High contrast ultraviolet fluorescence imaging system and method for pixel-level detection of crude oil in drilling cuttings

Abstract

Systems and methods are provided for imaging drilling cuttings that employ a UV source that includes a UV LED configured to irradiate a sample volume with UV radiation that interacts with oil-containing cuttings to induce fluorescence emission. The camera system is configured to capture at least one image of the chip based on the fluorescence emission. In another aspect, there is provided a method for characterizing oil content in cuttings, the method involving capturing at least one WL image of a drilling cuttings irradiated by white light, capturing at least one UV image of the drilling cuttings based on fluorescence emission from UV radiation, processing the at least one WL image to determine a first pixel count for all cuttings, and determining a second pixel count for all cuttings based on the first pixel count. The at least one UV image is processed to determine a second pixel count of the oily chips, and a parameter representative of the oil content of the chips is determined based on the first pixel count and the second pixel count.

Description

[0001] Cross-reference paragraphs

[0002] This application claims the benefit of U.S. Provisional Application No. 63 / 365,444, filed on May 27, 2022, entitled “HIGH-CONTRAST ULTRAVIOLETFLUORESCENCE IMAGING SYSTEMS AND METHODS FOR PIXEL-LEVEL DETECTION OF CRUDEOIL IN DRILL CUTTINGS,” the disclosure of which is incorporated herein by reference. Technical Field

[0003] The present disclosure relates to systems and methods for acquiring surface measurements at a wellsite as part of a mud-logging campaign. Background Art

[0004] Mud-logging activities can employ fluid analysis to provide potential successful indications for locating oil-bearing formation rocks while drilling. Typically, mud-logging activities employ infrared (IR) spectroscopy measurements. However, due to the high absorptivity of crude oil, IR is limited by challenges associated with absorption band overlap and signal saturation. On the other hand, efforts to advance fluorescence-based methods have considered two performance criteria: high sensitivity and faster measurement. Unfortunately, UV fluorescence imaging, which has been used for the past sixty years to detect crude oil during mud-logging activities, does not score well on either criterion. The traditional detection system, known as a fluoroscope, provides an eyepiece window to observe the interaction of high-power UV radiation and drilling cuttings, but it has low robustness and bulky equipment. Fluoroscopes typically provide weak contrast images, which are of limited use for reliably applying modern machine learning tools to detect oil in drilling cuttings. In addition, the UV radiation emitted by the mercury-based bulb of the fluoroscope may pose a high health and safety risk to the operator. Therefore, fluoroscopes are not suitable for integration into automated mud-logging workflows at the well site. Summary of the invention

[0005] This Summary is provided to introduce a selection of concepts that are further described below in the Detailed Description. This Summary is not intended to identify key or essential features of the claimed subject matter, nor is it intended to be used as an aid in limiting the scope of the claimed subject matter.

[0006] In an embodiment, an imaging system for use with drilling cuttings is provided. The imaging system includes an ultraviolet (UV) source, the ultraviolet (UV) source including a UV light emitting diode. The UV source is configured to irradiate a sample volume with UV radiation, the UV radiation interacting with crude oil bound to the drilling cuttings located in the sample volume to cause fluorescent emission of photons in the visible region of the electromagnetic spectrum. The imaging system also includes a camera system, the camera system including an image sensor. The camera system is configured to capture at least one image of the drilling cuttings located in the sample volume based on the fluorescent emission from the drilling cuttings.

[0007] In an embodiment, a UV source illuminates the sample volume with UV radiation within at least one predetermined UV wavelength band that interacts with crude oil bound to drill cuttings located in the sample volume to induce fluorescent emission of photons in the visible region of the electromagnetic spectrum.

[0008] In an embodiment, the UV light emitting diode may be configured to emit UV radiation in a UV wavelength band between 265 nm and 400 nm, for example with a peak intensity at a wavelength of 365 nm.

[0009] In an embodiment, the UV source may further include an optical bandpass filter configured to transmit UV radiation within at least one predetermined UV wavelength band and block other UV wavelength bands. The at least one predetermined UV wavelength band may interact with the crude oil bound to the drilling cuttings located in the sample volume to cause fluorescent emission of photons in the visible region of the electromagnetic spectrum. In an embodiment, the optical bandpass filter may be configured to transmit UV radiation in a wavelength band between 265nm and 400nm, and substantially block the transmission of UV radiation of wavelengths outside the wavelength band for transmission. In an embodiment, the wavelength band for transmission of the optical bandpass filter extends over a narrow UV wavelength range of 100nm or less (i.e., the narrow UV wavelength range does not cover more than 100nm).

[0010] In an embodiment, the imaging system may further include an optical filter in the optical path between the sample volume and the camera system.

[0011] In an embodiment, the imaging system may further include a white light source configured to illuminate the sample volume with white light.The camera system may further be configured to capture at least one image of the drill cuttings located within the sample volume and illuminated by the white light.

[0012] In another aspect, the imaging system can be combined with a data processing system for use in a mud-logging workflow. The data processing system can be configured to process at least one image of drill cuttings located within the sample volume and illuminated by white light as captured by the imaging system, and further process at least one image generated by fluorescent emissions from drill cuttings located in the sample volume as captured by the imaging system to quantitatively characterize the oil content of the cuttings.

[0013] In an embodiment, the data processing system may be configured to:

[0014] i) processing at least one image of drill cuttings captured by the imaging system and located within the sample volume and illuminated by white light to determine a first pixel count of all cuttings;

[0015] ii) processing at least one image captured by the imaging system generated by fluorescent emissions from drill cuttings located in the sample volume to determine a second pixel count of oil-containing cuttings; and

[0016] iii) determining a parameter indicative of oil content of the drill cuttings based on the first pixel count and the second pixel count.

[0017] In another aspect, there is provided a method of characterizing oil content in cuttings, comprising:

[0018] capturing at least one image of the drill cuttings illuminated by white light;

[0019] capturing at least one image of the drill cuttings based on fluorescent emissions produced by exposure to UV radiation; and

[0020] Configuring a data processing system to perform operations including

[0021] i) processing at least one image of the drill cuttings illuminated by white light to determine a first pixel count of all the cuttings;

[0022] ii) processing at least one image of the drilling cuttings based on the fluorescence emission to determine the oil-containing cuttings;

[0023] A second pixel count of crumbs; and

[0024] iii) determining a parameter indicative of oil content of the drill cuttings based on the first pixel count and the second pixel count.

[0025] In an embodiment, the data processing system of the present disclosure may be configured to determine a parameter indicative of the oil content of the cuttings based on a ratio of the second pixel count to the first pixel count.

[0026] In an embodiment, the data processing system of the present disclosure may employ a first machine learning model to determine a first pixel count for all chips, and the data processing system may employ a second machine learning model to determine a second pixel count for oily chips.

[0027] In an embodiment, a first machine learning may be trained to determine a first pixel count of all cuttings from images of drilling cuttings of different textures, colors, and oils from different APIs and associated label data, and a second machine learning may be trained to determine a second pixel count of oil-containing cuttings from images of drilling cuttings of different textures, colors, and oils from different APIs and associated label data.

[0028] In an embodiment, a first machine learning model may be trained to determine a first pixel count of all cuttings from label data obtained from binarization of a drilling cuttings image illuminated by white light, and a second machine learning model may be trained to determine a second pixel count of oil-containing cuttings from label data obtained from binarization of a UV image illuminated by UV radiation, wherein the UV radiation causes fluorescent emission from the oil-containing cuttings.

[0029] In an embodiment, the first machine learning model may be a convolutional encoder-decoder neural network trained to determine a first pixel count of all drilling cuttings from an image of drilling cuttings illuminated by white light, and / or the second machine learning model may be a convolutional encoder-decoder neural network trained to determine a second pixel count of oil-containing cuttings from an image of drilling cuttings illuminated by UV radiation, wherein the UV radiation causes fluorescent emissions from the oil-containing cuttings.

[0030] In an embodiment, at least one image of drill cuttings illuminated by white light and at least one image of drill cuttings based on fluorescent emissions may be captured by an integrated WL / UV imaging system having a white light source and a UV source configured to illuminate a sample volume and a camera system configured to acquire images of drill cuttings located in the sample volume and illuminated by the white light source or the UV source without moving the drill cuttings.

[0031] In other embodiments, at least one image of drill cuttings illuminated by white light and at least one image of drill cuttings based on fluorescent emissions may be captured by separate and distinct white light imaging systems and UV imaging systems.

[0032] In an embodiment, an integrated WL / UV imaging system (or UV imaging system) may include a UV source and a camera system, the UV source including a UV light emitting diode, the camera system including an image sensor. The UV source may be configured to irradiate a sample volume with UV radiation, the UV radiation interacting with crude oil bound to drilling cuttings located in the sample volume to cause fluorescence emission of photons in the visible region of the electromagnetic spectrum. The imaging system also includes a camera system, the camera system including an image sensor. The camera system is configured to capture at least one image of drilling cuttings located in the sample volume based on the fluorescence emission from the drilling cuttings.

[0033] In an embodiment, the UV source may be configured to illuminate the sample volume with at least one predetermined UV wavelength band that interacts with crude oil bound to drill cuttings located in the sample volume to induce fluorescent emission of photons in the visible region of the electromagnetic spectrum.

[0034] In an embodiment, the UV light emitting diode may be configured to emit UV radiation in a UV wavelength band between 265 nm and 400 nm, for example with a peak intensity at a wavelength of 365 nm.

[0035] In an embodiment, the UV source may further include an optical bandpass filter configured to transmit UV radiation within at least one predetermined UV wavelength band and block other UV wavelength bands. The at least one predetermined UV wavelength band may interact with the crude oil bound to the drilling cuttings located in the sample volume to cause fluorescent emission of photons in the visible region of the electromagnetic spectrum. In an embodiment, the optical bandpass filter may be configured to transmit UV radiation in a wavelength band between 265nm and 400nm, and substantially block the transmission of UV radiation of wavelengths outside the wavelength band for transmission. In an embodiment, the wavelength band for transmission of the optical bandpass filter extends over a narrow UV wavelength range of 100nm or less (i.e., the narrow UV wavelength range does not cover more than 100nm).

[0036] In an embodiment, the UV source may further include an optical bandpass filter configured to transmit UV radiation within at least one predetermined UV wavelength band and block other UV wavelength bands. The at least one predetermined UV wavelength band may interact with the crude oil bound to the drilling cuttings located in the sample volume to induce fluorescent emission of photons in the visible region of the electromagnetic spectrum.

[0037] In an embodiment, the optical bandpass filter may be configured to transmit UV radiation in a wavelength band centered around 365 nm and substantially block transmission of UV radiation of wavelengths outside the wavelength band for transmission. In an embodiment, the wavelength band for transmission of the optical bandpass filter may extend over a narrow UV wavelength range of 100 nm or less (i.e., the narrow UV wavelength range does not cover more than 100 nm).

[0038] In an embodiment, the integrated WL / UV imaging system (or UV imaging system) may further include an optical filter in the optical path between the sample volume and the camera system. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a schematic diagram illustrating an exemplary wellsite embodying aspects of the present disclosure.

[0040] Figure 2A is a schematic diagram of an exemplary imaging system that may be used as part of a mud-logging workflow according to the present disclosure.

[0041] Figure 2B Depicted is a graph representing the fluorescence spectra of different crude oils in the visible region.

[0042] Figure 2C It means it can be used as Figure 2A A graph of the transmittance of an exemplary optical bandpass filter that is part of an imaging system.

[0043] Figure 3 Shown by Figure 2A An exemplary white light (WL) image of a drill cuttings sample captured by an imaging system.

[0044] Figure 4A Shown by Figure 2A The imaging system captures Figure 3 An exemplary ultraviolet (UV) image of the same sample of drill cuttings.

[0045] Figure 4B Shows Figure 4A A zoomed selection of an exemplary (UV) image of .

[0046] Figure 5 Shown by Figure 2A An exemplary white light (WL) image of another sample of drill cuttings captured by the imaging system of FIG.

[0047] Figure 6 Shown by Figure 2A The imaging system captures Figure 5 An exemplary ultraviolet (UV) image of the same sample of drill cuttings.

[0048] Figure 7 An exemplary white light (WL) image of yet another sample of drill cuttings captured by a white light imaging system is shown.

[0049] Figure 8 shows a photo captured by an imaging system that uses UV LEDs instead of fluorescent bulbs Figure 7 Exemplary images of the same sample of drilling cuttings.

[0050] Fig. 9 An exemplary white light (WL) image of yet another sample of drill cuttings captured by a white light imaging system is shown.

[0051] Fig.10 shows a UV fluorescent ring light captured by an imaging system Fig. 9 Exemplary images of the same sample of drilling cuttings.

[0052] Fig.11 An exemplary white light (WL) image of another sample of drill cuttings captured by a white light imaging system is shown.

[0053] Fig.12 shows the image captured by a digital microscope Fig.11 The digital microscope is configured with a UV light emitting diode, an optical bandpass filter, an optical filter and an image sensor, similar to Figure 2A part of the system.

[0054] Fig.13 is a schematic diagram of an exemplary workflow for quantitatively characterizing the oil content of chips using correlative imaging processing of WL images and UV images.

[0055] Fig.14 is used Figure 2A The imaging system follows Fig.13 Figure 3 is a flow chart of an exemplary experimental workflow of the general workflow.

[0056] Fig.15 An exemplary workflow for training a machine learning (ML) model to detect pixels of all chips in an arbitrary WL image (ie, pixels that are part of any chip in the WL image) is depicted.

[0057] Fig.16 Depicted is an exemplary workflow for training another machine learning (ML) model for detecting pixels containing oily chips in an arbitrary UV image (ie, pixels that are part of any oily chips in the UV image). DETAILED DESCRIPTION

[0058] Crude oil is a complex mixture of organic compounds, including saturates, aromatics, polar compounds, and asphaltenes. Most downhole spectral fluid measurements are expensive and cumbersome. Due to the presence of different amounts of gases, such as CO 2 ,CO,O 2 、N 2 , H 2 , CH 4etc., which are particularly challenging. The present disclosure is motivated by a long-standing need to obtain surface measurements at the well site as part of mud-logging activities. This capability can bring huge benefits by automating and optimizing the mud-logging workflow based on rapid measurements of drilling cuttings.

[0059] In addition, the present disclosure relies on the visible fluorescence of crude oil. When excited by ultraviolet (UV) radiation, visible fluorescence is triggered by the absorption of electron energy. Photons of UV radiation collide with electrons, promoting the excitation of electrons from the ground state level to a higher energy level (excited state). Subsequently, during deactivation, the electron energy is released back to the ground state by fluorescence. Unsaturated organic structures containing π electrons, such as aromatic hydrocarbons and conjugated polyenes, are the main causes of visible fluorescence in crude oil.

[0060] Figure 1 A wellsite system according to the present disclosure is shown in which drill cuttings from a hydrocarbon-bearing formation are collected and analyzed to characterize the oil content of the cuttings (and thus the oil content of the hydrocarbon-bearing formation corresponding to the drill cuttings). In the system, the wellsite is depicted as being on land (onshore). In other embodiments, the wellsite may be offshore (using a similar platform on the water). In the illustrative wellsite system, a borehole 111 is formed through a hydrocarbon-bearing underground formation 131 by rotary drilling in a well-known manner. A drill string 112 is suspended within the borehole 111 and has a bottom hole assembly 100, which includes a drill bit 105 at its lower end. The surface system includes a platform and a derrick assembly 110 positioned above the borehole 111. The assembly 110 includes a rotary table 116, a kelly 117, a hook 118, and a swivel 119. The drill string 112 is rotated by a rotary table 116, powered by means not shown, which engages a kelly 117 at the upper end of the drill string 112. The drill string 112 is suspended from a hook 118 by the kelly 817. A swivel 119 allows the drill string 112 to rotate relative to the hook 118. As is well known, a top drive system may be used instead.

[0061] exist Figure 1In an exemplary embodiment of the present invention, the surface system also includes a drilling fluid or mud 126 stored in a pit 127 formed at the well site. A pump 129 delivers the drilling fluid 126 to the interior of the drill string 112 via a port in the swivel 119, causing the drilling fluid to flow downward through the drill string 112, as indicated by directional arrows 108. The drilling fluid exits the drill string 112 via a port in or near the drill bit 105 and then circulates upward through the annular region between the exterior of the drill string 112 and the wall of the borehole 111, as indicated by directional arrows 109. In this well-known manner, when the drilling fluid 126 returns to the pit 127 for recirculation, the drilling fluid lubricates the drill bit 105 and carries drill cuttings (which are small pieces of rock that are broken away by the action of the drill bit 105) upward to the surface. The bottom hole assembly 100 of the illustrated embodiment may include modules for downhole measurements such as logging while drilling (LWD) modules 120, 120A and measurement while drilling (MWD) modules 130, a rotary steerable system and motor, and a drill bit 105. Directional drilling may also be used with embodiments of the subject disclosure.

[0062] According to some embodiments, drilling cuttings (also referred to herein as cuttings) can be extracted from the drilling fluid 126 returned to the surface (e.g., by a vibrator), cleaned and / or otherwise prepared for analysis, and placed in a sample tray ( Figure 1 The UV / white light imaging system 152 is configured to capture at least one image of the cuttings on the sample tray while illuminating the cuttings on the sample tray with white light. Such image(s) are Figure 1 The UV / white light imaging system 152 is labeled 154 and is referred to herein as a white light image or WL image. The UV / white light imaging system 152 is also configured to acquire at least one image 156 of the cuttings on the sample tray while irradiating the cuttings on the sample tray with ultraviolet (UV) radiation. UV radiation is a form of electromagnetic radiation having a wavelength of 10 nm to 400 nm, which is shorter than the wavelength of visible light but longer than X-rays. Such an image is Figure 1 The WL image 154 and the UV image 156 are both processed and interpreted by a data processing unit 160 in order to characterize the oil content of the cuttings on the sample tray (and therefore the oil content of the hydrocarbon-bearing subterranean formation 131 or a portion thereof from which they originate). The data processing unit 160 includes one or more central processing units 162, a storage system 164, a communications and input / output module 166, a user display 168, and a user input system 169. The storage system 164 may include a memory such as a semiconductor memory device (e.g., RAM, ROM, PROM, EEPROM, or flash programmable RAM), a magnetic storage device (e.g., a magnetic disk or fixed disk), an optical storage device (e.g., a CD-ROM), a PC card (e.g., a PCMCIA card), or another memory device.

[0063] According to some embodiments, the imaging system 152 and the data processing unit 160 may be located at the well site, such as a logging truck or some other location at the well site. In this case, the image acquisition and image processing operations as described herein may be performed in real time during the drilling process. Providing this type of rapid formation evaluation has potential benefits in completion decisions, for example, for optimal placement of perforations and hydraulic fractures. There may also be potential applications for geosteering of the drill bit 105.

[0064] In other embodiments, the data processing unit 160 and / or the imaging system 152 may be located at one or more locations away from the well site, such as at a remote laboratory. In the remote laboratory, additional and more accurate / sophisticated characterization methods may be employed to provide better accuracy and resolution of formation properties.

[0065] Figure 2A , an example embodiment of an imaging system 152 is shown, which includes a UV light source including a UV light emitting diode 201 and an optical bandpass filter 203, which cooperate to direct UV radiation within at least one predetermined UV wavelength band to a sample volume coincident with the field of view of a camera system 205. The optical bandpass filter 203 is positioned in an optical path between the UV light emitting diode 201 and the sample volume, and is configured to transmit UV radiation within at least one predetermined UV wavelength band while blocking other UV wavelength bands. The UV light emitting diode 201 emits UV radiation within one or more UV wavelength bands that include or correspond to the at least one predetermined UV wavelength band of the optical bandpass filter 203. The UV radiation emitted by the UV light emitting diode 201 is directed to the optical bandpass filter 203 for optical filtering and irradiating the sample volume with the UV radiation within the at least one predetermined UV wavelength band transmitted by the optical bandpass filter 203. The camera system 205 includes an image sensor (e.g., a CCD or CMOS image sensor) and a lens configured to capture an image of a sample 207 (i.e., cuttings on a sample tray) supported on a substrate 208 and located within a sample volume. An optical filter 209 may be positioned in an optical path between the sample volume (i.e., the location of the sample 207) and the camera system 205. UV radiation within at least one predetermined UV wavelength band illuminating the sample volume interacts with the drilling cuttings on the sample tray and excites electrons of fluorophores in the crude oil bound to the cuttings, resulting in emission primarily in the visible region of the electromagnetic spectrum, such as Figure 2BThe fluorescence spectrum of crude oil is generally composed of broadband radiation in the visible region. This broadband radiation in the visible region is caused by overlapping emissions from different fluorophores present in the sample. With the UV LED 201 turned on (and the white light source 211 as described below turned off), the camera system 205 can be configured to capture an image of visible photons from the sample fluorescence emission, which is referred to herein as a UV image (in Figure 1 156).

[0066] The imaging system 152 also includes a white light source 211 (also referred to as a broadband light source) that directs white light (i.e., broadband light within the visible spectrum) to a sample volume that coincides with the field of view of the camera system 205. In an embodiment, the white light source 211 may be an LED white light source. With the white light source 211 turned on (and the UV light emitting diode 201 turned off), the camera system 205 may be configured to capture an image of the sample 207 illuminated by the white light source 211, which is referred to herein as a white light image or WL image (in Figure 1 154).

[0067] The imaging system 152 may also include a light blocking shield 213 that may cover the system and block ambient light from passing therethrough, thereby eliminating spectral noise caused by ambient light that would otherwise adversely affect the imaging operations performed by the system. In other embodiments, the imaging system 152 may be operated in a dark room with minimal ambient light, and the light blocking shield 213 may be omitted.

[0068] In an embodiment, the optical filter 209 is optional, and its use may depend on the sensitivity of the imaging sensor of the camera system 205. In an embodiment, the optical filter 209 may be used as a bandpass filter configured to pass the fluorescence spectrum of crude oil in the visible spectrum so as not to affect the UV image acquisition of the camera system 205. In an embodiment, the bandpass filter 209 may be configured to block UV light from passing therethrough and entering the camera system 205.

[0069] In an embodiment, the UV light emitting diode 201 may be configured to emit UV radiation in a wavelength band between 265 nm and 400 nm. In an embodiment, the UV light emitting diode 201 may be configured to emit UV radiation in a wavelength band centered around 365 nm, with a peak intensity at a wavelength of 365 nm. The optical bandpass filter 203 may be configured to transmit UV radiation in a central pass UV wavelength band (e.g., a transmittance greater than 30% and possibly higher), and substantially block transmission of UV radiation of wavelengths outside the pass UV wavelength band for transmission (e.g., a transmittance less than 15% and possibly lower). In an embodiment, the pass UV wavelength band for transmission of the optical bandpass filter 203 may extend over a narrow UV wavelength range of 100 nm or less in the UV wavelength band of 265 nm to 400 nm (i.e., the narrow UV wavelength range does not cover more than 100 nm). In an embodiment, the optical bandpass filter 203 may provide a Figure 2C In this case, as shown in the figure, the narrow UV wavelength band for transmission of the optical bandpass filter 203 extends over a narrow UV wavelength range of 100 nm or less centered at 365 nm. In addition, the optical bandpass filter 203 substantially blocks the transmission of UV radiation of wavelengths outside the UV wavelength band for transmission.

[0070] UV LED 201 may also emit light at the tail end of the visible spectrum (e.g., visible light with wavelengths between 400 nm and 700 nm). Such light will produce highly variable spectral noise, which may negatively affect the imaging operations performed by the system. In an embodiment, optical bandpass filter 203 may be configured to block the transmission of such light in the tail end of the visible spectrum (e.g., visible light with wavelengths between 400 nm and 700 nm), such as Figure 2C This removes the spectral noise generated by this light and improves the contrast of the images of the cuttings acquired by the system.

[0071] In an embodiment, camera system 205 may include a model D850 camera with a macro lens commercially sold by Nikon Inc., Melville, NY.

[0072] In an embodiment, all components of the imaging system 152 may be purchased or acquired as off-the-shelf parts.

[0073] Figure 3 and Figure 4A Shown by Figure 2A An image of the cuttings on the sample tray is acquired by the imaging system 152. The cuttings consist of non-oil-bearing (or dry) rock pieces and oil-bearing rock pieces. Figure 3 An exemplary WL image 154 is shown, and Figure 4A An exemplary UV image 156 is shown. It should be noted that Figure 4A In the UV image 156 of FIG. 1 , the oily chips are bright and visible under UV illumination due to fluorescence from oil trapped in surface accessible pores of the chips, whereas the non-oiled chips are not bright and particularly visible in the UV image. Figure 4B yes Figure 4A A zoomed image of a selected portion of the UV image 156 is shown, highlighting visible features of the oily chips.

[0074] Figure 5 and Figure 6 Shown by Figure 2A An image of the cuttings on the sample tray is acquired by the imaging system 152. The cuttings consist of non-oil-bearing (or dry) rock pieces and oil-bearing rock pieces. Figure 5 An exemplary WL image 154 is shown, and Figure 6 An exemplary UV image 156 is shown. It should be noted that Figure 6 In the UV image 156 , the oily chips are bright and visible under UV illumination due to the fluorescence from the oil trapped in the surface accessible pores of the chips.

[0075] It is important to note that a mixture of different colored chips was used as a sample to test Figure 2A The color and reflectance variability of the imaging system 152. Importantly, in the UV image (see Figure 4A and Figure 6 ), the pixels of non-oil cuttings are dark, making non-oil cuttings invisible; the pixels of oil-containing cuttings are captured with bright luminescence. Also note that longer exposure times or higher UV powers only saturate the detector and do not degrade the contrast between cuttings with and without oil. This contrast is important for quickly identifying oil-containing cuttings in the mud-logging workflow.

[0076] Another option being considered is to replace the fluorescent bulbs of fluoroscopes with UV LEDs for mud-log oil detection. Figure 7 and Figure 8 The results of this option are shown. Figure 7 An exemplary WL image is shown, and Figure 8 shows an example image acquired using a UV LED. Note that by comparing Figure 6 UV images and Figure 8 Image, when using UV LED ( Figure 8 ), the contrast ratio between oily and non-oiled chips is Figure 6 The UV images obtained by the system of , which uses UV LEDs and optical passband filters to illuminate the sample, are much weaker.

[0077] Another option considered was to replace the bulky box of the fluoroscope with a small UV fluorescent ring light. In principle, this could allow white light and UV light imaging to be integrated into a unified system. Two limitations made this idea less practical: (i) the presence of mercury in the fluorescent ring light source, and (ii) low quality data. Fig. 9 and Fig.10 The results of this option are shown. Fig. 9 An exemplary WL image is shown, and Fig.10 An example image acquired using fluorescent ring light (e.g., 356 nm fluorescent ring light) is shown. Note that by comparing Figure 6 UV images and Fig.10 Image, using a UV fluorescent ring light ( Fig.10 ) creates a contrast between oil-containing and oil-free drilling cuttings that is greater than that obtained from Figure 6 The UV images obtained by the system (which uses UV LEDs and optical passband filters to illuminate the sample) are much weaker. This lack of contrast is a significant obstacle to making high-confidence predictions about oil based on modern machine learning methods (or even manual data processing).

[0078] In other embodiments, separate imaging systems may be used to acquire WL and UV images of the chips. For example, the imaging system may be configured with a WL source, an optical filter, and a camera / image sensor, similar to Figure 2A The system 152 is part of the system 152 and is used to capture WL images of chip samples. A digital microscope (such as the digital microscope commercially sold by Sanxo of Budapest, Hungary) can be configured with UV light emitting diodes, optical bandpass filters, optical filters and a camera / image sensor, similar to Figure 2A The digital microscope is a part of a system 152 and is used to capture UV images of the same chip sample. The digital microscope can be used with or without a light blocking shield. When used without a light blocking shield, any light in the microscope's surrounding environment can be turned off or dimmed, and the acquisition time of the imaging can be controlled, for example, to less than 1 second, in order to reduce environmental noise. Fig.11 and Fig.12 The results of this option are shown. Fig.11 An exemplary WL image is shown, and Fig.12 An exemplary UV image acquired with a digital microscope without a light blocking shield is shown. Note that by comparing Figure 6 UV images and Fig.12 In the UV images, the contrast between the drill cuttings with and without oil is very similar for both options.

[0079] In an embodiment, Figure 1The data processing unit 160 may be configured with imaging processing functionality (referred to herein as an "imaging toolbox") that processes and interprets the WL images and UV images acquired by the workflow in order to characterize the oil content of the cuttings and, therefore, the oil content of the hydrocarbon-bearing subterranean formation 131 or a portion thereof from which the cuttings originate. In an embodiment, the imaging toolbox utilizes the high-quality data of the UV images.

[0080] In an embodiment, the imaging toolbox can be configured to enable pixel-level detection of oil from chips, suitable for a fully automated workflow. The main goal of the imaging toolbox is to provide correlative image processing of WL images and UV images to quantitatively characterize the oil content of chips, such as Fig.13 As shown. The workflow includes the following steps: In step 1301, one or more WL images of the cuttings are obtained, whereby the cuttings are illuminated by white light. In step 1303, one or more UV images of the cuttings are obtained, whereby the cuttings are illuminated by UV radiation to produce visible fluorescence of crude oil trapped at the surface of the drilling cuttings, which results in oil-containing cuttings with high contrast. In step 1305, the WL image of 1301 and the UV image of 1303 are processed by correlation image processing of the WL and UV images to provide a quantitative assessment of the oil-containing cuttings. In an embodiment, the operation of 1305 may involve processing the WL image to determine the pixel count (i.e., the number of pixels) of all cuttings (both oil-containing cuttings and non-oil-containing cuttings), processing the UV image to determine the pixel count (i.e., the number of pixels) of the oil-containing cuttings, and determining a parameter representing the oil content of the cuttings based on the ratio of the pixel count of the oil-containing cuttings relative to the pixel count of all cuttings.

[0081] Fig.14 is used Figure 2A The imaging system 152 follows Fig.13 Figure 3 is a flow chart of an exemplary experimental workflow of the general workflow.

[0082] The workflow begins in block 1401 by obtaining a sample of cuttings. In an embodiment, the cuttings sample may be extracted from the drilling fluid returning to the surface (eg, by a shaker). Optionally, the cuttings may be cleaned, dried, and / or otherwise prepared for analysis.

[0083] In block 1403, the chip samples of block 1401 are loaded onto a sample tray. In an embodiment, the chips may be loaded onto the sample tray in a non-stacked configuration such that the chips are spatially distributed on the sample tray and are not located on top of each other.

[0084] In block 1405 , a WL image of the chip sample on the tray is acquired with the white light source 211 turned on (and the UV source 201 turned off).

[0085] In block 1407, the white light source 211 is turned off.

[0086] In block 1409 , a dark control image of the chips on the tray is acquired with the WL source 211 off and the UV source 201 off.

[0087] In block 1411 , a UV image of the chip sample on the tray is acquired with the UV source 201 turned on (and the WL source 211 turned off).

[0088] In block 1413 , the dark control image of 1409 may be used to subtract the contribution of ambient light from the UV image of 1411 .

[0089] In box 1415, the WL image of 1405 is input to a machine learning (ML) model that is trained to detect pixels of all chips in the WL image (i.e., these pixels are part of any chip in the WL image) and output a pixel count of all chips in the WL image.

[0090] In box 1417, the UV image of 1413 is input to another machine learning (ML) model, which is trained to detect pixels of oily chips in the UV image (i.e., these pixels are part of any oily chips in the UV image) and output a pixel count of oily chips in the UV image.

[0091] In block 1419 , a parameter representing the oil content of the chips (eg, labeled “Percentage of chips containing oil”) is determined based on a ratio of the pixel count of the oily chips of block 1417 relative to the pixel count of all chips of block 1415 .

[0092] In an embodiment, the workflow can be configured to minimize any movement of the chip sample when acquiring the white light image and the UV image and between the acquisition times of both the white light image and the UV image. In this way, the chip remains in the same position in the field of view of the imaging system while acquiring both the white light image and the UV image.

[0093] In other embodiments, the acquisition of the dark control image at 1409 and the processing of the UV image using the dark control image at 1413 may be omitted. In this case, the UV image at 1411 may be input to the ML model at 1417 to detect pixels of oily chips in the UV image.

[0094] In other embodiments, the operations of 1405 to 1413 of the workflow can be adapted to use separate WL and UV imaging systems to acquire WL images and UV images as described herein. In this case, the workflow can be configured to locate the chips at the same position in the corresponding fields of view of the separate imaging systems while acquiring both the white light image and the UV image.

[0095] In other embodiments, an image preprocessing operation may be performed on the WL image input to the machine learning (ML) model in block 1405. Such image preprocessing operations may provide image correction (such as adjusting the brightness or color balance of the WL image or filtering the WL image), transformation of the color space of the WL image, or other desired image processing operations on the WL image. Similarly, in block 1407, an image preprocessing operation may be performed on the UV image input to the additional machine learning (ML) model. Such image preprocessing operations may provide image correction (such as adjusting the brightness or color balance of the UV image or filtering the UV image), transformation of the color space of the UV image, or other desired image processing operations on the UV image.

[0096] Fig.15 An exemplary workflow for training a machine learning (ML) model to detect pixels of all chips in an arbitrary WL image (ie, pixels that are part of any chip in the WL image) is depicted.

[0097] In box 1501, drilling cuttings of different textures, colors and oils of different APIs are selected. The image background is selected to ensure good color contrast. Importantly, this step can be configured so that the drilling cuttings of the selected images are in different arrangements to ensure variability of the data. Different background colors and textures can also be included to capture degradation aspects in different well site environments.

[0098] Block 1503 is a data collection operation that involves acquiring a WL image of the cuttings of block 1501 .

[0099] Block 1505 involves labeling the WL images, i.e., for each WL image, labeling all pixels of the cuttings as "1" and labeling background pixels (not part of the cuttings) as "0". In an embodiment, the labeling of the pixels of a given WL image may involve binarization of the given WL image. There are many possible methods that may be implemented for binarization of a WL image.

[0100] For example, a WL image can be transformed from the color space of a WL image as captured by the system 152 (which is typically an RGB color space) to a hue-saturation-value (HSV) color space. The transformed WL image in the HSV color space is referred to herein as an HSV WL image. The RGB color space uses three components: red (which is typically in the range of 0-255), green (which is typically in the range of 0-255), and blue (which is typically in the range of 0-255). The RGB color space is an additive color space. In other words, red, green, and blue values ​​(referred to as the three primary colors) are combined to reproduce other colors. For example, the color "red" can be represented as [R=255, G=0, B=0], "purple" can be represented as [R=238, G=130, B=238], and so on. The HSV color space uses three components: hue (which is usually in the range of 0 to 360° and represents the color type); saturation (which is usually in the range of 0 to 100% and represents the intensity of the color; and value (which is usually in the range of 0 to 100% and represents the brightness of the color). The HSV color space is a nonlinear transformation of the RGB color space. In other words, colors are not defined as simple combinations (addition / subtraction) of primary colors, but as mathematical transformations. The resulting HSV WL image (or one or more components or channels H, S or V of the HSV WL image) can be processed by a suitable clustering or segmentation method (e.g., k-means clustering), which segments the pixels of the HSV WL image into two classes: a class "1" for pixels that are part of the chips, and another class "0" for background pixels (not part of the chips). In this way, pixels of class "1" correspond to all chips in the HSV WL image, and pixels of class "0" correspond to background pixels (not part of the chips) in the HSV WL image. HSV Pixels of class "1" in the WL image may be labeled as "1" and correspond to all chips in the HSV WL image, and corresponding pixels in the WL image may be labeled as "1" and correspond to all chips in the WL image. Pixels of class "0" in the HSV WL image may be labeled as "0" and correspond to background pixels (not part of the chips) in the HSV WL image, and corresponding pixels in the WL image may be labeled as "0" and correspond to background pixels in the WL image.

[0101] In another example, the WL image may be processed by a suitable edge-based segmentation method that segments the pixels of the WL image into two classes: class "1" for pixels that are part of the chips, and another class "0" for background pixels (not part of the chips). Edge-based segmentation employs one or more edge detection operators for edge detection and contour processing on the WL image, which contour processing locates pixels inside the edge boundary and thereby determines the perimeter or edge boundary pixels of each chip. In the case where multiple drilling cuttings are in contact with each other (or overlap with each other) in the WL image, the edge boundaries of the multiple cuttings may be merged together as a single chip. Pixels of the WL image within the edge boundary of the chip are assigned to class "1", while pixels of the WL image outside the edge boundary of the chip are assigned to class "0". In this way, pixels of class "1" correspond to all chips in the WL image, and pixels of class "0" correspond to background pixels in the WL image (not part of the chips). Pixels of class "1" in the WL image may be labeled "1" and correspond to all chips in the WL image, and pixels of class "0" in the WL image may be labeled "0" and correspond to background pixels (not part of the chips) in the WL image. Details of an exemplary edge-based segmentation method are described in U.S. Patent Publication No. 2021 / 0248428, which is commonly assigned to the assignee of the present application and is incorporated herein by reference in its entirety.

[0102] In other examples, other image binarization methods may be used to label pixels of the WL image.

[0103] Box 1507 involves training an ML model using the WL image of box 1503 and the associated label data of box 1505. For example, the ML model can be a convolutional encoder-decoder machine learning algorithm. The basic trainable encoder-decoder network is topologically like a multi-layer convolutional neural network. The encoder module aggregates multiple levels of features and downsampling of the data to obtain a coarse feature abstraction, and the decoder module upsamples these features to restore fine spatial classification. Typically, the encoder takes an input image and generates a high-dimensional feature vector. On the other hand, the role of the decoder is to map the high-dimensional (but low-resolution) feature vector to a feature map with the resolution of the original input, and thus achieve pixel-by-pixel labeling. The building blocks of the encoder module include convolution filters, element-by-element nonlinearity, maximum pooling, or downsampling based on strided convolutions - a sequence of these steps builds a representation of the potential features. The decoder module includes upsampling and transposed convolutions of the "skip connection" of the filters of the corresponding blocks from the encoder module. Fig.15 An example convolutional encoder-decoder neural network is shown in FIG. In other embodiments, other machine learning models may also be used.

[0104] Once passed Fig.15 The trained ML model can be deployed to interpret WL images, for example, Fig.14 This interpretation involves inputting an arbitrary WL image into the ML model. The ML model detects the pixels of all chips (both oily and non-oily) in the WL image and outputs the pixel count of all chips in the WL image.

[0105] In other embodiments, image preprocessing operations may be performed on the WL image that is subjected to image binarization in block 1505 and / or the WL image that is input to a machine learning (ML) model in block 1507 for training. Such image preprocessing operations may provide image correction (such as adjusting the brightness or color balance of the WL image or filtering the WL image), transformation of the color space of the WL image, or other desired image processing operations on the WL image.

[0106] Fig.16 Depicted is an exemplary workflow for training another machine learning (ML) model for detecting pixels containing oily chips in an arbitrary UV image (ie, pixels that are part of any oily chips in the UV image).

[0107] In block 1601, drill cuttings of different textures, colors, and oils of different APIs are selected. The image background is selected to ensure good color contrast. Importantly, this step can be configured so that the drill cuttings of the selected images are in different arrangements to ensure variability in the data. Different background colors and textures can also be included to capture degradation aspects in different wellsite environments. This step can be combined with Fig.15 The operation of box 1501 is shared.

[0108] Block 1603 is a data collection operation involving acquiring a UV image of the chips of block 1601 .

[0109] Block 1605 involves marking pixels of oily chips as "1" and marking other pixels (not part of oily chips, and therefore part of non-oily chips or background) as "0" for each UV image. In an embodiment, the marking of pixels of a given UV image may involve binarization of the given UV image. There are many possible methods that may be implemented for binarization of UV images.

[0110] For example, the UV image can be transformed from the color space of the UV image as captured by the system 152 (which is typically an RGB color space) to a hue saturation value (HSV) color space. The transformed UV image in the HSV color space is referred to herein as an HSV UV image. The resulting HSV UV image (or for one or more components or channels H, S, or V of the HSV UV image, such that the V channel of the HSV UV image) can be processed by a suitable clustering or segmentation method (e.g., k-means clustering) that segments the pixels of the HSV UV image into two classes: a class "1" for bright pixels that are part of the oily cuttings, and another class "0" for all other pixels (which are part of non-oily cuttings or background pixels). In this way, pixels of class "1" correspond to all oily cuttings in the HSV UV image, and pixels of class "0" correspond to non-oily cuttings and background pixels (not part of the cuttings) in the HSV UV image. Pixels of class "1" in the HSV UV image may be labeled as "1" and correspond to oily chips in the HSV UV, and corresponding pixels in the UV image may be labeled as "1" and correspond to oily chips in the UV image. Pixels of class "0" in the HSV UV image may be labeled as "0" and correspond to non-oily chips and background pixels in the HSV UV image, and corresponding pixels in the UV image may be labeled as "0" and correspond to non-oily chips and background pixels in the UV image.

[0111] In another example, the UV image may be processed by a suitable edge-based segmentation method that segments the pixels of the UV image into two classes: class "1" for pixels that are part of an oily chip, and another class "0" for pixels that are part of a non-oily chip or background pixels (not part of a chip). Edge-based segmentation employs one or more edge detection operators for edge detection and contour processing on the UV image, which locates pixels inside the edge boundary and thus determines the perimeter or edge boundary pixels of each chip. In the case where multiple chips are in contact with each other (or overlap each other) in the UV image, the edge boundaries of the multiple chips may be merged together as a single chip. Pixels of the UV image within the edge boundary of the chip are assigned to class "1", while pixels of the UV image outside the edge boundary of the chip are assigned to class "0". In this embodiment, pixels of class "1" correspond to oily chips in the UV image, and pixels of class "0" correspond to non-oily chips and background pixels (not part of a chip) in the UV image. Pixels of class "1" in the UV image may be labeled as "1" and correspond to oily chips in the UV image, and pixels of class "0" in the WL image may be labeled as "0" and correspond to oily chips and background pixels (not part of the chips) in the UV image. Details of an exemplary edge-based segmentation method are described in U.S. Patent Publication No. 2021 / 0248428, commonly assigned to the assignee of the present application and incorporated herein by reference in its entirety.

[0112] In another example, the UV image can be processed by a suitable threshold method that divides the pixels of the UV image into two classes: class "1" for pixels that are part of the oily cuttings, and another class "0" for pixels that are part of the non-oily cuttings or background pixels (not part of the cuttings). In an embodiment, the threshold pixel values ​​for the two classes can be determined from a trial and error method or by the Otsu method as described by Nobuyuki Otsu (1979). "A threshold selection method from gray-level histograms". IEEE Trans. Sys. Man. Cyber. 9(1): 62-66. Otsu's method is applicable to UV images with a bimodal distribution of pixel values.

[0113] In other examples, other image binarization methods may be used to label pixels of the WL image.

[0114] Block 1607 involves training an ML model using the UV image and associated label data of block 1605. For example, the ML model may be a convolutional encoder-decoder machine learning algorithm. Fig.16An example convolutional encoder-decoder neural network is shown in FIG. In other embodiments, other machine learning models may also be used.

[0115] Once passed Fig.16 The trained ML model can be deployed to interpret UV images, for example, Fig.14 The interpretation involves inputting the UV image into the ML model. The ML model detects pixels of oily chips in the UV image. The ML model outputs a pixel count of oily chips in the UV image.

[0116] In other embodiments, image preprocessing operations may be performed on the UV image that is subjected to image binarization in block 1605 and / or the UV image that is input to a machine learning (ML) model in block 1607 for training. Such image preprocessing operations may provide image correction (such as adjusting the brightness or color balance of the UV image or filtering the UV image), transformation of the color space of the UV image, or other desired image processing operations on the UV image.

[0117] Various systems and processes of the present disclosure may be described herein in the general context of software or program modules, or techniques and modules may be implemented in pure computing hardware. Software generally includes routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The implementation of these modules and techniques may be stored on or transmitted across some form of tangible computer-readable media. Computer-readable media may be any available data storage medium or medium that is tangible and can be accessed by a computing device. Therefore, computer-readable media may include computer storage media. "Computer storage media" specifies tangible media and includes volatile and non-volatile, removable and non-removable tangible media implemented to store information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical storage device, cassette, tape, disk storage device or other magnetic storage device, or any other tangible medium that can be used to store the required information and can be accessed by a computer. Some of the above methods and processes may be performed by a processor. The term "processor" should not be construed to limit the embodiments disclosed herein to any particular device type or system. The processor may include a computer system. The computer system may also include a computer processor (e.g., a microprocessor, a microcontroller, a digital signal processor, a general-purpose computer, a special-purpose machine, a virtual machine, a software container, or a device) for performing any of the above methods and processes.

[0118] The computer system may also include memory, such as semiconductor memory devices (e.g., RAM, ROM, PROM, EEPROM, or flash programmable RAM), magnetic memory devices (e.g., magnetic disks or fixed disks), optical memory devices (e.g., CD-ROMs), PC cards (e.g., PCMCIA cards), or other memory devices.

[0119] Alternatively or additionally, the processor may include discrete electronic components coupled to a printed circuit board, an integrated circuit (e.g., an application specific integrated circuit (ASIC)), and / or a programmable logic device (e.g., a field programmable gate array (FPGA)). Such logic devices may be used to implement any of the above methods and processes.

[0120] Some of the above methods and processes can be implemented as computer program logic for use with a computer processor. The computer program logic can be embodied in various forms, including source code form or computer executable form. The source code can include a series of computer program instructions in various programming languages ​​(e.g., object code, assembly language, or high-level languages ​​such as C, C++, or JAVA). Such computer instructions can be stored in a non-transitory computer readable medium (e.g., memory) and executed by a computer processor. The computer instructions can be distributed in any form as a removable storage medium with an accompanying printed or electronic document (e.g., shrink-wrapped software), pre-installed with a computer system (e.g., on a system ROM or fixed disk), or distributed from a server or electronic bulletin board via a communication system (e.g., the Internet or the World Wide Web).

[0121] Although only a few example embodiments have been described in detail above, those skilled in the art will readily appreciate that many modifications may be made to the example embodiments without materially departing from the invention. For example, the methods and processes of the present disclosure may also be performed on organic matter separated from a bulk formation sample to determine the properties of the organic matter.

[0122] The methods and systems of the present disclosure provide advantages over conventional methods and systems used for mud-logging activities.

[0123] First, the methods and systems of the present disclosure can be designed with a modular system to target multiple potential applications. In particular, the combination of UV LED filter camera / microscope provides a flexible (easily reconfigurable) design that allows two modes of operation: (i) low UV power for an open system, and thus the physical barrier between the operator and the UV source is very small (and still safe); and (ii) a housing-based closed system for higher power UV light.

[0124] Secondly, the methods and systems of the present disclosure can adopt a modular system design that provides integration and automation of white light and UV imaging operations. In this design, the same sensor (camera / microscope) is used to capture both WL and UV images of drilling cuttings, while eliminating the need to move the sample. This enables pixel overlap between image sequences, which is necessary for quantitative analysis.

[0125] Third, the methods and systems of the present disclosure can employ a modular system design that uses UV LEDs instead of fluorescent bulbs or other fluorescent light sources. UV LEDs are solid-state electronic devices that do not contain mercury (which is part of the fluorescent light source). This improves multiple HSE factors, including mercury-free systems, lower energy consumption, and reduced heat release, especially in hydrocarbon environments.

[0126] Fourth, the methods and systems of the present disclosure provide improved data quality in images of drilling cuttings, which results in better performance of machine learning methods for accurate pixel-level oil detection in drilling cuttings. For example, Figures 3 to 6 The images of drilling cuttings show that the method and system can be a key lever to realize the data-centric engineering pipeline of the well site and help estimate key sustainability indicators (such as oil detection and water quality measurement). In fact, data quality is the most important factor in measurement automation, machine learning-based detection, and easy interpretation of results. The method and system of the present disclosure use modern machine learning algorithms (especially deep learning methods) to fill important gaps in the data-centric pipeline and apply machine learning algorithms to challenging field environments with high variability of rock and fluid samples. By achieving high-quality data acquisition, the method and system of the present disclosure make machine learning models easy to implement and also improve detection performance.

[0127] Although only a few example embodiments have been described in detail above, it will be readily appreciated by those skilled in the art that many modifications may be made in the example embodiments without substantially departing from the present invention. Therefore, all such modifications are intended to be included within the scope of the present disclosure as defined in the appended claims. In the claims, means-plus-function clauses are intended to cover structures described herein as performing the described function, and not only structural equivalents, but also equivalent structures. Thus, although nails and screws may not be structural equivalents because nails employ cylindrical surfaces to secure wooden parts together, while screws employ helical surfaces, nails and screws may be equivalent structures in the context of fastening wooden parts. It is the express intention of the applicant not to invoke 35 U.S.C. § 112, paragraph 6, for any limitation of any claim herein, except for those where the claims expressly use the term "means for..." and the associated functions.

Claims

1. An imaging system for use with drilling cuttings, the imaging system comprising: an ultraviolet (UV) source, the UV source comprising a UV light emitting diode, wherein the UV source is configured to illuminate the sample volume with UV radiation that interacts with crude oil bound to drill cuttings located in the sample volume to induce fluorescent emission of photons in the visible region of the electromagnetic spectrum; and A camera system includes an image sensor, wherein the camera system is configured to capture at least one image of drill cuttings located in the sample volume based on fluorescent emissions from the drill cuttings.

2. The imaging system of claim 1, wherein: The UV source is configured to illuminate the sample volume with at least one predetermined UV wavelength band, wherein the at least one predetermined UV wavelength band interacts with crude oil bound to drill cuttings located in the sample volume to induce fluorescent emission of photons in the visible region of the electromagnetic spectrum.

3. The imaging system of claim 1, wherein: The UV light emitting diode is configured to emit UV radiation in a UV wavelength band centered around 365 nm with a peak intensity at a wavelength of 365 nm.

4. The imaging system of claim 1, wherein: The UV source also includes an optical bandpass filter configured to transmit UV radiation within at least one predetermined UV wavelength band and block other UV wavelength bands, wherein the at least one predetermined UV wavelength band interacts with crude oil bound to the drill cuttings located in the sample volume to cause fluorescent emission of photons in the visible region of the electromagnetic spectrum.

5. The imaging system of claim 4, wherein: The optical bandpass filter is configured to transmit UV radiation in a wavelength band centered around 365 nm and substantially block transmission of UV radiation of wavelengths outside the wavelength band for transmission; and / or The wavelength band of transmission for the optical bandpass filter extends over a narrow UV wavelength range of 100 nm or less; and / or The optical bandpass filter is configured to block visible light in a wavelength band between 400 nm and 700 nm.

6. The imaging system of claim 1, further comprising: An optical filter is in the optical path between the sample volume and the camera system.

7. The imaging system of claim 1 , further comprising: a white light source configured to illuminate the sample volume with white light; Wherein the camera system is further configured to capture at least one image of drill cuttings located within the sample volume and illuminated by white light.

8. A system for a mud-logging workflow, the system comprising: The imaging system according to claim 7; and a data processing system operably coupled to the imaging system, wherein the data processing system is configured to process at least one image of drill cuttings located within the sample volume and illuminated by white light as captured by the imaging system, and to process at least one image generated by fluorescent emissions from drill cuttings located in the sample volume as captured by the imaging system, so as to quantitatively characterize the oil content of the drill cuttings.

9. The system according to claim 8, wherein: The data processing system is configured to: i) processing at least one image of drill cuttings captured by the imaging system and located within the sample volume and illuminated by white light to determine a first pixel count of all cuttings; ii) processing at least one image captured by the imaging system generated by fluorescent emission from drill cuttings located in the sample volume to determine a second pixel count of oil-containing cuttings; and iii) determining a parameter indicative of oil content of the drill cuttings based on the first pixel count and the second pixel count.

10. The system of claim 9, wherein: The data processing system is configured to determine a parameter indicative of oil content of the drill cuttings based on a ratio of the second pixel count to the first pixel count.

11. The system of claim 9, wherein: The data processing system employs a first machine learning model to determine a first pixel count for all chips; and The data processing system employs a second machine learning model to determine a second pixel count of the oily cuttings.

12. The system of claim 11, wherein: training a first machine learning to determine a first pixel count of all cuttings from images of drilling cuttings of different textures, colors, and oils of different APIs and associated label data; and The second machine learning is trained to determine second pixel counts of oil-containing cuttings from images of drilling cuttings of different textures, colors, and oils of different APIs and associated label data.

13. The system of claim 11, wherein: A first machine learning model is trained to determine a first pixel count of all cuttings having label data derived from binarization of an image of drill cuttings illuminated by white light.

14. The system of claim 13, wherein: Binarization of the image employs i) clustering of hue, saturation or value image data obtained by transforming the color space of the image, or ii) edge-based segmentation of the image.

15. The system of claim 11, wherein: A second machine learning model is trained to determine a second pixel count of the oily cuttings, wherein the label data is derived from binarization of an image illuminated by UV radiation that causes fluorescent emission from the oily cuttings.

16. The system of claim 15, wherein: The binarization of the image employs i) clustering of hue, saturation or value image data obtained by transforming the color space of the image, or ii) edge-based segmentation of the image, or iii) thresholding of the image.

17. The system of claim 11, wherein: a first machine learning including a convolutional encoder-decoder neural network trained to determine a first pixel count of all cuttings given an image of drill cuttings illuminated by white light; and / or The second machine learning model includes a convolutional encoder-decoder neural network trained to determine a second pixel count of oil-bearing cuttings given an image of drill cuttings illuminated by UV radiation.

18. A method for characterizing the oil content in cuttings, the method comprising: capturing at least one image of the drill cuttings illuminated by white light; capturing at least one image of the drill cuttings based on fluorescent emissions produced by exposure to UV radiation; and Configuring a data processing system to perform operations including i) processing at least one image of the drill cuttings illuminated by white light to determine a first pixel count of all the cuttings; ii) processing the at least one image based on the fluorescent emission to determine a second pixel count of the oily cuttings; and iii) determining a parameter indicative of oil content of the drill cuttings based on the first pixel count and the second pixel count.

19. The method of claim 18, wherein: The data processing system is configured to determine a parameter indicative of oil content of the drill cuttings based on a ratio of the second pixel count to the first pixel count.

20. The method of claim 18, wherein: The data processing system employs a first machine learning model to determine a first pixel count for all chips; and The data processing system employs a second machine learning model to determine a second pixel count of the oily cuttings.

21. The method of claim 20, wherein: training a first machine learning to determine a first pixel count of all cuttings from images of drilling cuttings of different textures, colors, and oils of different APIs and associated label data; and The second machine learning is trained to determine second pixel counts of oil-containing cuttings from images of drilling cuttings of different textures, colors, and oils of different APIs and associated label data.

22. The method of claim 20, wherein: A first machine learning model is trained to determine a first pixel count of all cuttings having label data derived from binarization of an image of drill cuttings illuminated by white light.

23. The method of claim 22, wherein: Binarization of the image employs i) clustering of hue, saturation or value image data obtained by transforming the color space of the image, or ii) edge-based segmentation of the image.

24. The method of claim 20, wherein: A second machine learning model is trained to determine a second pixel count of the oil-bearing cuttings, wherein the label data is derived from binarization of an image of the drilling cuttings illuminated by UV radiation, which causes fluorescent emission from the oil-bearing cuttings.

25. The system of claim 24, wherein: The binarization of the image employs i) clustering of hue, saturation or value image data obtained by transforming the color space of the image, or ii) edge-based segmentation of the image, or iii) thresholding of the image.

26. The method of claim 20, wherein: a first machine learning including a convolutional encoder-decoder neural network trained to determine a first pixel count of all cuttings given an image of drill cuttings illuminated by white light; and / or The second machine learning model includes a convolutional encoder-decoder that is neurally trained to determine a second pixel count of the oil-containing cuttings when an image of the drill cuttings illuminated by UV radiation causes fluorescent emissions from the oil-containing cuttings.

27. The method of claim 18, wherein: At least one image of drill cuttings illuminated by white light and at least one image based on fluorescent emissions are captured by an imaging system having a white light source and a UV source configured to illuminate a sample volume and a camera system configured to acquire images of drill cuttings located in the sample volume and illuminated by the white light source or the UV source without moving the drill cuttings.

28. The method of claim 18, wherein: At least one image of the drill cuttings illuminated by white light and at least one image based on fluorescent emissions are captured by separate and distinct white light imaging and UV imaging systems.

29. The method of claim 18, wherein: The UV radiation is generated by a UV source comprising a UV light emitting diode.

30. The method of claim 29, wherein: The UV light emitting diode is configured to emit UV radiation in a UV wavelength band centered around 365 nm with a peak intensity at a wavelength of 365 nm.

31. The method of claim 29, wherein: The UV source is configured to illuminate the sample volume with at least one predetermined UV wavelength band, wherein the at least one predetermined UV wavelength band interacts with crude oil bound to drill cuttings located in the sample volume to induce fluorescent emission of photons in the visible region of the electromagnetic spectrum.

32. The method of claim 29, wherein: The UV source also includes an optical bandpass filter configured to transmit UV radiation within at least one predetermined UV wavelength band and block other UV wavelength bands, wherein the at least one predetermined UV wavelength band interacts with crude oil bound to the drill cuttings located in the sample volume to cause fluorescent emission of photons in the visible region of the electromagnetic spectrum.

33. The method of claim 32, wherein: The optical bandpass filter is configured to transmit UV radiation in a wavelength band centered around 365 nm and substantially block transmission of UV radiation of wavelengths outside the wavelength band for transmission; and / or The wavelength band of transmission for the optical bandpass filter extends over a narrow UV wavelength range of 100 nm or less; and / or The optical bandpass filter is configured to block visible light in a wavelength band between 400 nm and 700 nm.

Images