forexample, in a typical drilling workflow the collar location of the borehole is often specified at the pre-drilling stage. Inaccuracies in the hole drilling data may occur due to practical effects occurring during borehole drilling, such as for example as a result of the drill bit moving off a starting location that was recorded as the collar location by the operator in the pre-drilling state or prior to the start of drilling. This results in inaccuracies in the corresponding depth corrections that are performed using the hole drilling data based collar location methodology. Ground plane reduced levels

[0080] In an alternative approach, the collar location of the borehole is determined by calculating a ground plane reduced level (“RL”) of the borehole relative to a known reference level, or corresponding positional coordinate (e.g., the z cartesian coordinate). For example, the ground plane may be determined in a coordinate space of a localization or position tracking system (e.g., GNSS) by determining a reference location of a deployment system (e.g., a robot or vehicle) that deploys the measurement device during logging.

[0081] Fig.1a schematically illustrates an apparatus 10 of the prior art, comprising a measurement device 12 and deployment vehicle 16 configured to determine a location of a collar A of borehole 101 in a ground surface 19 by a conventional ground plane RL determination process. A collar RL is determined by calculating a RL of the true ground plane 20 based on a reference plane 22 determined relative to the deployment vehicle (the “vehicle plane”). For example, the vehicle plane 22 extends from the vehicle 16 to the collar A and is calculated from a known position on the vehicle 16, such as a position at a base of the vehicle 16 (e.g., tracks if an excavator) that touches the ground surface 19.

[0082] Respective locations of points within the vehicle plane 22 can be determined from a reference location of the vehicle 16. A localization system (e.g., a GNSS or combined GNSS and INS Fusion system) provides a global location of a fixed component of the vehicle 16 (e.g., antenna 14) which may be used to determine thevehicle plane 22. An elevation of the antenna 14 relative to the vehicle plane 22 is compensated for by applying a calibrated offset value in the coordinate space of the localization system, which allows the vehicle RL to be determined. Then, an estimated collar location A’ is calculated by performing a vector projection along an axis 12’ of the measurement device 12 to find a point of intersection between the measurement device 12 and the vehicle plane 22.

[0083] Determination of the collar location enables the generation of depth values by comparing the collar location to a corresponding position coordinate of the measurement device prior to commencement of and / or during the logging. However, the accuracy of the generated depth values is conditional on an alignment between the reference plane 22 and the true ground plane 20 of the collar. In the ideal case shown in Fig.1a, the reference plane 22 on which the deployment system 16 is positioned is perfectly coplanar with the ground plane 20 that surrounds the borehole collar location A. In this case, the “inferred as-logged” collar RL matches the true (pre-drilled) collar RL (i.e., the estimated collar location A’ is aligned, at least in the vertical coordinate, with true collar location A).

[0084] A drawback of the conventional use of a deployment reference plane to determine the collar RL is its sensitivity to elevation (or depression) of the reference plane 22, relative to the true ground plane 20, when the deployment system is positioned in a region around the borehole.

[0085] Fig.1b schematically illustrates prior art apparatus 10 of Fig.1a configured to determine a collar location A by performing a conventional ground plane RL determination in response to a build-up of debris 15 (referred to as “tailings”) on the ground surface 19 in a region around and / or surrounding the borehole collar. In order to get close to the borehole, the vehicle will be situated as close to the borehole as possible, consequently the vehicle will be driven onto the tailings. The tailings 15 cause an elevation or depression of the vehicle 16. Consequently, the reference plane 22 is elevated from the true ground plane 20 for which it is desired to measure the depth. This leads to several undesirable effects. First, as the measurement device 16 ismounted to the deployment vehicle 16, the elevation (or depression) of the vehicle 12 causes a corresponding offset in the position of the measurement device relative to the true collar location A. This affects the ability to position the measurement device 12 close to the collar prior to the commencement of logging. Further, the tailings 15 is often uneven across the ground surface 19 around the borehole collar, resulting in variations to the offset depending on the exact location of the deployment vehicle 16 relative to the borehole 101. This leads to inaccurate depth values, and inconsistency in the depth values at which particular measurements are recorded depending on the pathing, and / or eventual location, of the vehicle in positioning the measurement device.

[0086] Second, the true ground plane 20 and the reference plane 22 at times due to the ground conditions are not coplanar. Thus the inferred as-logged collar RL has an elevation that differs significantly from the true pre-drilled collar RL value. Fig.1b illustrates this issue occurring with a 2D offset between the estimated collar location A’ and the true collar location A. However, it is also possible that the ground 20 and reference 22 planes deviate in 3D, for example with a relative pitch or roll offset. The pitch or roll offset may not be common with any corresponding deviation experienced during the determination of the MWD data. In this case, the logging data does not correlate with the as drilled (MWD) data which uses a predetermined drilling start position as the collar location (i.e., which is not subject to deviations from the true collar location due to pitch or roll offset).

[0087] Third, the presence of the tailings 15 that results from drilling the borehole 101 often obscures the collar location A in the true (pre-drilled) ground plane 20. In some cases, the tailings 15 is built-up to an extent that it becomes difficult to determine the pre-drilled ground plane 20 from data of the borehole 101 obtained in its post- drilled state. These drawbacks lead to an inaccurate determination of the collar location and thereby on the ability to log the borehole with a desired degree of accuracy and efficiency. It is desired to develop devices, apparatus and methods that address one or more of these problems, or that at least provide a useful alternative.Overview

[0088] Disclosed herein are methods, devices, and apparatus for locating the collar of a borehole 101 as drilled on a surface, as depicted by Figs.2a-2g. Determining a location of the borehole collar according to the proposed approaches involves using one or more sensors 104 to scan a region near to, around and / or surrounding the borehole collar. The region near to, around and / or surrounding the borehole collar (referred to herein as the “collar region”) includes at least a portion of the collar of the borehole. Preferably, the region is defined in 3D space and includes one or more areas on the surface 19 near to, around and / or surrounding the collar. In some examples, the region includes one or more volumes above or below, near to, around and / or surrounding the collar. In some examples, the collar region extends beyond the collar of the borehole to include, for example, one or more portions of an outer surface 19 and a drill tailings pile 15 associated with the collar. In some examples, the collar region includes one or more other features that may visually obscure the collar on the surface. A computer processing device (the “controller”) is configured to receive imaging data that is generated of the collar region with the one or more sensors 104. Preferably the sensors carry this out by scanning, imaging, or otherwise capturing data representing the area in and around the collar region. This data capturing process performed by operating the sensors is generally referred to as “scanning” herein irrespective of the type of sensor(s) used. The controller is configured to process the imaging data to generate an estimate of a pre-drilled ground plane 24 of the collar. The controller then determines a reduced level (RL) of the collar from the estimated pre-drilled ground plane 24.

[0089] Advantageously, the determined RL of the collar automatically accounts for the presence of tailings 15 or other artefacts around and / or surrounding the borehole 101 in its post-drilled state on surface 19, thereby correcting for any elevation error that would otherwise be present in the determined collar RL when using a reference plane 22 calculated from a vehicle 16 deployed in the region (e.g., as part of the control system 120). An accurate location of the borehole collar 107 (e.g., an absolute or relative position in a known global coordinate reference frame) can be subsequently calculated from the determined RL. Measurement data obtained from the borehole 101,whether as MWD data or from post-drilling logging activity, can be generated with, or corrected to, the determined collar location thereby improving the borehole logging capability.

[0090] The proposed approach to locating the collar involves: (1) obtaining the pre- drilled ground plane estimate 24 by extracting a continuous segment from a geospatial (i.e., positional and orientation based) model of the collar region, as provided by the imaging data generated locally by a detection system 105 comprising the one or more sensors 104; and (2) translating the pre-drilled ground plane estimate from a local reference frame 205’ of the detection system 105 into a global reference frame 205 from which the collar RL is determined. The proposed approach thereby advantageously utilizes the imaging capabilities of the detection system positioned in the vicinity of the region surrounding the collar, while accounting for any positional offset, and / or deviation in orientation (e.g., roll / pitch), due to local conditions of the system to eliminate or at least reduce error in the determined collar RL. Generating imaging data

[0091] Scanning of the collar region by the one or more sensors 104 involves measuring at least geospatial information of the collar region. The one or more sensors 104 may include a single one or multiple types of sensors, scanners, cameras, emitters or vision capture devices (collectively referred to as “sensors” herein) such as, but not limited to light detection and ranging (LIDAR) devices, laser scanning devices, stereo cameras, and radar devices. It will therefore be apparent to the skilled reader that the “imaging data” refers to data generated from such operations collectively, rather than being limited to any specific type of visual data. The one or more sensors 104 collectively generate sensor output data representing information within an observable or sense-able area in 3D space, referred to as the “field of view” 106 of the one or more sensors 104. Sensor output data is produced during one or more scanning operations in which the field of view 106 of the sensor(s) 104 is configured to include at least the collar region such that data representing the same is captured by the sensor(s) 104.

[0092] The sensor output data (referred to as “scanning data”) is processed to generate the imaging data. The imaging data provides a representation of the physical characteristics of the collar region, in the post-drilled state of the borehole 101, from which an estimated collar location A’ can be determined. Various types and forms of the imaging data may be generated from the scanning data. For example, the imaging data may include two dimensional images (e.g., disparity maps) and / or three dimensional images (e.g., depth images). In some embodiments, the imaging data is generated from the scanning data generated by a plurality of sensors of different types. For example, the imaging data may include a point cloud representation of the collar region. Alternatively, or in addition, the imaging data may include a disparity map representation of the collar region.

[0093] Although some embodiments described herein perform the scanning and generation of imaging data for a collar region including the borehole collar and related features, it will be appreciated that the sensor(s) 104 may also be configured to scan one or more objects, including, for example, one or more components of a measurement probe deployment system, that are positioned within the field of view 106. That is, the field of view 106 of the sensor(s) 104 may include at least the collar region and optionally one or more objects, such that data representing the same is included within the imaging data (e.g., for the purpose of determining a reference location of a measurement probe deployed to log the borehole via the collar).

[0094] In some embodiments, the imaging data is generated by a control unit local to the sensor(s) 104 in response to the scanning operation, where the control unit and the sensor(s) 104 form a collar detection system 105 deployed to evaluate the borehole 101. In other embodiments, the controller 222 includes one or more processors configured to directly or indirectly communicate with the one or more sensors 104 to receive scanning data from the sensor(s) and to generate the imaging data from the same.Forming the pre-drilled ground plane

[0095] Formation of the pre-drilled ground plane estimate involves processing the imaging data to generate a continuous ground plane segment. In some embodiments, the processing is conducted by a computing or processing device (e.g., the controller) configured to perform feature extraction and classification (recognition) operations on the imaging data to identify one or more features of the collar region associated with the continuous ground plane segment. In one implementation, the features of the collar region include a tailings feature indicating a drill tailings pile at least partially surrounding the collar, a ground plane feature indicating one or more portions of a post- drilled ground plane surrounding the drill tailings pile, and the actual drill hole (and its centre point). In some embodiments, the processing device is configured to identify one or more other features of the collar region including otherwise unspecified features (e.g., detected as an unspecified abnormality in the region).

[0096] In some embodiments, the controller is configured to identify the features of the collar region using machine learning. For example, the controller may perform a probabilistic detection of a tailings feature by accessing a borehole collar model trained with a priori data gathered from one or more other boreholes in the post-drilling state (e.g., a deep neural network trained using supervised or unsupervised learning on the training data).

[0097] In other embodiments, the controller is configured to identify the features of the collar region by analytical processing of the imaging data. For example, the controller may identify the presence of one or more objects in a point cloud or disparity map representation of the collar region based on the geometry of the one or more objects.

[0098] The controller estimates the pre-drilled ground plane as a continuous segment in the model representation space based on the features extracted from the imaging data. This approach is based on an assumption that the ground plane beneath the drill tailings may be modelled as a continuous surface prior to that same ground being drilled to form the borehole. The pre-drilled ground plane estimate produced accordingto the proposed techniques is correctly aligned with the true ground plane of the borehole, irrespective of the occurrence of tailings in the post-drilled state, as opposed to using a projected machine plane to calculate the same. For example, the controller may perform semantic segmentation of imaging data to generate a continuous surface, or a corresponding approximation, of the ground plane from one or more detected features of the collar and / or the region surrounding the collar. Data conversion from local to global reference

[0099] The controller represents the continuous pre-drilled ground plane segment with ground plane data comprising values in a local reference frame of the sensor(s) (referred to as “local ground plane data”). For example, the local ground plane data may include respective data points each specifying a position and orientation by: a 3D cartesian coordinate vector; a roll offset; and a pitch offset. In some embodiments, the position values within the local ground plane data are expressed using other non- Cartesian coordinate systems (e.g., polar coordinates).

[0100] The local ground plane data is converted to ground plane data having corresponding position and orientation values in a global reference frame (referred to as “global ground plane data”). The conversion is achieved by determining a position and orientation of the one or more sensors 104 in the global reference frame 205 (referred to as the “sensor position and orientation”). The sensor position and orientation indicates the true global origin position and orientation for the sensor (local) reference frame 205’ that is used to generate imaging data from scanning performed at a given time. A transformation is applied to the local ground plane data based on the sensor position and orientation to produce the global ground plane data.

[0101] In some examples with a plurality of sensors 104, the sensor position and orientation is measured with respect to a single point (“detection point” 104’) that is representative of the set of the sensors 104 (e.g., a centre point of the sensors as arranged on the support member or vehicle). That is, the sensor position and orientation may provide a global position vector and a global orientation vector of the sensor set 104, thereby defining the local measurement frame used by the detection system 105 togenerate the imaging data relative to the detection point. The sensor position and orientation is determined by a localization and pose system associated with the one or more sensors 104, and a spatial arrangement of the detection system relative to the localization and pose system.

[0102] In some embodiments, the localization and pose system is coupled to a deployment system 16, such as a vehicle or structure configured to deploy the detection system to perform the scanning of the collar region. The spatial arrangement therefore describes a relative relationship between the sensors(s) 104 and portion of the deployment system 16, and is either: (i) fixed, where the sensors are mounted to the deployment system 16 itself or (ii) moveable, where the sensors are mounted to a portion of the deployment system 16, or another member such as a support member 110, that changes its position and / or orientation over time, relative to a base portion of the deployment system 16 to which the localization and pose system is coupled.

[0103] The form and representation of the spatial arrangement varies according to the configuration of the one or more sensors 104 relative to the deployment system 16. For a fixed relative relationship, the spatial arrangement may be a function indicating a position and orientation offset of the one or more sensors to a part of the localization and pose system (e.g., an antenna). For a moveable relative relationship, the spatial arrangement is a function of time and accounts for changes in the position and / or orientation of the one or more sensors with respect to the localization and pose system. For example, the deployment system 16 may include a kinematics linkage system configured to dynamically track the position and orientation of the one or more sensors relative to a point on the deployment system 16.

[0104] Use of a spatial arrangement function in combination with position and orientation measurements provided by the localization and pose system enables the controller to determine the sensor position and orientation at each instant in which imaging data is generated, irrespective of any movement of the sensor(s) relative to the deployment vehicle 16, and / or any movement of the vehicle itself (i.e., between or during the scanning operations). The local ground plane produced from processing theimaging data can then be converted to the global reference frame 205 to produce the desired (pre-drilled) ground plane estimate. Applications to logging

[0105] In some embodiments, the borehole collar locating devices and apparatus proposed herein are also configured to enable logging of the borehole 101, such as for example by supporting and / or controlling a measurement probe 102 configured to generate geological data in response to its deployment into the borehole 101. For example, the one or more sensors 104 may be coupled to a support member 110 configured to support the sensor(s) 104 and to optionally house a measurement probe for subsequent logging of the borehole 101.

[0106] An exemplary practical use of the proposed approach is to enable the determination and / or correction of depth values associated with geological data generated by a logging of the borehole 101. For example, the determined collar RL may be provided to a computing device that is configured to receive geological data generated by a measurement probe 102 in response to deployment of the measurement probe 102 into the borehole 101 via the collar. The computing device processes the determined collar RL to generate depth values in a subsequently performed logging activity, or to adjust or correct depth values generated by earlier logging activities, given the determined collar RL.

[0107] The proposed approach to locating the collar of a borehole thereby advantageously enables depth values generated during a post-drilling logging of the borehole to be aligned with corresponding depth values determined during drilling. This facilitates the integration of the two independently obtained data sets, thereby promoting improved accuracy in data analysis activities. Borehole collar locating apparatus

[0108] Fig.2a illustrates an exemplary collar locating apparatus 100 according to the described embodiments. The apparatus 100 comprises one or more sensors 104, and acontrol system 120. Optionally, the apparatus comprises a support member 110 connected between the control system 120 and the sensor(s) 104. In some embodiments, the support member 110 is configured to hold the one or more sensors 104 in an adjustable position and orientation around the region surrounding the collar 107 of the borehole 101. Optionally, the collar locating apparatus 100 includes a measurement probe 102 used to perform logging of the borehole 101 with the determined reduced level, or a location, of the collar 107. Sensor devices and detection system

[0109] The one or more sensors 104 include one or more remote sensing or visualisation components, such as individual sensors 104a, 104b, 104c, configured to detect the physical characteristics of the collar region by measuring its reflected and emitted radiation at a distance. In some embodiments, sensor set 104 is configured with a field of view 106 (indicated in dotted lines in Fig.2a) that extends substantially over the collar region, including the collar 107, and at least part of tailings 15 and outer surface 19. The field of view 106 is three-dimensional and includes one or more volumes above, below and / or near to the collar 107, as well as a surface area on surface 19. Fig.2a depicts a representation of the collar region extending between boundaries located at points C and C’ of a slice of the surface 19. In some examples, the field of view 106 is an aerial or elevated view that allows the sensor(s) to capture data representing at least the outer edges of the tailings pile and other areas near to, around and / or surrounding the collar 107 on the surface 19.

[0110] Although Fig.2a depicts a set of three sensors 104a, 104b, 104c it will be appreciated that any arbitrary number and type of sensor may be implemented within or on the apparatus 100 to generate the imaging data 230.

[0111] Although Fig.2a depicts the measurement probe 102 and support member 110 as being located outside of the field of view 106 of the one or more sensors 104, it will be appreciated that in some embodiments various other objects, such as the measurement probe 102, support member 110 and one or more other components of the deployment system, may be positioned within the field of view 106, such that theapparatus 100 is configured to generate imaging data 230 of both the collar region, and the one or more other objects.

[0112] In the described embodiments, the one or more sensors 104 are included within a collar detection system 105 that further comprises a scanning control unit 103. Scan control unit 103 is in electronic communication with each individual sensor 104a, 104b, 104c to perform one or more control operations associated with imaging the collar 107 including: activating one or more of the sensor(s) to perform a scan of the collar 107 and the region surrounding, in response to receiving a control signal; receiving, from one or more of the sensors 104a, 104b, 104c, scanning data in response to performing a scan of the collar 107; and processing the scanning data received from the sensors 104a, 104b, 104c to generate imaging data of the collar region. Scan control unit 103 is configured to generate imaging data 230 as local data, including positions and orientations of objects, entities, or points in the collar region, and the field of view 106 generally, with respect to a local reference frame 205’ of the sensors 104.

[0113] In some embodiments, the scan control unit 103 is configured as an embedded system with a processor and memory implemented as an integrated microcontroller with a RISC architecture, and the sensors 104a-104c are configured as peripheral devices providing data to, and receiving control data from, the microcontroller. In other embodiments, the scan control unit 103 may be implemented as one or more full-scale computer systems, such as an Intel Architecture computer system. In other embodiments, the one or more sensors 104 may be integrated with the scan control unit 103, thereby enabling an exchange of data between operational modules of the scan control unit 103 and the sensors 104 via an internal controller bus or similar structure.

[0114] The scan control unit 103 is configured to communicate with the control system 120 via the communications network 150 which may include wireless and / or wired transmission media. The scan control unit 103 may include a modem or transceiver configured to access a communications module (e.g., a protocol stack) to perform a data transfer via network 150. In one embodiment a wired connection is established between the scan control unit 103 and the controller 222 of the controlsystem 120, such as via an Ethernet cable. For example, where the controller 222 is located on a deployment vehicle 16, the cable may be housed within a cable enclosure and passed through the deployment mechanism (not shown) or directly connected to the scan control unit 103 when said unit is mounted directly on the deployment vehicle. In other embodiments, the scan control unit 103 may have a network interface implementing the IEEE 802.xx family of networking protocols enabling the exchange of information wirelessly with the controller device 222 (e.g., over technologies such as Wi-Fi).

[0115] In some embodiments, the one or more sensors 104 of the detection system 105 are configured to be in direct communication with an external control device (i.e., external system 180 or controller 222 of the control system 120). In such embodiments, the external control device 180, 222 issues control signals directly to the sensor(s) 104 of the detection system 105 to perform the imaging control operations.

[0116] The one or more sensors 104 are configured to scan the collar region to generate scanning data that enables the creation of a profile or model of the characteristics of the collar 107 and its physical surroundings. Imaging data is generated from the scanning data to form a geospatial data source describing the collar region (e.g., the tailings 15, the surface 19, the collar 107, and / or any other areas near to, around and / or surrounding the collar), for example as a 3D model-based representation of the collar 107 of the borehole 101, and any physical features of the surface around the borehole 101. In some embodiments, the imaging data is generated from the scanning data capturing the collar region and one or more other objects in the field of view 106, such as one or more components of a deployment system configured to deploy the measurement probe into the borehole via the collar. In such embodiments, the imaging data may comprise a 3D model-based representation of the collar 107 of the borehole 101 together with the one or more other objects in the field of view 106.

[0117] For example, the one or more sensors 104 may include a LiDAR device comprising a laser configured to emit light onto the collar 107 and a reflector configured to collect measurements (scanning data) from the reflections. Each LiDARdevice produces a continuous series of point by point measurements including range, azimuth, and elevation. In some configurations, the data points are transformed into a point cloud representation of the scanned terrain, and optionally of one or more other objects. The scanning data is processed and used to create imaging data 230 that forms a 3D representation of the collar 107 and its surroundings, and the one or more other objects.

[0118] In some embodiments, position and orientation values of the imaging data model are constructed using an integrated Inertial Measurement Unit (IMU) and GNSS receiver of the scan control unit 103, which allows each measurement, or points in the resulting point cloud or other type of scanning data, to be georeferenced (i.e., thereby forming the geospatial model of the collar region 107). In some embodiments, the IMU and GNSS receiver units are coupled to the deployment system with additional measurement devices (e.g. encoders) coupled to the support member 110 and / or detection system 105.

[0119] In some embodiments, the imaging data 230 may be part of a digital terrain (DTM) and / or a digital elevation model (DEM) of the collar region. The DTM and / or DEM may be constructed from imaging data obtained from one or more scanning operations previously conducted on the collar region. Alternatively, or in addition, the DTM may be created from external datasets, such as high resolution digital surface models (DSMs) of the collar region, via the application of image processing techniques (e.g., to filter artefacts and other objects).

[0120] In some embodiments, the one or more sensors 104 include one or more radar devices configured to determine positions and orientations of one or more objects by emission and detection of electromagnetic (i.e., radio) waves. The radar devices may include at least one high-resolution, long-range sensor such as a 4D imaging radar device. Each 4D imaging radar device produces measurements including at least range, azimuth, and elevation data as per a LiDAR device. In some configurations, the use of 4D radar devices enables the detection system 105 to image the collar region with improved range and field of view compared to other sensing devices.

[0121] In some embodiments, the one or more sensors 104 include optical sensing elements and / or vision capture devices configured to generate representations of the collar region, and optionally of one or more other objects in the field of view 106, as image frames or video streams. For example, any one or more of the sensors 104a, 104b, 104c may include optical sensors, image sensors, and / or other camera-based devices that provide images and / or a sequence of video frames to the scan control unit 103 (i.e., as scanning data).

[0122] The camera-based devices may be configured with a plurality of separate image sensors enabling the generation of 3D images (e.g., stereo camera devices). For example, scan control unit 103 may be configured to generate a model of the collar region based on the generation of one or more disparity maps generated from pairs of stereo images of the scanning data. The measure of disparity between each data point in the imaging data 230, as calculated based on the horizontal displacement of a point’s projections between the images of the pair, is then used to determine a distance between the corresponding real-world object (e.g., a feature of the collar region) and the sensor(s) 104. Control system

[0123] With reference to Fig.2a, the control system 120 is configured with a localization and pose system 226 including components to receive a location and pose signal (e.g., an antenna, not shown), and components to process received location and orientation data. In one example, the localization and pose system 226 includes a global navigation satellite system (GNSS) and inertial navigation (IMU) system configured to determine an exact position and orientation of a portion of the deployment system (e.g., an antenna 14) in a global reference frame 205. In some embodiments, the localization and pose system also includes a kinematic linkage positioning and measurement system configured to determine relative positions and orientations of one or more components of the apparatus 100, including for example the deployment system 16 and / or the one or more sensors 104 (e.g., at the point 104’) inthe global reference frame 205 at all times, and during any movement of the sensor(s) 104 (e.g., via movement of the support member 110).

[0124] As shown in Fig.2a, the control system 120 further includes a controller 222 configured to receive data from the detection system 105, including imaging data 230 representing the collar region of the borehole 101 as generated from the scanning operation. The imaging data 230 may be received as a live data stream provided by the detection system 105 in real-time in response to one or more scanning operations performed on the collar region. The controller 222 processes the imaging data 230 to determine a reduced level ^^^^^^of the collar 107 from an estimate of the pre-drilled ground plane 24, as described below. In some embodiments, the controller 222 is configured to determine a position of the collar 107, such as a center position A or an estimate of A’ thereof.

[0125] In the described embodiments, controller 222 is implemented as a standalone computing device, and comprises a central system bus (not shown), a memory system 223, one or more processors 221, and communications module 224. The processor(s) 221 may be any microprocessor which performs the execution of sequences of machine instructions, and may have architectures consisting of a single or multiple processing cores such as, for example, a system having a 32- or 64-bit Advanced RISC Machine (ARM) architecture (e.g., ARMvx). The processor(s) 221 issues control signals to other device components via the system bus, and has direct access to at least some form of the memory system 223.

[0126] The memory system 223 provides internal media for the electrical storage of the machine instructions required to execute the user application. The memory system 223 may include random access memory (RAM), non-volatile memory (such as ROM or EPROM), cache memory and registers for fast access by the processor(s) 221, and high volume storage subsystems such as hard disk drives (HDDs), or solid state drives (SSDs).

[0127] The processes executed by the controller 222 are implemented as programming instructions of one or more software modules stored on non-volatile storage of the memory system 223. In some other embodiments, the processes may be executed by one or more dedicated hardware components, such as field programmable gate arrays (FPGAs) and / or application-specific integrated circuits (ASICs).

[0128] In the described embodiments, the one or more software modules include: a ground plane estimation module 213 which is configured to process imaging data 230 generated by the detection system 105 and generate ground plane data 232 providing local and global pre-drilled ground plane estimates 24’ and 24 of the collar; a frame transformation module 215 configured to transform a set of input positions and orientations from a first reference frame (e.g., local reference frame 205’ of the sensor(s) 104) to a second reference frame (e.g., global reference frame 205 of the control system 120); and a collar locator module 214 configured to determine collar location data indicating at least the collar reduced level from the ground plane data 232.

[0129] In some embodiments, the software modules further comprise a mode generator module 216 configured to set a mode of calculating the reduced level of the collar. In some examples, the mode generator module 216 is configured to switch the controller 222 between a mode that calculates the reduced level of the collar according to the techniques proposed herein (e.g., via method 300), and a mode that uses conventional ground plane RL determination (e.g., where the module 216 detects the absence of any significant degree of elevation (or depression) of the reference plane 22 of the deployment system 16, relative to the true ground plane 20).

[0130] Memory 223 may also include one or more general application programs providing methods, data structures or other software services that define data or perform functions as required by the controller 222 (e.g., an operating system). The data and instructions may reside in multiple parts of the memory system 223, including registers, cache, main memory, and high volume storage.

[0131] In some embodiments, the controller 222 comprises an I / O device interface (not shown) that provides functionality enabling the user to interact with the controller 222 via one or more I / O devices. In some embodiments, the device interface includes one or more onboard input devices such as a touchpad or touch screen enabling a user to interact with the controller 222. The I / O device interface also provides functionality for the controller 222 to instruct output peripherals, which may include displays, and audio devices.

[0132] In the some embodiments, the detection system 105 is connected to the controller 222 via a specialized I / O connector enabling the transfer of imaging data 230 (and / or scanning data) to the controller 222 in real-time, or substantially real-time. In some embodiments, the controller 222 is configured to store the received or otherwise obtained imaging data 230 values as a function of time in order to enable post- processing of the data. In other embodiments, the received imaging data values are only processed dynamically in real-time, for example by the invocation of the ground plane estimation module 213 with the received data and the subsequent invocation of the collar locator module 214.

[0133] Communications module 224 is a modem or transceiver device configured to enable the establishment of a logical connection between the controller 222 and other computing devices through a wireless or wired transmission media. For example, in some embodiments the controller 222 is configured to receive additional data representing prior developed models of the borehole collar 107 from the external system 180 via intermediate network 150. In some embodiments, the additional data received by the controller 222 includes imaging data, ground plane data, reference data, and / or data representing the collar position or an estimate of the same, as generated by the controller of a similar collar locating apparatus.

[0134] The controller 222 implements one or more service modules including a data storage and retrieval module (not shown) enabling data to be stored in, and retrieved from, a data store 208. In some embodiments, the data store 208 includes, for example, an SQL database and / or a file management system. In some embodiments, the datastore 208 is formed within the memory system 223 and includes data tables, or other structures, configured to store, for the borehole 101: imaging data 230, generated in response to scan(s) conducted by the detection system 105; ground plane data 232, including the local and global pre-drilled ground plane estimates 24’ and 24 of the collar; sensor arrangement data 234, including data representing the relative position and orientation of the spatial arrangement of the one or more sensors 104, and data representing the sensor point 104’; and reference data 236, including data indicating the position and orientation values, and other attributes such as the size and / or shape of one or more components, of a deployment system and sensor(s) 104 in global reference frame 205.

[0135] In some embodiments, the controller 222 is further configured to control logging operations using the probe 102 including: configuring the probe 102 for measurement of the borehole 101; instructing the deployment and / or extraction of the probe 102; receiving data representing live measurements (“measurement data”) of at least one geological parameter of the borehole 101 collected by the probe 102; evaluating the measurement data of the borehole 101; and logging the borehole 101 using at least the same. In some embodiments, the controller 222 is further configured to selectively control the mode of operation of the probe 102 used to generate measurement data, and corresponding logging data, for the borehole 101.

[0136] In such embodiments, the data store 208 stores logging data obtained from conducting measurement of the borehole 101 using probe 102. In some embodiments, the logging data includes one or more types of correction data associated with the measurement data, such as for example corrected depth values that account for the determined error, as generated by the controller 222.

[0137] The controller 222 may operate as an edge-processing device configured to receive data and signals, including imaging data, from the detection system 105, retrieve data from one or more remote devices, such as for example from devices of an external system 180, and transmit corresponding data and signals to or from one or more other devices (e.g., probe 102). In other embodiments, the functionality of thecontroller 222 to generate and process imaging data 230 may be performed by analogous components of the detection system 105. Generation of the ground plane data describing local and global pre-drilled ground plane estimates 24’ and 24 may be performed in real-time, or substantially real-time, by the controller 222 with the generation of the imaging data by the detection system 105 enabling a dynamic and ad- hoc assessment to be performed of the collar region of the borehole, during or shortly after scanning operations on the same.

[0138] The skilled person in the art will appreciate that many other embodiments may exist including variations in the hardware configuration of device 222, and the distribution of program data and instructions to execute the borehole logging methods described herein. Configuration of the detection system

[0139] With reference to Fig.2a, the one or more sensors 104 of the detection system 105 are configured in an arrangement to perform scanning of the region surrounding the collar 107. For example, the plurality of sensors 104a, 104b, 104c depicted in Fig. 2a may be arranged as a linear array or in a circular configuration about a common centre point. In some embodiments, the arrangement of the plurality of sensors describes a relative relationship of the sensors 104 with respect to a deployment system (not shown) configured to deploy the detection system 105 to perform the scanning of the collar region. The deployment system may comprise, for example, a vehicle, such as an automated, autonomous, or semi-autonomous ground vehicle (AGV), or a structure.

[0140] In some embodiments, at least the one or more sensors 104 of the detection system 105 are connected to a support member 110. Support member 110 is configured to support the one or more sensors 104 of the detection system 105 in an adjustable position and orientation around the region surrounding the collar 107 of the borehole 101 at least for the duration of a scanning operation conducted on the collar region. In some embodiments, the support member 110 is moveable relative to the deploymentsystem, or a portion thereof, such that the position and orientation of the one or more sensors 104 varies over time relative to the deployment system.

[0141] In some embodiments, the support member 110 is coupled to the deployment system via a mechanical interface which attaches to the connector 113. In some embodiments, the control system 120 includes one or more actuators (not shown) configured to move the support member 110 to a selected (and adjustable) position and orientation around the region surrounding the collar of the borehole 101. In other embodiments, the support member 110 is moved to the selected position and orientation by manual adjustment and without any mechanical or robotic operation being performed by the control system 120. In such embodiments, the connector 113 and mechanical interface are configured to maintain the support member 110 at the selected position and orientation at least during one or more scanning operations performed on the collar region.

[0142] Many different arrangements of the sensors 104 and support member 110 may be implemented according to various examples of the collar locating apparatus 100. Figs.2b-2d illustrate configurations of a support member 110 that is used with some exemplary collar locating apparatus 100. The shape, configuration, and properties of the support member 110 determine in part the possible configurations and arrangements of the sensor(s) 104 that are coupled to the member 110. In the depicted embodiments, the support member 110 is formed with a cylindrical or tubular shape. In other embodiments, the support member 110 may have an alternative shape, layout, or configuration. For example, support member 110 may be constructed as a rectangular, or similarly shaped, beam such that the one or more sensors 104 are arranged in a linear array at the first end 110’. In particular, the support member 110 may be formed from a beam or armature of a pre-existing vehicle or structure to which the sensors(s) 104 are coupled.

[0143] In some embodiments, additional components of the detection system 105, such as scan control unit 103, are optionally coupled to the support member 110 (e.g.,by attachment onto a surface of the member). In other embodiments, the one or more sensors 104 are located remotely to the other components of the detection system 105.

[0144] In the cylindrically shaped embodiment of Figs.2b-2d, the support member 110 has a longitudinal dimension D1 aligned with the cylindrical axis, and a non- longitudinal dimension D2 perpendicularly oriented to the same. The one or more sensors 104 are coupled, either removably or permanently, to the support member 110 at a first end 110’ of the support member 110. Circular surface 114 of the first end 110’ is faced in a direction towards the collar region to provide the sensor(s) 104 with a line of view of the collar region during a scanning operation. A second end 110’’ of the support member 110 includes a connector 113 which enables a connection between the support member 110 and the control system 120. The arrangement of the sensor(s) 104 at the first end 110’ determines a sensor origin position 104’ from which the positions and orientations of the imaging data are relative to.

[0145] Fig.2b shows an example of the support member 110 in which a single sensor 104a is coupled directly to the surface 114 of the support member 110. In one configuration, sensor 104a is located along axis 111 in the longitudinal dimension D1 such that the sensor origin position 104’ and sensor 104a are co-incident or are located in the center of the surface 114. In other configurations, the sensor 104a may be located non-centrally on the surface 114, or on the cylindrical surface 112 adjacent to the surface 114 (i.e., at an edge of the area defined by the surface 114). In these other configurations, the sensor origin position 104’ is determined at the fixed position of the single sensor 104a.

[0146] Fig.2c shows an example of the support member 110 in which the one or more sensors 104 includes four sensors 104a-104d arranged radially with uniform spacing around the edge of the surface 114 at first end 110’. Each sensor 104a-104d is offset from the center of surface 114 by a distance of D3, where sensor origin position 104’ resides at the center of surface 114. In the embodiment depicted by Fig.2c, the sensors 104a-104d are coupled to the cylindrical surface 112 at the edge of the surface 114. In other embodiments, the sensors 104a-104d may be offset from the edge ofsurface 114 by a fixed distance (e.g., to reduce the likelihood or severity of unintended contact between the sensors and other objects).

[0147] Fig.2d illustrates an embodiment in which the support member 110 of the collar locating apparatus 100 is formed with an internal volume configured to receive the probe 102. In the depicted examples, the cylindrical shape of the support member 110 corresponds to the shape of an outer housing of the probe 102, with the first end 110’ being open through which the probe 102 may be inserted into the internal volume of the member 110. The four sensors 104a-140d are positioned collectively in a fixed relative arrangement on the surface 112 at the open first end 110’. In some embodiments, longitudinal dimension D1 is such that the measurement probe 102 is sufficiently contained within the internal volume of the support member 110 to avoid any part of the probe 102 obstructing the line of view between the sensor(s) 104 and the collar region during scanning.

[0148] In such embodiments, the support member 110 functions as a removable protective shell or cover of the probe 102. This advantageously allows a logging platform to use the support member 110 to locate the collar A of the borehole according to the techniques described, and also to sheath the probe 102 between deployments to measure each borehole 101 (or other boreholes) on the site.

[0149] In some embodiments, the support member 110 is comprised of one or more rigid materials, such as metals, and / or a plastics (e.g., polypropylene, polystyrene, nylon, polycarbonate and methacrylate or similar compounds). The material composition of the support member 110 may be selected based at least in part on the functionality of the support member 110. For example, in some embodiments where the support member 110 functions as a removable protective shell or cover of measurement probe 102, the particular material composition may be determined to, for example, minimize electrical interference with components of the probe 102.

[0150] In some embodiments, the one or more sensors 104 of the detection system 105 are fixed or coupled directly to the deployment system. That is, the one or moresensors 104 are held in a predetermined position and orientation relative to a base portion of the deployment system (e.g., a chassis of the vehicle). This is advantageous in that the predetermined position and orientation of the one or more sensors 104 may be selected such that the generation of imaging data is not obstructed or disturbed by the deployment system and / or the measurement probe 102. Deployment systems

[0151] Fig.2e illustrates an exemplary collar locating apparatus 100 configured to determine a location of a collar of a borehole 101 drilled into surface 19 at a site, such as a mine site. As shown in Fig.2e, apparatus 100 includes a set of one or more sensors 104, and a control system 120 configured to control the operation of the one or more sensors 104, and the position and the orientation of the sensor(s) 104 in a region of space around the borehole collar. In this example, the one or more sensors 104 are coupled to a support member 110, which is in turn coupled to a deployment system 16. The one or more sensors 104 have a field of view 106 that extends to include the collar region C-C’ as a wide area around the collar 107 and tailings 15, and the outer surface 19, such as to enable the capture of data representing the same.

[0152] In some embodiments, the deployment system 16 further comprises a probe deployment system 115 having one or more components configured to deploy the measurement probe 102 into the borehole 101 via the collar. The one or more components of the probe deployment system 115 include a pivot connector 16a, a plurality of deployment arms 16b and a probe deployment device 16d. The probe deployment device 16d may be a wireline device configured to lower the measurement probe 102 (e.g., a device comprising a cable and a winch, or a similar device) through the support member 110 and into the borehole 10 during deployment. In some embodiments, the probe deployment system 115 comprises the support member 110.

[0153] In the embodiment depicted in Fig.2e, the deployment system is a vehicle 16 that is positioned near to the borehole 101. The vehicle 16 may be part of a drilling rig used to form the borehole 101, or another special purpose vehicle configured to position itself at position B on the surface 19 prior to operating the sensor(s) 104. Thevehicle 16 may comprise one or more electrical, electronic, mechanical, and / or robotic components that maneuverer the one or more sensors 104, via movement of the support member 110, and hold the sensors 104 at the sensor point 104’ for scanning of the collar region.

[0154] In the described embodiments, the control system 120 includes a controller 222, in the form of a computing device including one or more processors configured to execute computer readable instructions enabling the controller 222 to receive, process, generate, and transmit data for locating the collar of the borehole 101. The controller 222 is configured as an edge-processing device, coupled or affixed to the deployment vehicle 16, that controls the operation of the apparatus 100 to locate a collar of the borehole 101 based on input imaging data. The input imaging data is generated by the one or more sensors 104 that are configured to scan the collar region C-C’ of the borehole 101, including the area near to, around and / or surrounding the borehole 101 on the surface 19. For example, the sensor(s) 104 may generate imaging data 230 by taking a series of individual scans or a continuous scan with at least one corresponding field of view 106. Control system 120 operates the controller 222 to process the generated imaging data to produce an estimate of a pre-drilled ground plane 24 of the collar.

[0155] In some embodiments, scanning of the collar region C-C’ involves moving the sensor set 104, and the corresponding field of view 106, via movement of the support member 110. For example, the support member 110 may be moveably coupled to the vehicle 16 via the pivot connector 16a and deployment arms 16b. Alternatively, or in addition, movement of the support member 110 may be achieved by a corresponding movement of the deployment vehicle (i.e., from position B depicted) in the area of the borehole 101. In other examples, the support member 110 is connected to vehicle 16 with a means that holds the support member 110 stationary, thereby fixing the position and orientation of the set of sensors 104 relative to the body of the vehicle 16.

[0156] In some embodiments, the control system 120 is in communication with a detection system of the one or more sensors 104 and / or a logging apparatus (notshown). For example, the logging apparatus may be configured to facilitate the operation of a measurement probe 102 to take measurements of the borehole 101, in response to corresponding control signals from the control system 120. In the depicted embodiments, the measurement probe 102 is housed within the support member 110 to facilitate use of the apparatus to perform logging activities on the borehole 101 following determination of the location of the collar.

[0157] In the described embodiments, the control system 120 is implemented on, or integrated with, the deployment vehicle 16, for example, as a standalone computing device that performs computational operations for the vehicle 16 (i.e., as a “computing box” or “communications box” of the vehicle 16). In some embodiments, the control system 120 is detachable from the deployment vehicle 16 for implementation on, or coupling to, another like vehicle. In some embodiments, the controller 222 of the control system 120 communicates to one or more external systems such as for example computing systems that are remote to the collar locating apparatus 100. For example, external computing system 180 may be configured as a bench management system (BMS) to receive, store and process data associated with boreholes 101 or other boreholes within the bench within the surface 19.

[0158] Various other configurations exist in which the one or more sensors 104 are deployed about the borehole 101 for the purpose of scanning the collar region C-C’, thereby enabling the corresponding control system 120 to determine a location of the collar 107.

[0159] Fig.2f illustrates another exemplary collar locating apparatus 100’ including a control system 120 and a set of one or more sensors 104 that are fixed or attached to the deployment vehicle 16 without a support member. For example, the one or more sensors 104 may be permanently or non-permanently fastened to a panel 16c forming part of the chassis of vehicle 16. Both the control system 120 and the sensors 104 are local to the deployment vehicle 16, and a direct connection may exist between the same (e.g., for the transmission and reception of imaging data).

[0160] Movement of the vehicle 16 in the area around the borehole 101 on surface 19 results in movement of the sensors 104, and therefore a change in the sensor position and orientation (and corresponding local reference frame R) from which the imaging data 230 of the collar region C-C’ is generated. This configuration may be advantageous in applications where it is desired to configure the deployment vehicle 16 specifically to perform scanning of the collar region (i.e., without subsequent logging), as the absence of the support member 110 results in a vehicle 16 that is more compact, lighter, and more efficient to control in movement around the surface 19.

[0161] The amount of the region C-C’ captured by the field of view 106 may vary depending on the position of the deployment vehicle 16. In the example depicted in Fig.2f, the vehicle 16 is positioned at B on the tailings pile 15 such that the field of view 106 of the one or more sensors 104 extends over a sub-part of the collar region C- C’. In some examples, the deployment vehicle 16 is positioned such that the one or more sensors 104 have a field of view 106 that captures the entire region C-C’, or a different portion of the same. Other configurations of the deployment vehicle 16 may be used in accordance with the proposed techniques. For example, the deployment vehicle 16 may be an aerial vehicle (e.g., a drone) with the sensor(s) 104 attached to a portion of the vehicle 16 that faces towards the collar 107 when the vehicle 16 is in flight. This allows scanning of the collar region C-C’ with one or more fields of view 106 that capture a wide surface area encompassing the collar 107, tailings 15, and outer surface 19 of the region C-C’.

[0162] In some embodiments, the control system 120 is not coupled to a deployment vehicle 16, but is instead deployed at a fixed position on the surface 19. Fig.2g illustrates another exemplary collar locating apparatus 100’’ in which the control system 120 of apparatus 100 is coupled to, or contained within, a non-movable structure 17 (e.g., a base station or outpost) having a fixed position B on the surface 19. Sensor(s) 104 are configured to have a field of view 106 that extends to include the collar region C-C’ as a wide area around the collar 107, the tailings 15, and the outer surface 19. The control system 120 is integrated with, or deployed on the structure 17 thereby permitting the collar locating apparatus 100 to function without vehicle 16.

[0163] Irrespective of its form and structure, the control system 120 is configured to maintain a precise indication of its position and orientation within a global reference frame. In some embodiments, the vehicle 16 or structure 17 includes a receiver component 14 configured to receive location data from one or more transmitter devices (e.g., GNSS and IMU satellites). The receiver 14 is part of the localization and pose system 226 configured to process the location data in real-time to determine a position and orientation of the receiver component 14 (referred to as a deployment system position and orientation). In some embodiments, the deployment system and / or sensor positions and orientations are provided as Real-Time Kinematic GPS (RTKGPS) location values, such as 3D spatial co-ordinate values.

[0164] The control system 120 is also configured to maintain a precise indication of the global position and orientation of the sensor(s) 104, i.e., the local reference frame LRF 205’, by knowledge of the spatial arrangement of the one or more sensors 104 relative to the localization and pose system. For example, in the configuration of Fig.2f the controller 222 is programmed with the position and orientation offset of the sensor(s) 104 relative to the deployment vehicle 16. The offset remains constant over time, changing only when the attachment of the sensor(s) 104 to the vehicle 16 is altered. In the example of Fig.2e, the controller 222 monitors the spatial arrangement by tracking the movement of the support member 110 relative to the body of the vehicle 16. This enables the control system 120 to have updated knowledge of the sensor positions and orientations in response to the real-time kinematics of the support member 101 as may be experienced during the scanning of the collar region. Method of locating a borehole collar

[0165] Collar locating apparatus 100 depicted in Fig.2a operates to locate the collar 107 (i.e., determine collar position A, or an estimate A’ thereof) of the borehole 101 by: optionally, configuring sensor(s) 104 for scanning the collar region C-C’ of the borehole 101 in its post-drilling state, where the scanning involves capturing data by one or more sensors 104 having a field of view 106 encompassing the collar 107, tailings 15, and outer surface 19 of the collar region C-C’; placing sensor(s) 104 at aposition and orientation; operating the sensor(s) 104 to scan the collar region from the position and orientation; generating imaging data 230 from scanning data of the scan; receiving or otherwise obtaining the imaging data 230 at the controller 222 in response to the scan; processing the imaging data 230 to generate an estimate of a pre-drilled ground plane 24 of the collar; and determining an indication of a reduced level of the collar from the estimate of the pre-drilled ground plane 24. Sensors, positions and orientations

[0166] Further descriptions of example methods and techniques of locating a collar of a borehole using the collar locating apparatus 100 are provided below. The following terms are used in the further descriptions.

[0167] The sensor set 104 refers to a group of the one or more sensors 104 of the deployment system 105 that are active during a scanning operation. The sensor set 104 is configured to have a sensor point 104’ being a relative location in space that is representative of the sensor set 104 for the purpose of performing a scanning operation on the collar region. For example, the sensor point 104’ may be indicated as a fixed numerical offset relative to a known location of the one or more sensors 104 as attached to either a support member 110 or a portion of a deployment system 16.

[0168] The sensor position and orientation describes the local reference frame 205’ (LRF) used collectively by the detection system 105 to generate imaging data 230 using the sensor set 104. In some embodiments, the sensor position and orientation is quantified by respective sensor position orientation vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^. The sensor position and orientation is defined in global space and at the sensor point 104’. That is, for given a ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^, the detection system 105 generates images using the set of sensors 104 according to the corresponding LRF 205’ with an origin at the sensor point 104’, such that the generated images are local to that point. In some embodiments, the sensor position and orientation vectors, and therefore the LRF 205’, are time-varying in response to movement of the set of sensors 104 during the scanning of the collar region. In other examples, the sensor position and orientation vectors remain constant.

[0169] The deployment system position and orientation describes the global reference frame 205 (GRF) defined at a fixed known point on a portion or component of the deployment system 16. Deployment system position and orientation vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^provide an indication of the location of the system 16 in global space which is constantly updated over time by the controller 222. For example, in the control system 120 of Fig.2a the values of ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^may provide an indication of the GRF 205 at antenna 14 as determined by the localization and pose system 226.

[0170] The sensor spatial arrangement defines the relative positional and orientational relationship between the sensor set 104, taken at the sensor point 104’, and a fixed known point on the portion of the deployment system 16 providing the GRF 205 (e.g., a point of connection of antenna 14). In some embodiments, the sensor spatial arrangement is represented by relative position and orientation vectors ^^^^^^^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^^^^^^^.

[0171] As described below, the controller 222 advantageously performs continuous real-time tracking of the (global) deployment system vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^, and the relative position and orientation vectors ^^^^^^^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^^^^^^^, during the scanning operations. Controller 222 processes these time-varying values to determine respective sensor position and orientation vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^during scanning, and stores these values (e.g., in data store 208). This enables the controller 222 to maintain accurate global sensor position and orientations ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^, therefore corresponding local reference frames of image sets of the imaging data 230, irrespective of any movement of the set of sensors 104 and / or the deployment system 16.

[0172] As described below, the controller 222 uses the sensor position and orientation vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^, to apply a transformation ^^ to convert data points of the local pre-drilled ground plane segment 24’ to data representing the global pre-drilled ground plane segment 24. In this way, any positional or orientational difference of the deployment system 16 relative to the surface 19 can be accounted for by producing aglobally relevant pre-drilled reference plane 24 from which a position, or reduced level, of the collar 107 can be determined. Configuring a scanning operation

[0173] Fig.3a illustrates a flow diagram of a method 300 for locating the collar of borehole 101 as performed by the exemplary collar locating apparatus 100. At step 301, apparatus 100 is optionally configured for performing a scanning operation on the collar region. In some embodiments, configuration involves determining one or more of: (i) a number of scans comprising the scanning operation; (ii) a corresponding sensor position and orientation to conduct each scan; and (iii) a set of the sensor(s) 104 that are to be active during each scan (i.e., sensor set 104).

[0174] In some embodiments, the controller 222 determines the number of scans based on an instruction or direction from a user of the apparatus 100. In some embodiments, the number of scans performed in a scanning operation is determined dynamically based on the characteristics of the borehole, and / or the collar region.

[0175] In some embodiments, the controller 222 determines a set of sensor positions and orientations to conduct respective one or more scans of the scanning operation based on a location of the deployment system 16 on the surface 19, and an estimate of the location of the collar on the surface 19.

[0176] In some embodiments, the controller 222 selects one or more of the total available sensors 104a, 104b, 104c of the detection system 105 to produce scanning data during a scan of the collar region. The set of sensors 104 may comprise particular sensors 104a 104b 104c of the sensor set 104 as determined based on one or more criteria. The criteria may relate to the type of the sensor, the imaging data produced from output of the sensor, and / or one or more characteristics of the borehole 101, collar region C-C’, and / or surface 19.

[0177] In some embodiments, the sensor set 104 is determined based on user input provided to the controller 222 (e.g., via a peripheral device, or a data transfer) viacommunications module 224. Alternatively, or in addition, the controller 222 may be programmed to automatically activate or deactivate particular sensors, thereby forming the sensor set 104, in response to the occurrence of predetermined events. For example, individual sensors 104a, 104b, 104c of the sensor set 104 may be activated or deactivated for a collar scanning operation based on the occurrence of physical and / or environmental conditions of the collar region (e.g., the amount of light on the surface 19 near the collar 107, the presence of rain or other weather effects in proximity to the sensor set 104, etc.).

[0178] In some embodiments, configuration step 301 is omitted from method 300. For example, the collar locating apparatus 100 may perform a scanning operation on the collar region with a default number of sensors (e.g., all sensors 104a 104b 104c) and from a predetermined set of sensor positions and orientations. Scanning the collar region

[0179] At step 302, the controller 222 receives imaging data of the region surrounding the collar 107, where the imaging data is generated by scanning the region with the one or more sensors 104. Fig.3b illustrates an exemplary scanning operation performed with a collar locating apparatus 100 comprising a support member 110. Prior to commencing the scanning operation, the deployment vehicle 16 is positioned at a location B on the surface 19 near to the collar 107.

[0180] In the example depicted by Fig.3b, the scanning operation involves performing at least four scans at respective time instants ^^0, ^^1^^2and ^^3. The respective scans generate scanning data associated with collar region C-C’ of the surface 19. The movement of the support member 110 and / or vehicle 16 causes the sensor point 104’ of the sensor set 104 to move in a region above the collar region during the scanning operation. Movement of the support member 110 may be continuous in time, with the scans being performed at predetermined intervals during the motion of the sensor set 104, and / or in response to the sensor set 104 having a position and orientation as predetermined by the controller 222.

[0181] In some embodiments, the field of view 106 of the sensor set 104 for each scan is configured to comprise all or part of the collar region C-C’ such that a wide area around the collar 107, the tailings 15, and the outer surface 19 is sensed during the scan. This enables the sensors 104 to capture data associated with the collar 107, the tailings 15, the outer surface 19 and any other relevant features within the region C-C’. In the example depicted by Fig.3b, the individual scans generate data of the collar region C-C’ with an overlapping sensor field of view 106 that encompasses the entire region C-C’. This enables the scanning process to incorporate redundancy in the imaging data subsequently generated (i.e., to assist with avoiding a failure to capture features that are relevant to the determination of the collar position A).

[0182] In other embodiments, the scanning operation is configured such that at least one individual scan involves the one or more sensors 104 having a field of view 106 that encompasses one or more of the collar 107, tailings 15, and outer surface 19 beyond the tailings of a sub-part of the collar region C-C’ in order to capture data of the same. In one example, each individual scan may involve the one or more sensors 104 having a field of view 106 that encompasses each of the collar 107, tailings 15, and outer surface 19 beyond the tailings of a sub-part of the collar region C-C’. In another example, the field of views 106 of a plurality of individual scans each encompass one of, or any combination of two or more of, the collar 107, tailings 15, and / or outer surface 19 of the collar region C-C’, such that the scans collectively capture data of each of the same. This enables the controller 222 to combine data captured from respective scans that individually do not extend over the complete collar region (i.e., where the respective scan extends partially between C and C’ in the cross-sectional representation). At least in this way, by capturing data representing a plurality of sub- parts of the collar region C-C’ in respective scans, the controller 222 is able to represent the collar 107, tailings 15, and outer surface 19 of the entire region C-C’.

[0183] Individual scans are performed with the sensor set 104 to generate imaging data capturing the collar region with LRFs R0, R1, R2 and R3 each defined by sensorposition and orientation vector sets {(^^^^^^^^^^^^^^ , ^^^^^^^^^^^^^^)^^}. The controller 222 isconfigured to determine the sensor position and orientation vectors, containing 3Dcoordinate and roll and pitch angles respectively, in the global reference frame 205 at least for each point in time instant ^^0, ^^1, ^^2and ^^3. For example, by obtaining (i) thedeployment system vectors (^^^^^^^^^^^^^^, ^^^^^^^^^^^^^^) and (ii) the time-varying sensor spatialarrangement vectors (^^^^^^^^^^^^^^^^^^^^, ^^^^^^^^^^^^^^^^^^^^), the sensor position and orientation vectors^^^^^^^^^^^^^^) are accurately determined both in response to vehicle 16 remainingstationary, and for movement of the vehicle from the position B. In some embodiments, the controller 222 is configured to record the sensor position and orientation vector sets{(^^^^^^^^^^^^^^, ^^^^^^^^^^^^^^)^^} together with the received imaging data to facilitate thetransformation of associated local ground plane data to a global reference frame (i.e., at step 304).

[0184] Figs.3c-f illustrate an exemplary scanning operation performed with a collar locating apparatus 100 comprising a sensor set 104 that is affixed to the deployment vehicle 16 (i.e., without support member 110). The scanning operation involves capturing data of collar region C-C’, depicted at respective time instants ^^0, ^^1, ^^2and ^^3. In this example, the sensor set 104, having a sensor point 104’, is moved via movement of the vehicle 16 from position B0 (shown in Fig.3c), to position B1 (shown in Fig.3d), to position B2 (shown in Fig.3e), and then finally to position B3 (shown in Fig.3f). The individual scans depicted in Figs.3c-f each involve the one or more sensors 104 having a field of view 106 that encompasses one or more of the collar 107, tailings 15, and outer surface 19 beyond the tailings of a sub-part of the collar region C- C’. In other embodiments, the sensor field of view 106 for each scan encompasses the entire region C-C’.

[0185] Movement of the vehicle 16 may be continuous between positions B1 and B2 and / or between B2 and B3. In some embodiments, scanning is performed continuously (e.g., to generate scanning data at a predetermined period) as the vehicle 16 moves between points B1 and B3. Alternatively, the vehicle 16 may stop at or at any point between the respective positions. Movement of the vehicle 16 causes the sensor set 104 to assume respective sensor positions and orientations corresponding to LRFs R0, R1,R2 and R3 at times ^^0, ^^1, ^^2and ^^3. Scans are performed at least at each time instant ^^0, ^^1, ^^2, and ^^3, with the sensor set 104, as described for the example of Fig.3b.

[0186] In some embodiments, one or more sets of images ^^0, ^^1, ^^2, ^^3comprise the imaging data 230 generated by performing scans with the detection system 105 at times ^^0, ^^1, ^^2, ^^3and with respective sensor position orientation vectors ^^^^^^^^^^^^^^(^^) and ^^^^^^^^^^^^^^(^^) of the sensor set 104. The controller 222 resolves individual LRFs R0, R1, R2, R3 associated with each image set ^^0, ^^1, ^^2, ^^3to a ground plane LRF 205’ that is associated with the continuous pre-drilled ground plane segment 24’ determined from processing the image sets ^^0, ^^1, ^^2, ^^3(as imaging data) in accordance with the techniques described herein.

[0187] With reference to Fig.2a, scan control unit 103 is configured to initiate a scan of the collar region C-C’ by transmitting a control signal to the one or more sensors 104a, 104b and 104c that comprise the sensor set 104. In some embodiments, scan control unit 103 initiates a scan in response to an instruction from the control system 120.

[0188] Sensors 104a 104b 104c of the sensor set 104 are operated during a scan to produce output data representing one or more sensed parameters. The parameters and output data values produced by each sensor 104a 104b 104c may depend on the type of sensor and any sensor configuration data provided to the respective sensor by the scan control unit 103.

[0189] Scan control unit 103 is configured to receive output data generated by each sensor of the sensor set 104 performing the scan, and process the output data to generate imaging data to image the collar region and optionally one or more other objects in the field of view 106. In some embodiments, the output data provided to the scan control unit 103 as a real-time data stream is in a format specific to the type of the sensor 104a 104b 104c generating the same. Scan control unit 103 is configured to apply one or more data processing methods to generate imaging data from the individual sensor output data streams.

[0190] Each output data stream may be a form of image data specific to the corresponding sensor 104a 104b 104c (e.g., stereo image output from at least one stereo camera sensor). For example, the scan control unit 103 is configured to generate disparity images by calculating a difference in location of an object or artefact of the collar region in corresponding images produced by two sensors in the sensor set 104.

[0191] A scanning operation performed by the detection system 105 involves one or more scans initiated by the scan control unit 103. In some embodiments, the detection system 105 is controlled directly by the control system 120, such as by the transmission of real-time data and control signals between the controller 222 and the scan control unit 103. Alternatively, or in addition, the operation of the scan control unit 103 is controlled by a program uploaded to the unit 103 prior to the collar locating activity.

[0192] The controller 222 of the control system 120 is configured to receive imaging data 230 from the scan control unit 103 associated with scanning operations performed by the detection system 105 (i.e., at step 302 of Fig.3a). In some embodiments, the scan control unit 103 transmits the imaging data 230 to the controller 222 in real-time as the imaging data 230 is generated (i.e., during or shortly after each individual scan). In other examples, the scan control unit 103 is configured to buffer the imaging data 230 within the detection system 105 and to transmit the data at a predetermined time (e.g., once all individual scans of a scanning operation have completed). One or more processors 221 of the controller 222 are configured to process the imaging data 230 in preparation for determination of ground plane estimates and / or detection of one or more other objects. In some embodiments, the processing of the imaging data 230 includes associating the imaging data 230 with sensor position and orientation data, such as sensor and / or deployment reference data 236, and / or storing the imaging data 230 in data store 208 or memory 223. Generating pre-drilled ground plane estimates

[0193] With reference to Fig.3a, at step 304 the controller 222 processes the imaging data 230 of the scanned region surrounding the borehole collar 107 to generate one or more estimates of a pre-drilled ground plane 24 of the collar 107. To generate theground plane estimates 24, controller 222 invokes ground plane estimation module 213 with imaging data 230 associated with the collar region C-C’.

[0194] Fig.4a illustrates a method 400 for processing imaging data 230 based on the determination of continuous segments of the ground plane on which the collar 107 resides. At step 402, the controller 222 detects, from the imaging data 230, one or more features of the collar region C-C’. Then, at step 404, the controller 222 processes the one or more features to determine a continuous pre-drilled ground plane segment 24’ of the collar 107.

[0195] In some embodiments, the one or more features detected by the controller 222 comprise at least: one or more tailings features 420 (see Fig.4e) indicating a drill tailings pile 15 near to, around and / or at least partially surrounding the collar 107; and one or more ground plane features 430 indicating one or more portions of a post-drilled ground plane associated with the drill tailings pile 15. In some embodiments, the controller 222 detects one or more additional features 440 from the imaging data 230 such as foreign objects or artefacts in the vicinity of the collar 107.

[0196] In some embodiments, controller 222 is configured with one or more detection models indicating the expected characteristics of the tailings feature(s) 420 and the ground plane feature(s) 430 as represented in imaging data generated by scanning the collar region.

[0197] Fig.4b illustrates an example of a detection model 450 implemented as a neural network (also referred to as a “detection network”). Detection network 450 is executed on input imaging data (I) 230 to produce classified image data (CI) 460, as shown in Fig.4b. A mapping ^^(^^) of the network 450 is defined by parameters or weights 455 of the one or more layers of the neural network. Network 450 produces an output 460 image by generating values of weights of output nodes 458 corresponding to the input image data 230. In some embodiments, the classified output image data (CI) 460 represents a classified or labelled version of the input image data 230.

[0198] In some embodiments, the detection network 450 is a deep neural network such as a multilayer perceptron (MLP), a recurrent neural network (RNN), or a convolutional neural network (CNN). In some embodiments, the neural network is configured with a single input layer 456 and a single output layer 458, and with multiple hidden layers 457. Controller 222 invokes the ground plane estimation module 213 to detect the one or more features 420, 430, 440 from the received imaging data 230 using the detection network 450 (see Fig.4e).

[0199] Detection network 450 is trained on a set of training data specific to the one or more features to be detected from the input image data 230. Training of the detection network 450 may be performed with, for example, a set of tailings feature labels 451, ground plane feature labels 452 and feature labels indicating other features 453 within a set of training images. In some examples, other types of machine learning and / or pattern recognition models may be used as the detection model 450 to detect one or more features from the imaging data 230.

[0200] Alternatively, or in addition, the controller 222 may be configured to perform feature detection based on input received from a user of the control system 120. For example, a human operator may perform a manual inspection of the imaging data 230 and provide the control system 120 with input data identifying or flagging regions in the images that correspond to respective drill tailings and an associated ground plane. In some examples, the controller 222 is configured to adjust, modify, or otherwise correct classifications or detections of features made by the use of the detection model 450 based on user input.

[0201] Various forms of the classified image data 460 may be generated by the detection model 450 according to the techniques proposed. For example, the controller 222 may apply semantic segmentation to the imaging data 230 to generate the classified image data 460. In some examples, the imaging data 230 is three dimensional and performing semantic segmentation on the imaging data 230 involves: performing a first segmentation step in two dimensions of the imaging data 230; and then mapping the classified / segmented pixels identified from the first segmentation step to acorresponding measured depth value of the data 230 (i.e., to obtain an estimate of the ground plane in 3D space). In other examples, the imaging data 230 is two dimensional such that the controller 222 is configured to apply semantic segmentation to the imaging data values directly and to combine the segmented pixels with one or more depth values provided externally to the imaging data 230 to generate a 3D ground plane estimate.

[0202] Fig.4c illustrates a method 410 executed by the controller 222 to perform semantic segmentation on the imaging data 230 based on the one or more features of the collar region C-C’.

[0203] At step 412, the ground plane estimation module 213 is invoked by the controller 222 with input imaging data 230 obtained from a scanning operation (e.g., as described by Figs.3b-3e) on the collar region. In some examples, the ground plane estimation module 213 performs one or more pre-processing operations (i.e., at step 414) on the imaging data 230 to facilitate the segmentation. For example, the pre- processing operations performed by the ground plane estimation module 213 may include one or more of thresholding, filtering, smoothing, subtraction, cropping, or any other image enhancement process, as applied to part or all of the imaging data 230.

[0204] At step 416, the ground plane estimation module 213 performs semantic segmentation on the pre-processed data. In some embodiments, the ground plane estimation module 213 is configured to use at least one detection network 450 to perform the segmentation on images represented by the imaging data 230. For example, the detection network 450 may be a convolutional neural network (CNN) configured with a convolutional layer, a pooling layer, and a non-linear activation function to form a convolutional encoder-decoder. In other examples the detection network 450 may be a U-net, a Pyramid Scene Parsing Network (PSPNet), or any other CNN-based network.

[0205] In some examples, the ground plane estimation module 213 is configured to perform one or more enhancement and / or optimization processes to improve theparameters or variables of the detection network 450, and / or any other models, used to perform the segmentation. For example, the ground plane estimation module 213 may perform parameterization on detection network 450 to optimize and / or enhance the weights 455 such that the model 450 is able to more accurately identify and locate objects within an image of the imaging data 230. Adaptation of the weights 455 of the neural network layers used in the model 450 may be made with respect to aspects such as the size, position, and shape of bounding boxes around objects.

[0206] In some embodiments, the ground plane estimation module 213 performs a parameterization process to fine-tune existing or pre-defined variables of the detection model 450, such as values initialized during a training phase. The parameterization process enables the detection model 450 to better learn and generalize patterns that represent various object classes. In this sense, parameterization is used by the ground plane estimation module 213 to actively adapt the model 450 to detect objects with known properties or characteristics. For example, a tailings pile feature may be classified by detecting objects with one or more of: a conical shape, an angle of repose between some set range, a height between some set range and a localization residing in a mostly flat area of the imaged region.

[0207] Additionally, the parameterization process may improve the precision with which objects associated with the borehole collar, and / or the collar region, are identified during detection tasks. In some configurations, the ground plane estimation module 213 adapts the parameters 455 of the network 450 to recognize diversity in object characteristics and environmental conditions. For example, the variables of detection network 450 may be parameterized to improve the identification of one or more features within the imaging data 230 captured with new or unseen conditions compared to the training data set (e.g., the existence of particular weather effects, such as rain or mist, present during scanning).

[0208] The CNN 450 extracts one or more features, including tailing 420 and ground plane 430 features, from the imaging data 230 before using them to separate the represented images into multiple segments. For each image of the image data 230 theground plane estimation module 213 creates a segmentation mask in which image pixels are labeled according to categories. Exemplary categories include ‘collar’ (value 0), ‘tailings’ (value 1), ‘ground plane’ (value 3) and ‘other’ (value 4). This enables the ground plane estimation module 213 to output (i.e., at step 416) classified image data 460 in the form of a semantic image map. The segmented image map transforms each pixel value, from 0 to 255, within images of the input image data 230 into a class label with a corresponding integer value (e.g., a value of 0,1,2 or 3).

[0209] Fig.4d illustrates classified 405 image data of the collar region C-C’ generated by processing imaging data 230 with a semantic segmentation method 410. The classified image data 405 depicted in Fig.4d is a two dimensional representation of 3D imaging data produced from outputs of sensor(s) 104 having an elevated position above the region C-C’. This results in a corresponding field of view 106 that captures the edges of the tailings and the ground 19 surrounding the collar in the region C-C’. The classified image data 405 is a segmented map that identifies collar feature 408, surrounded by a drill tailings pile 407, and the outer surface 409.

[0210] Alternatively, or in addition, the controller 222 is configured to detect and extract the one or more features 420, 430, 440 from the imaging data 230 without the use of machine learning. For example, ground plane estimation module 213 may be configured to process the imaging data 230 by applying one or more techniques based on determining gradient distributions or edge orientations, or applying scale-invariant transformation descriptors to detect, describe, and match one or more features 420, 430, 440 within images of the imaging data 230.

[0211] In some images of the collar region, as represented by the imaging data 230, the pre-drilled ground plane on which the collar 107 resides may be difficult to identify, such as for example due to the presence of the drill tailing pile on the surface 19 which may obscure part or all of the collar 107. To improve the accuracy by which the pre- drilled ground plane 24’ is determined, the ground plane estimation module 213 is configured to identify portions of an estimate of the ground plane (“ground planeportions”), and to process the identified portions to form the continuous ground plane segment.

[0212] Fig.4e illustrates an example model 419 of the features extracted from images of the collar region C-C’. The features of model 418 may be obtained by using the imaging data 230 as input data to a detection model or network 450. For example, the ground plane estimation module 213 may execute semantic segmentation method 410 to extract ground plane features 430 that are separated into N portions 431- 43N (i.e., during the feature detection and extraction stage of step 404).

[0213] Fig.5a illustrates a method 500 for converting a ground plane portion, as detected from imaging data of the collar region, to a continuous pre-drilled ground plane segment using semantic segmentation. At step 502, the ground plane estimation module 213 determines one or more portions of the pre-drilled ground plane of the collar. For example, a post-drilled ground plane portion is identified by a feature extraction process (e.g., as part of the semantic segmentation of method 410).

[0214] At step 504, the ground plane estimation module 213 is configured to extend the identified selected portion of the post-drilled ground plane to approximate a continuous surface beneath the drill tailings pile 15.

[0215] Fig.5b illustrates the extension of a ground plane portion selected in step 504 to generate an approximated continuous surface 23 that continues beneath tailings 15. In some embodiments, the controller 222 is configured to generate a plurality of approximated continuous surfaces by extending the portions of respective identified post-drilled ground planes and selecting one of the plurality of approximated continuous surfaces as surface 23 that is processed to determine the continuous pre- drilled ground plane segment.

[0216] The ground plane estimation module 213 is configured to determine the continuous pre-drilled ground plane segment 24’ from the approximated continuous surface 23. In some examples, the controller 222 forms the ground plane segment 24’by applying a bounding function to geometrically limit the size of the continuous surface 23 to a set of known or estimated values. The controller 222 is configured to store at least the continuous pre-drilled ground plane segment 24’ as (local) ground plane data 232 in the data store 208 and / or memory 203. Local to global ground plane conversion

[0217] With reference to Fig.2a, the imaging data 230 is produced by detection system 105 when the sensor set 104 has one or more corresponding sensor position orientation vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^. As a result, the ground plane data 232 generated by processing the imaging data 230, for example by the semantic segmentation method 410, is local imaging data representing the continuous pre-drilled ground plane segment in a local reference frame 205’ of the detection system 105.

[0218] In some embodiments, the controller 222 is configured to transform the local ground plane data 232 to represent the (local) continuous pre-drilled ground plane segment 24’ in global reference frame 205.

[0219] Fig.6a illustrates a transformation 600 executed by the controller 222 to transform the local ground plane 24’ to a global ground plane 24. The controller 222 invokes the frame transformation module 215 with the local ground plane data 232 to generate global ground plane data, representing global ground plane estimate 24, by: (i) determining a position and an orientation of the one or more sensors in the global reference frame 205; and (ii) applying a transformation to positions and orientations of the local ground plane data, based on the position and orientation of the one or more sensors in the global reference frame.

[0220] At step (i), the controller 222 provides the frame transformation module 215 with position and orientation vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^, describing the origin of the local reference frame 205’ in global space (e.g., by extracting the vectors from the reference data 236). Then, at step (ii), the frame transformation module 215 constructs a transformation function ^^ to convert the local ground plane 24’ data points to corresponding data points of global ground plane 24.

[0221] With reference to the example depicted in Fig.6a, let local ground plane 24’be defined by data points with 3D cartesian coordinate vectors ^^′^^^^^^^^^^ = (^^’, ^^’, ^^’) androll and pitch offset vectors ^^′^^^^^^^^^^ = (^^′, ^^′) in local reference frame 205’. Frametransformation module 215 produces position and orientation vectors ^^^^^^^^^^^^, ^^^^^^^^^^^^of the continuous ground plane estimate 24 in global reference frame 205 by applying the transformation function ^^ that is specific to the associated local reference frame:

[0222] Controller 222 determines and continuously updates the sensor position and orientation vectors using: location updates received from localization and pose system 226, indicating the real-time position and orientation of the deployment system (stored as deployment reference data 236); and the sensor spatial arrangement data 234, which is continuously updated by a kinematics tracking system in examples where sensor set 104 is moveable relative to the deployment vehicle 16. In some embodiments, controller 222 is configured to store data representing the continuous ground plane estimate 24 in the ground plane data 232 of data store 208.

[0223] In some embodiments, the imaging data 230 is produced by detection system 105 in the global reference frame 205 (i.e., where the local and global reference frames are aligned). In response to detecting an alignment of the local and global reference frames, the controller 222 is configured to generate the global ground plane 24 directly from the local ground plane 24’ (i.e., without performing the transformation depicted in Fig.6a).

[0224] With reference to Fig.3a, at step 306 the controller 222 is configured to determine a reduced level of the borehole collar 107 from the true (global) pre-drilled ground plane estimate 24. In some examples, the controller 222 invokes the collar locator module 214 with ground plane data 232 representing global plane 24. The collar locator module 214 is configured to process the global ground plane estimate 24 to determine collar location data including a reduced level of the collar 107.

[0225] The reduced level of the collar 107 is determined from one or more coordinates of the ground plane estimate 24 in the global reference 205. Fig.6b illustrates the determination of the reduced level of a collar 107 using the ground plane estimate 24 generated by the collar locating apparatus 100. In this example, the reduced level ^^^^^^of collar 107 is determined as the value of the Cartesian z-coordinate of the pre-drilled ground plane estimate 24 in the global reference 205.

[0226] In some embodiments, the collar locator module 214 is configured to determine an estimated position A’ of the collar 107 in the global reference frame 205 from the pre drilled ground plane 24. The collar locator module 214 may be preconfigured to select a particular location as the estimated position of the collar 107, relative to the geometry of the collar 107 on the surface 19. In the example of Fig.6b, the collar locator module 214 is configured to estimate center position A of the collar 107. In other examples, the collar locator module 214 may estimate the collar location at another relative location such as on the boundary of the collar 107.

[0227] In one example, the collar locator module 214 determines the estimated collar position A’ by calculating an intersection point between a line vector projected from the detection system 105 and the estimated pre drilled ground plane 24. The line vector is projected from a reference point of the one or more sensors 104, such as sensor point 104’, and towards a target on the estimated pre drilled ground plane 24. The collar locator module 214 is preconfigured with a target determined from at least one feature of the imaging data 230 used to generate the ground plane estimate 24. For example, the target may be configured as one or more drill hole features visually indicating a center point of the collar (e.g., feature 406 of Fig.4d). Correcting logging data

[0228] With reference to Fig.3a, at step 308 the controller 222 is optionally configured to determine or correct logging data associated with the borehole 101 using the reduced level of the collar 107. In some embodiments, the logging data is generated by conducting a survey of the borehole 101 with measurement probe 102, or a similar device. Measurement probe 102 is configured to generate measurement data includinggeological data indicating values of at least one geological parameter, and corresponding depth data indicating depths at which the values of the at least one geological parameter are obtained. Determination of the reduced level of the collar 107 according to the described techniques advantageously enables the controller 222, or another computing device, to more accurately determine, and / or to correct or adjust, at least the depth data of logging data associated with the borehole 101.

[0229] Fig.7a illustrates a measurement probe 102 configured to measure a borehole 101 in response to the use of a collar locating apparatus 100 to locate the collar 107 of the borehole 101. Measurement probe 102 is configured to generate measurement data, including depth data, relative to a position M indicating a location of the measurement probe 102 (referred to as the “probe reference location”). The probe reference location M is a point in global space that is used as a reference for the generation of measurement data of the borehole 101 by the measurement probe 102. For example, the probe reference location M may be the position at which the probe 102 activates measurement sensors to measure the geological parameter values and / or corresponding depths of the borehole 101. Alternatively, the probe reference location M may be any reference position from which the location of the measurement probe 102 can be resolved at any point in time following the commencement of measurement of the borehole 101 using the probe 102. That is, the probe reference location M provides a global reference location from which the global position of the measurement probe 102 may be determined (e.g., for the purpose of determining depths of the measurement probe 102 relative to the borehole 101). For example, the probe reference location M may be a point on the collar locating apparatus 100 (e.g., on a bottom edge of the support member 110 of the probe deployment system 115, as shown in Figs.7a and 8d- e), which advantageously permits calculation of the global position of the measurement probe 102 based on a known offset between the location of the measurement probe 102 and the location of position M relative to the (known) location of the probe deployment system 115 (e.g., using data provided by a probe deployment device 16d as shown in Fig.2e, such as a wireline encoder). During a measurement activity, the measurement probe 102 is deployed from the position M through the collar 107 of the borehole 101 as geological data and corresponding depth data are generated.

[0230] Vertical displacement between the probe reference location M and the collar position A, or estimated position A’, results in errors in the depth data values generated during the measurement activity performed with probe 102.

[0231] Fig.7b illustrates a method 700 performed by the controller 222, or another computing device, for determining or correcting the depth data values generated by measurement probe 102 for borehole 101. At step 702, the controller 222 is configured to receive or determine the probe reference location of measurement probe 102. In some embodiments, the probe reference location of measurement probe 102 is at spatial position M that is predetermined and provided to the controller 222 prior to the borehole measurement activity (e.g., by an operator of control system 120, or from external system 180).

[0232] In some embodiments, the controller 222 determines the probe reference location of the measurement probe 102 dynamically at or just prior to the time of the measurement activity. For example, the controller 222 may be configured to determine the probe reference location of the measurement probe 102 by using position and / or orientation data provided by the kinematic linkage positioning and measurement system to locate and track the one or more components of a probe deployment system 115.

[0233] In some embodiments, the controller 222 may be configured to calculate a spatial position M of the probe 102 based on knowledge of a position of the support member 110 of the collar locating apparatus 100 from which the measurement probe 102 is deployed. In the example depicted by Fig.7a, the position M coincides with the sensor reference point 104’, which is tracked in real-time by the controller 222 in some configurations.

[0234] For example, with reference to Fig.2e the controller 222 may be configured to track the position and orientation of deployment arms 16b and support member 110, in real-time in response to a movement of the components of the probe deployment system 115, relative to the deployment system position and orientation. Using thelocalization and pose system 226 to determine and track the deployment system position and orientation (i.e., as 3D spatial co-ordinate values), the probe reference location (e.g., position M) can be determined in real-time (e.g., based on the controller 222 maintaining knowledge of any positional and / or orientational offset between the relative location of M with respect to the components of the probe deployment system 115).

[0235] In some embodiments, the probe reference location is determined by processing location data obtained from a locator device, such as a GNSS receiver within or coupled to the measurement probe 102. Measurement probe 102 may be configured to determine the spatial position M, and transmit the probe reference location M to the controller 222 via a wired or wireless communications network such as network 150.

[0236] In some embodiments, determining the probe reference location includes converting an absolute location, for example as specified by GPS latitude and longitude values, to a value representing the measurement probe 102 relative to the surface 19. The probe reference location may therefore have a coordinate representation that facilitates calculations of the depth of the probe 102 during measurement activities. For example, a Universal Transverse Mercator (UTM) conversion may be applied to convert GPS latitude and longitude values of a location on the curved surface of the Earth to flattened (i.e., planar) Cartesian UTM coordinate values. A depth offset value may be calculated based on the 3D Cartesian co-ordinate value of the position M and collar position A (or estimated position A’). In some embodiments, the Cartesian UTM coordinate positions are obtained by applying UTM conversion to other non-GPS coordinate values, such as those of a site-specific coordinate system (e.g., Cartesian grid coordinate values).

[0237] At step 704 the controller 222 uses the reduced level, and / or estimated position A’, of the collar 107 and the probe reference location to determine or adjust one or more depth values of the measurement data generated by the measurement probe 102. With reference to Fig.7a, controller 222 determines a depth offset value Δ^^^^as thevertical displacement between the probe reference location, given by spatial position M, and the reduced level of the collar 107: Δ^^^^ = ^^^^ – ^^^^^^

[0238] where ^^^^is the z-coordinate of the position M and ^^^^^^is the z-coordinate of the pre-drilled ground plane estimate 24, for representations using 3D Cartesian coordinate values.

[0239] In some embodiments, the controller 222 utilizes the depth offset value Δ^^^^to correct depth data against a difference between the probe reference location from which borehole measurements commence, and the location of collar 107 of the borehole 101, for example as described in International Patent Publication No. WO2023060313A1. In other examples, the controller 222 is configured to provide the depth offset value Δ^^^^to measurement probe 102, and / or an external system 180, to enable further measurement and / or logging of the borehole 101.

[0240] Advantageously, by using the collar locating apparatus 100 to determine the reduced level ^^^^^^, and / or a corresponding estimated collar position A’, the borehole 101 may be measured, and subsequently logged, with improved depth values compared to techniques that rely on prior knowledge of a collar location that be may subject to inaccuracy (e.g., as provided within the pre-drilling hole data). Determining the probe reference location

[0241] With reference to step 702 of method 700, in some embodiments the controller 222 is configured to determine the probe reference location (e.g., the reference position M) of the measurement probe 102 based on the data generated by the detection system 105 of the collar locating apparatus 100. For example, the controller 222 may be configured to determine both the estimated collar position A’ and the reference position M of the probe 102, in the global reference frame 205, by processing the imaging data generated by the sensor(s) 104. In such embodiments, the one or more sensors 104 are used to generate imaging data that provides a representation of the one or morecomponents of the probe deployment system 115, in addition to the physical characteristics of the collar region in the post-drilled state of the borehole 101, simultaneously in time.

[0242] In some embodiments, imaging data 230 of the type and form discussed above is generated from a scanning operation conducted when the probe deployment system 115 places the measurement probe 102 in a spatial position from which its deployment into the borehole 101 is to commence (e.g., at a position resulting from an alignment of the probe to the borehole). The controller 222 is configured to process the imaging data 230 to detect the probe deployment system 115, and to subsequently determine the probe reference location from the imaging data.

[0243] The proposed techniques for determining the probe reference location and the estimated collar location, or of a relative displacement between the locations, using the data generated by the detection system 105 provide improvements in the calculation of the depth offset value Δ^^^^. For example, determining the reference position M by processing the imaging data from scanning the region surrounding the collar and the probe deployment system 115 eliminates errors resulting from inaccuracies of the kinematic linkage and measurement system (e.g., due to miscalibration), and from a reliance on an assumption of the ideal functioning of the deployment components, which may exhibit non-ideal behaviour over time (e.g., due to joint pin wear causing slop in the joints, and other factors).

[0244] Further, in the proposed techniques the controller 222 is able to determine the probe reference location and the estimated collar location from imaging data generated within a single reference frame from sensor readings taken at the exact same time. This advantageously captures the relative relationship between the locations during the scanning operation, thereby avoiding errors in the depth offset that may result from independently determining the probe reference location from the collar location (e.g., when real-time updates of the deployment system position and orientation are not available for a kinematics based determination of the probe reference location).

[0245] Fig.8a illustrates the collar locating apparatus 100 of Fig.2a configured to further determine a probe reference location of a measurement probe 102 to log the borehole 101. The detection system 105 is configured with one or more sensors 104 having a field of view 106 that extends over the collar region C-C’ and also encompasses a probe deployment system 115 configured to deploy the measurement probe 102 into the borehole 101. In some embodiments, the controller 222 of the control system 120 is configured with modules that perform a detection of one or more objects in the field of view 106 (e.g., object detector 217), and that estimate a pose, including values of a position and orientation, of each detected object (e.g., object pose estimator 218).

[0246] In the embodiment shown in Fig.8a, the detection system 105 is located separately to the support member 110, for example to permit improved imaging of the support member 110 by the one or more sensors 104. For example, the detection system 105 may be coupled to a deployment vehicle or structure (not shown in Fig.8a) that is configured to place the probe deployment system 115 in view of the sensor(s) 104 during scanning.

[0247] Probe deployment system 115 comprises one or more components such as deployment arm(s) 16b, pivot connector(s) 16a, deployment mechanism 16d, and optionally support member 110 housing the measurement probe 102. In some embodiments, the field of view 106 captures all or a subset of the components of the probe deployment system 115 to determine the probe reference location (e.g., as a position M in global reference frame 205).

[0248] In various configurations of the apparatus 100, such as the example depicted by Fig.8a, the one or more sensors 104 of the detection system 105 are configured in an arrangement to perform scanning of: the region surrounding the collar 107; and one or more components of a probe deployment system 115, for example according to the field of view 106 of the sensor(s) 104. The generation of the imaging data 230 may proceed analogously to other example apparatus described above depending on the configuration of the scanning operations.

[0249] Fig.8b illustrates a flow diagram of a method 800 for locating the probe reference location of a measurement probe 102, as performed by the collar locating apparatus 100. At step 801, the apparatus 100 is optionally configured for performing a scanning operation to image the collar region and the one or more components of the probe deployment system 115.

[0250] Analogously to step 301 of method 300, in the configuration step 801 of method 800 the controller 222 may determine, for example, the number of scans, a set of sensor positions and orientations to conduct respective one or more scans, and / or one or more of the total available sensors 104a, 104b, 104c of the detection system 105 to produce scanning data. In configuring the scanning operation, the controller 222 may take into account a relative position and / or orientation of the one or more components of the probe deployment system 115 (e.g., to ensure that at least one component is imaged during the scanning to generate imaging data that is adequate to detect the probe deployment system 115 and subsequently determine the probe reference location M).

[0251] Fig.8c illustrates an example apparatus 100 with one or more sensors 104 including at least one 360 degree rotational Lidar sensor with a 90 degree vertical field of view (e.g., + / -45 degrees around horizontal). This advantageously allows a large section of the collar region C-C’ and the entire probe deployment system 115 to be captured in a single image frame.

[0252] At step 802, the controller 222 receives imaging data capturing the collar region C-C’ and one or more components of the probe deployment system 115, where the imaging data is generated by a scanning operation conducted with the one or more sensors 104. The imaging data also captures the region surrounding the collar 107 at the same respective imaging instants. Analogously to step 302 of method 300, in some embodiments the collar locating apparatus 100 is configured to perform a scanning operation in which each scan is configured to comprise all or part of the collar region C-C’, and to additionally comprise all or part of the probe deployment system 115.

[0253] In some embodiments, individual scans generate data of both the collar region C-C’ and the probe deployment system 115 with an overlapping sensor field of view 106 that encompasses the entire region C-C’ and the entire probe deployment system 115. In other embodiments, the scanning operation is configured such that at least one individual scan involves the one or more sensors 104 having a field of view 106 that encompasses one or more parts of the collar region C-C’, and one or more parts of the probe deployment system 115, in order to capture data of the same.

[0254] Figs.8d-e illustrate an exemplary scanning operation performed with a collar locating apparatus 100 configured to determine a collar location and a probe reference location for logging a borehole 101. The scanning operation involves capturing data of collar region C-C’, depicted at respective time instants ^^0and ^^1. In this example, the sensor set 104 is moved via movement of the vehicle 16 from position B0 (shown in Fig.8d) to position B1 (shown in Fig.8e). The individual scans depicted in Figs.8d-e each involve the one or more sensors 104 having a field of view 106 that extends to include part of the collar region C-C’ as a wide area around the collar 107 and tailings 15, and the outer surface 19, such as to enable the capture of data representing the same. The field of view 106 is three-dimensional and includes one or more volumes above, below and / or near to the collar 107, as well as a surface area on surface 19.

[0255] As shown in Figs.8d-e, the field of view 106 also includes one or more components of the probe deployment system 115. For example, the imaging data generated by the one or more sensors 104 captures one or more of the deployment arms 16b, the pivot connector 16a, deployment device 16d, and support member 110 housing the measurement probe 102. In some embodiments, the boundaries of the collar region C-C’ are determined by a maximum viewing distance of the one or more sensors 104 for the field of view 106. Reference position M of the measurement probe 102 and estimated collar position A are each defined by respective coordinates in the global reference frame 205.

[0256] Movement of the vehicle 16 may be continuous between positions B0 and B1. In some embodiments, scanning is performed continuously (e.g., to generate scanningdata at a predetermined period) as the vehicle 16 moves between points B0 and B1. Movement of the vehicle 16 causes the sensor set 104 to assume respective sensor positions and orientations corresponding to LRFs R0 and R1 at times ^^0and ^^1. Scans are performed at least at each time instant ^^0and ^^1, with the sensor(s) 104.

[0257] In some embodiments, the imaging data 230 comprises data generated by at least one scan performed when the probe deployment system 115 positions the measurement probe 102 at the probe reference location (e.g., when the probe deployment system 115 positions the probe 102 to be substantially vertically aligned with the collar 107, as shown in Fig.8e). For example, the scanning operation may be configured such that a final scan is performed to image the probe deployment system 115 in the spatial configuration from which deployment of the measurement probe 102 is to occur (e.g., from a position where the deployment vehicle is located as close as possible to the approximate location of the collar, such as position B1 of Fig.8e).

[0258] At step 804, the controller 222 processes the imaging data 230 to generate a detection of the probe deployment system 115. To perform an image-based detection of the probe deployment system 115, the controller 222 invokes the object detector 217 with input data comprising the imaging data 230.

[0259] In some embodiments, the object detector 217 uses machine learning to process the imaging data 230 to detect one or more features representing the probe deployment system 115. The object detector 217 extracts, from the reference data 236, model data indicating expected reference characteristics of the probe deployment system 115. In some embodiments, the model data comprises a set of parameters describing the expected reference characteristics of the probe deployment system 115. The expected reference characteristics are determinable from imaging the probe deployment system 115 and may include for example a shape or physical dimension or configuration of the probe deployment system 115.

[0260] In some embodiments, each deployment system detection model may be implemented as a detection network, analogously to the detection network 450, such asfor example as a deep neural network such as a multilayer perceptron (MLP), a recurrent neural network (RNN), or a convolutional neural network (CNN). Each deployment system detection model is trained on one or more features to be detected from the input imaging data 230. For example, training of a deployment system detection model may be performed with, for example, a set of visual features including one or more of: a geometric shape of one of more objects captured by the imaging data 230; edges and contours, as identified by the boundaries of the objects; corners and keypoints, as identified by detecting points where edges meet; texture, as determined by analyzing the surface patterns of objects to differentiate between similar shapes; size and scale and / or aspect ratio, as determined by comparing the width and height of objects to specific reference values; and / or symmetry, as determined by calculating measures of known symmetrical properties common in many geometric shapes.

[0261] The deployment system detection model is trained with a set of training images, for example using supervised learning to train a single CNN with a corresponding set of training labels (i.e., to map values of each visual feature to imaging data of the corresponding component(s) of the probe deployment system 115). In some examples, other types of machine learning and / or pattern recognition models may be used as the deployment system detection model to detect one or more features from the imaging data 230.

[0262] In response to obtaining the model data, the object detector 217 detects, from the imaging data, the probe deployment system 115 by determining one or more visual features within the images of the data. The object detector 217 executes the deployment system detection model with a set of input image frames extracted from the imaging data 230. The deployment system detection model produces an output indicating the presence or absence of the probe detection system 115 in the set of input image frames (e.g., by matching values of a shape feature in the image frame(s) of the imaging data 230 to a representative shape of at least one component of the probe deployment system 115).

[0263] For example, the object detector 217 may perform real-time object detection via a deep learning algorithm such as You Only Look Once (YOLO) using a CNN detection model. In some configurations, images of the imaging data 230 are each pre- processed for example to resize the image to a fixed dimension prior to input to the detection model. Object detector 217 generates detection data indicating one or more features of the probe deployment system 115 detected within the image frames. For example, the detection data may represent the detected features as bounding boxes of respective identified components of the probe deployment system 115, with the detection data specifying: a width and height of the bounding box relative to the image; a confidence score and / or accuracy of the detection; and local image coordinates of the bounding box.

[0264] At step 806, the controller 222 determines a probe reference location from the detection of the probe deployment system 115 within the imaging data 230. In some embodiments, the imaging data 230 comprises one or more sets of images generated by performing scans with the detection system 105 at corresponding times and with respective sensor position orientation vectors ^^^^^^^^^^^^^^(^^) and ^^^^^^^^^^^^^^(^^) of the sensor set 104. In some embodiments, the controller 222 is configured to determine the probe reference location by: generating, by processing the imaging data 230, detection data to identify one or more relevant components of the probe deployment system 115, each having local position and orientation data in a local reference frame 205’ of the detection system 105. In some embodiments, the controller 222 determines a component of the probe deployment system 115 as a relevant component if the determination of local position data of that component contributes to the determination of the probe reference location (e.g., according to a pre-configuration of the controller 222 to select particular surfaces or edges of the support member 110 on which the probe reference location is to reside).

[0265] In some embodiments, the controller 222 is configured to invoke the object pose estimator 218 to: (i) process the detection data and the imaging data to determine local position data representing at least a local position of the probe reference location from the identification of the one or more relevant components of the probedeployment system 115 (i.e., in the local reference frame 205’ of the one or more sensors 104); and (ii) transform the local position data of the probe reference location to determine a global position of the probe reference location (i.e., as reference position M for measurements obtained from deploying the measurement probe 102 into the borehole 101).

[0266] At step (i), the object pose estimator 218 generates position and orientation data for the identified one or more relevant components of the probe deployment system 115. The controller 222 provides the object pose estimator 218 with position and orientation vectors ^^^^^^^^^^^^^^and ^^^^^^^^^^^^^^, describing the origin of the local reference frame 205’ in global space (e.g., by extracting the vectors from the reference data 236). The object pose estimator 218 processes local image coordinates of the respective bounding boxes of the detection data to determine corresponding imaging data of the probe deployment system 115. For example, the object pose estimator 218 determines values of point cloud data corresponding to Lidar images of the probe deployment system 115 and processes the values to generate estimated position and orientation values of the identified components in the local reference frame 205’.

[0267] At step (ii), the object pose estimator 218 transforms position and orientation values of the identified one or more relevant components of the probe deployment system 115 from the local reference frame 205’ to global reference frame 205. In some embodiments, the object pose estimator 218 invokes frame transformation module 215 to perform the transformation analogously to the conversion of the local ground plane 24’ data points to corresponding data points of global ground plane 24 (as described above).

[0268] In some embodiments, the imaging data 230 is produced by detection system 105 in the global reference frame 205 (i.e., where the local and global reference frames are aligned). In response to detecting an alignment of the local and global reference frames, the object pose estimator 218 is configured to utilize the position and orientation values of the identified components of the probe deployment system 115, as generated from the imaging data, without performing further transformation.

[0269] In response to obtaining the local position and orientation values of the identified one or more relevant components of the probe deployment system 115, the controller 222 determines corresponding global position and orientation values of the identified one or more relevant components in the global reference frame 205. In some embodiments, the global position values may each be a spatial position expressed in a common coordinate set of the global reference frame 205. For example, the controller 222 may determine a global position in 3D cartesian coordinates (as shown in Fig.7a for global position M).

[0270] In some embodiments, the controller 222 is configured to determine the global probe reference location from the global position and orientation values of a predetermined position calculated from the identified one or more relevant components of the probe deployment system 115. For example, the predetermined position may reside on a bottom edge of the support member 110, such that the global probe reference location represents the location from which the measurement probe 102 exits the support member 110 when the measurement probe 102 is deployed to log the borehole 101. In some embodiments, the controller 222 is configured to determine the probe reference location from the (global) position and orientation values of the measurement probe 102 itself, such as in configurations where the object detector 217 identifies the measurement probe 102 as an object represented by the imaging data 230.

[0271] In the example of Fig.7a, the controller 222 is configured to set the probe reference position M to the global location representing the bottom edge of the support member 110. The controller 222 may subsequently determine a depth offset value Δ^^^^as depicted Fig.7a and described above. It will be appreciated that the controller 222 may be configured to use any other arbitrary predetermined relative position on image surfaces of the identified components of the probe deployment system 115 (e.g., a midpoint of the support member 110) to determine the global reference location.

[0272] In some embodiments, the controller 222 is configured to process imaging data 230 of the type and form discussed above to determine a relative displacement of the measurement probe 102 from an estimated collar position A’ (shown as alignedwith the true collar position A in Fig.7a), or a reduced level of the collar 107, without directly determining the probe reference location. For example, object detector 217 may be configured to operate with one or more detection models that are trained to identify a difference between probe reference and collar positions (i.e., by performing joint feature recognition of the respective components of the probe deployment system 115 and collar 107). Object pose estimator 218 may then be invoked with an input representing a detection of a probe deployment component-to-collar feature in one or more image frames, and generate an output indicating a corresponding displacement or distance between the probe deployment component and the collar (or its reduced level). The controller 222 may use the distance value, or the displacement value, as or to derive the depth offset value Δ^^^^in the method 700.

[0273] It will be appreciated by persons skilled in the art that numerous variations and / or modifications may be made to the above-described embodiments, without departing from the broad general scope of the present disclosure. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims

CLAIMS:

1. A method for determining a ground plane on which a collar of a borehole is located, the method comprising: receiving imaging data of a region surrounding the collar, the imaging data generated by scanning the region with one or more sensors; processing the imaging data to generate an estimate of a pre-drilled ground plane of the collar; and determining an indication of a reduced level of the collar from the estimate of the pre-drilled ground plane.

2. The method of claim 1, wherein the one or more sensors comprises at least one of: one or more LIDAR devices; one or more stereo cameras; and one or more radar devices.

3. The method of any of claims 1 or 2, wherein processing the imaging data comprises: detecting, from the imaging data, one or more features of the region surrounding the collar; and processing the one or more features to determine a continuous pre-drilled ground plane segment of the collar.

4. The method of claim 3, wherein the one or more features of the region surrounding the collar comprise at least: a tailings feature indicating a drill tailings pile at least partially surrounding the collar; anda ground plane feature indicating one or more portions of a post-drilled ground plane associated with the drill tailings pile.

5. The method of claim 4, wherein the one or more features are detected, from the imaging data, using a machine learning model.

6. The method of any of claims 4 to 5, wherein the continuous pre-drilled ground plane segment is determined by performing semantic segmentation on the imaging data based on the one or more features of the region surrounding the collar.

7. The method of claim 6, wherein performing semantic segmentation on the imaging data comprises: extending one of the one or more portions of the post-drilled ground plane to approximate a continuous surface beneath the drill tailings pile; and determining the continuous pre-drilled ground plane segment from the approximated continuous surface.

8. The method of any of claims 3 to 7, wherein processing the imaging data comprises: (i) generating local ground plane data representing the continuous pre-drilled ground plane segment in a local reference frame of the one or more sensors; and (ii) transforming the local ground plane data to represent the continuous pre- drilled ground plane segment in a global reference frame.

9. The method of claim 8, wherein transforming the local ground plane data comprises: determining a position and an orientation of the one or more sensors in the global reference frame; andapplying a transformation to positions and orientations of the local ground plane data, based on the position and an orientation of the one or more sensors in the global reference frame.

10. The method of claim 9, wherein the position and the orientation of the one or more sensors is determined using a localization and pose system associated with the one or more sensors, and a spatial arrangement of the one or more sensors relative to the localization and pose system.

11. The method of any of claims 8 to 10, wherein the local ground plane data comprises respective data points each having: a 3D cartesian coordinate vector; a roll offset; and a pitch offset.

12. The method of any of claims 1 to 11, the method further comprising: determining a position of the collar in the global reference frame from the estimated pre-drilled ground plane.

13. The method of claim 12, wherein the position of the collar is determined by calculating an intersection point between a line vector projected from the one or more sensors and the estimated pre drilled ground plane.

14. The method of claim 13, wherein the line vector is projected from a reference point of the one or more sensors towards a drill hole feature visually indicating a centre point of the collar, wherein the drill hole feature is determined from the imaging data.

15. The method of any of claims 1 to 14, further comprising: receiving or determining a probe reference location of a measurement probe to be deployed into the borehole via the collar; andusing the determined reduced level of the collar and the probe reference location to determine or adjust one or more depth values of measurement data generated by the measurement probe during its deployment.

16. The method of claim 15, wherein the imaging data is generated by scanning with the one or more sensors: the region surrounding the collar; and one or more components of a probe deployment system configured to deploy the measurement probe into the borehole via the collar, and wherein determining the probe reference location of the measurement probe comprises: processing the imaging data to detect the probe deployment system; and determining the probe reference location from the detection of the probe deployment system.

17. The method of claim 16, wherein determining the probe reference location further comprises: generating, from the imaging data, detection data identifying one or more relevant components of the probe deployment system; processing the detection data and the imaging data to determine local position data representing a local position of the probe reference location from the identification of the one or more relevant components of the probe deployment system; and transforming the local position data to determine a global position of the probe reference location as a reference position for measurements obtained from deploying the measurement probe into the borehole.

18. A device for determining a ground plane on which a collar of a borehole is located, the device comprising:a collar detection system having one or more sensors configured to scan a region surrounding the collar from a reference position and orientation; and a controller having one or more processors in communication with the collar detection system, the one or more processors configured to determine the ground plane of the collar by performing the method of any of claims 1 to 17.

19. An apparatus for determining a ground plane on which a collar of a borehole is located, the apparatus comprising: one or more sensors configured to scan a region surrounding the collar from a collar detection reference position and orientation; and one or more processors configured to: receive imaging data of the region surrounding the collar, the imaging data generated by the scanning of the region by the one or more sensors; process the imaging data to generate an estimate of a pre-drilled ground plane of the collar; and determine an indication of a reduced level of the collar from the estimate of the pre-drilled ground plane.

20. The apparatus of claim 19, wherein the one or more processors are further configured to: receive geological data generated by a measurement probe in response to deployment of the measurement probe into the borehole via the collar; and use the determined reduced level of the collar and a deployment reference location of the measurement probe to determine or adjust one or more depth values of the geological data.

21. The apparatus of claim 20, further comprising a support member configured to hold the one or more sensors in an adjustable position and orientation around the region surrounding the collar of the borehole.

22. The apparatus of claim 21, wherein the support member has an internal area configured to at least partially house the measurement probe, and an external surface configured to mount the one or more sensors.

23. The apparatus of any of claims 19 to 22, wherein the one or more sensors comprises at least one of: one or more LIDAR devices; one or more stereo cameras; and one or more radar devices.

24. The apparatus of any of claims 19 to 23, wherein processing the imaging data comprises: detecting, from the imaging data, one or more features of the region surrounding the collar; and processing the one or more features to determine a continuous pre-drilled ground plane segment of the collar.

25. The apparatus of claim 24, wherein the one or more features of the region surrounding the collar comprise at least: a tailings feature indicating a drill tailings pile at least partially surrounding the collar; and a ground plane feature indicating one or more portions of a post-drilled ground plane associated with the drill tailings pile.

26. The apparatus of claim 25, wherein the one or more features are detected, from the imaging data, using a machine learning model.

27. The apparatus of any of claims 25 to 26, wherein the continuous pre-drilled ground plane segment is determined by performing semantic segmentation on the imaging data based on the one or more features of the region surrounding the collar.

28. The apparatus of claim 27, wherein performing semantic segmentation on the imaging data comprises: extending one of the one or more portions of the post-drilled ground plane to approximate a continuous surface beneath the drill tailings pile; and determining the continuous pre-drilled ground plane segment from the approximated continuous surface.

29. The apparatus of any of claims 24 to 28, wherein processing the imaging data comprises: (i) generating local ground plane data representing the continuous pre-drilled ground plane segment in a local reference frame of the one or more sensors; and (ii) transforming the local ground plane data to represent the continuous pre- drilled ground plane segment in a global reference frame.

30. The apparatus of claim 29, wherein transforming the local ground plane data comprises: determining a position and an orientation of the one or more sensors in the global reference frame; and applying a transformation to positions and orientations of the local ground plane data, based on the position and orientation of the one or more sensors in the global reference frame.

31. The apparatus of claim 30, wherein the position and the orientation of the one or more sensors is determined using a localization and pose system associated with the one or more sensors, and a spatial arrangement of the one or more sensors relative to the localization and pose system.

32. The apparatus of any of claims 29 to 30, wherein the local ground plane data comprises respective data points each having: a 3D cartesian coordinate vector; a roll offset; and a pitch offset.

33. The apparatus of any of claims 19 to 32, wherein the one or more processors are further configured to determine a position of the collar in the global reference frame from the estimated pre drilled ground plane.

34. The apparatus of claim 33, wherein the position of the collar is determined by calculating an intersection point between a line vector projected from the one or more sensors and the estimated pre drilled ground plane.

35. The apparatus of claim 34, wherein the line vector is projected from a reference point of the one or more sensors towards a drill hole feature visually indicating a centre point of the collar, wherein the drill hole feature is determined from the imaging data.

36. The apparatus of any of claims 19 to 35, wherein the one or more processors are further configured to: receive or determine a probe reference location of a measurement probe deployed into the borehole via the collar; and use the determined reduced level of the collar and the probe reference location to determine or adjust one or more depth values of measurement data generated by the measurement probe during its deployment.

37. The apparatus of claim 36, wherein the imaging data is generated by scanning with the one or more sensors: the region surrounding the collar; and one or more components of a probe deployment system configured to deploy the measurement probe into the borehole via the collar, and wherein the one or more processors are further configured to determine the probe reference location by: processing the imaging data to detect the probe deployment system; and determining the probe reference location from the detection of the probe deployment system.

38. The apparatus of claim 37, wherein the one or more processors are further configured to determine the probe reference location by: generating, from the imaging data, detection data identifying one or more relevant components of the probe deployment system; processing the detection data and the imaging data to determine local position data representing a local position of the probe reference location from the identification of the one or more relevant components of the probe deployment system; and transforming the local position data to determine a global position of the probe reference location as a reference position for measurements obtained from deploying the measurement probe into the borehole.

Images