Cable inspection and analysis based on hyperspectral imaging

Abstract

The present disclosure is directed to systems and techniques for cable inspection and analysis using hyperspectral image data. For example, a method can include obtaining one or more hyperspectral images associated with a cable, wherein each hyperspectral image associated with the cable includes respective spectral information for a plurality of different spectral bands. The method can further include analyzing the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable. The plurality of different spectral bands can be continuous spectral bands of the hyperspectral wavelength range. Each hyperspectral image can be obtained for a different location along a longitudinal length of the cable and / or for a same location along a longitudinal length of the cable.

Description

CABLE INSPECTION AND ANALYSIS BASED ON HYPERSPECTRAL IMAGINGCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of priority to U.S. Provisional Patent Application No. 63 / 522,936 filed June 23, 2023, the disclosure of which is hereby incorporated by reference in its entirety and for all purposes.TECHNICAL FIELD

[0002] The present disclosure generally relates to line management, including in rope and cable systems. For example, aspects of the present disclosure are related to systems and techniques for cable inspection and analysis using hyperspectral image data.BACKGROUND

[0003] As used herein, the terms “cable”, “rope”, and “rope and cable” may be used interchangeably. A cable can be considered a tensile strength member, in that a cable can transmit tensile forces but not compressive forces. For example, flexible cables can be connected between two components and used to transmit a tensile force between the two components. Cables often include end-fittings configured to transmit a load. The assembly of an end-fitting and the portion of the cable to which it is attached can be referred to as a “termination.”

[0004] The usable lifetime of a cable can depend on various factors, including characteristics of the particular deployment of the cable and material properties of the cable itself. For example, cables can be used to perform various tasks, such as fastening, lowering, lifting, etc., of various objects. Cables can be used in various environments, such as maritime environments (e g., where the cable may be at least partially submerged), land-based environments, indoor environments, etc. The usable lifetime of a cable can depend on the type(s) of loading cycle(s) imparted to the cable. For example, cyclical loads are often imposed on cables deployed in a maritime environment, and can cause accelerated fatigue in parts of the cable. Failure of a cable can have undesirable consequences, particularly a failure of a cable while under tension or load.

[0005] In addition to using loading information of a cable to measure or estimate fatigue, remaining usable lifetime, and / or various other safety-related characteristics, loading information of a cable can also be used to drive one or more cable management actions. For instance, differentactions may be taken when a cable is in a loaded (e.g., tensioned) state as compared to when a cable is in an unloaded (e.g., non-tensioned state). Similarly, different actions may be taken when a cable is an overloaded state (e.g., current load exceeding one or more load threshold values) as compared to when a cable is not in an overloaded state (e.g., current load not exceeding one or more load threshold values). Load information of a cable can be utilized as discrete measurements corresponding to discrete points in time and / or can be utilized as a time-series of measurements corresponding to changes in load over a period of time.SUMMARY

[0006] Systems and techniques are described for performing cable inspection and analysis using hyperspectral image data obtained from one or more locations along a longitudinal length of a cable. According to at least one illustrative example, a method is provided, the method comprising: obtaining one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyzing the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

[0007] In some aspects, each hyperspectral image of the one or more hyperspectral images is associated with a hyperspectral wavelength range that is wider than a respective wavelength range associated with a visible light spectrum.

[0008] In some aspects, the plurality of different spectral bands are continuous spectral bands of the hyperspectral wavelength range.

[0009] In some aspects, each hyperspectral image is obtained for a different location along a longitudinal length of the cable.

[0010] In some aspects, each hyperspectral image is obtained for a same location along a longitudinal length of the cable.

[0011] In some aspects, analyzing the one or more hyperspectral images includes analyzing the respective spectral information for the plurality of different spectral bands in combination.

[0012] In some aspects, analyzing the one or more hyperspectral images includes analyzing at least a portion of the respective spectral information for the plurality of different spectral bandsindependently from a remaining portion of the respective spectral information for the plurality of different spectral bands.

[0013] In some aspects, a first state of wear of the cable or a first type of damage to the cable is associated with a first pattern of spectral information in a particular spectral band.

[0014] In some aspects, the method further includes analyzing the one or more hyperspectral images to detect the first state of wear of the cable or the first type of damage to the cable.

[0015] In some aspects, the analyzing is based on comparing the respective spectral information for the particular spectral band to the first pattern of spectral information associated with the first state of wear of the cable or the first type of damage to the cable.

[0016] In another illustrative example, an apparatus is provided that includes a memory (e.g., configured to store data, such as virtual content data, one or more images, etc.) and one or more processors (e.g., implemented in circuitry) coupled to the memory. The one or more processors are configured to and can: obtain one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyze the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

[0017] In another illustrative example, a non-transitory computer-readable medium is provided, the non-transitory computer-readable medium having stored thereon instructions which, when executed by one or more processors, cause the one or more processors to perform operations including: obtaining one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyzing the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

[0018] This summary is not intended to identify key or essential features of the claimed subject matter, nor is it intended to be used in isolation to determine the scope of the claimed subject matter. The subject matter should be understood by reference to appropriate portions of the entire specification of this patent, any or all drawings, and each claim.

[0019] The foregoing, together with other features and embodiments, will become more apparent upon referring to the following specification, claims, and accompanying drawings.BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Details of one or more aspects of the subject matter described in this disclosure are set forth in the accompanying drawings and the description below. However, the accompanying drawings illustrate only some typical aspects of this disclosure and are therefore not to be considered limiting of its scope. Other features, aspects, and advantages will become apparent from the description, the drawings, and the claims.

[0021] FIG. 1A depicts an example of plastic objects having various colors and material compositions, as imaged using a visual light imaging system;

[0022] FIG. IB depicts an example of the plastic objects of FIG. 1A, as imaged using a hyperspectral imaging system; and

[0023] FIG. 2 depicts another example of visual light (e g., RGB) imaging vs. hyperspectral imaging.DETAILED DESCRIPTION

[0024] Certain aspects and embodiments of this disclosure are provided below. Some of these aspects and embodiments may be applied independently and some of them may be applied in combination as would be apparent to those of skill in the art. In the following description, for the purposes of explanation, specific details are set forth in order to provide a thorough understanding of embodiments of the application. However, it will be apparent that various embodiments may be practiced without these specific details. The figures and description are not intended to be restrictive.

[0025] The ensuing description provides example embodiments only, and is not intended to limit the scope, applicability, or configuration of the disclosure. Rather, the ensuing description of the exemplary embodiments will provide those skilled in the art with an enabling description for implementing an exemplary embodiment. It should be understood that various changes may be made in the function and arrangement of elements without departing from the spirit and scope of the application as set forth in the appended claims.

[0026] Various aspects of the present disclosure are described below.H perspectral Imaging

[0027] Visible light images are captured based on collecting and processing information within the visible light spectrum, spanning wavelengths between approximately 400 - 700 nanometers(nm). For instance, visible light images can encode an intensity value and a color value for each pixel of a plurality of pixels in the visible light image. The intensity value represents the relative “brightness” of the particular pixel, while the color value represents a particular color within the visible light spectrum. Visible light images can use various color models. For instance, the YUV color model encodes luminance information separately from chrominance information. The RGB color model combines luminance and chrominance information, and represents each pixel using a red (R) component, green (G) component, and blue (B) component. Various other color models can be used to encode visible light images corresponding to the ~ 400 - 700 nm range of wavelengths.

[0028] In hyperspectral imaging (HSI), information is collected and processed both within the visible light spectrum and beyond the visible light spectrum. For example, a hyperspectral image can include information from the visible light spectrum and the ultraviolet (UV) spectrum (e.g., ~400-700nm and <400 nm); can include information from the visible light spectrum and the infrared spectrum (e.g., ~400-700nm and >700nm). In many cases, hyperspectral imaging can be performed based on extending the imaged spectrum beyond the visible light spectrum on both ends (e.g., below ~400nm and above ~700nm).

[0029] More generally, hyperspectral imaging can be understood to collect and process information from across the electromagnetic spectrum. Incident light striking each pixel is broken down or decomposed into many different spectral bands, and the spectral information can provide additional information on the scene or object that is being imaged. For instance, in hyperspectral imaging, a unique color signature of an individual object can be detected (e.g., based on hyperspectral images including or distinguishing a full color spectrum in each pixel). In some cases, hyperspectral images can be understood to augment conventional 2D or spatial images with spectral information of the scene or object being imaged. That is, HSI techniques can be understood to provide a combination of both spatial and spectral information of a scene or objects being imaged, such that, for each spatial pixel, a series of corresponding spectral information representing the chemical properties of the local point is obtained. In some examples, hyperspectral imaging data can include a plurality of spatial pixels (e.g., having dimensions of (x, y )), wherein the hyperspectral imaging data further includes a corresponding continuous spectral curve, with tens to hundreds of narrow bands, for each pixel of the plurality of pixels.

[0030] For instance, FIG. 1A depicts an RGB (e g., visual light) image of various different plastic objects. The plastic objects have different material compositions and different colors (e.g., the plastic objects are made of various different types of plastics - PET, PP, PVC, HDPE, and PS - and using various different dye colors - white, green, and red). The visual light / RGB image of FIG. 1A corresponds approximately to how the human eye would perceive the arrangement of plastic objects. In particular, the visual light / RGB image is able to differentiate the plastic objects based on their dye colors (white, green, or red), but provides no information indicative of the material composition of the plastic objects.

[0031] By comparison, FIG. IB depicts a hyperspectral image of the same plastic objects contained in the visual light / RGB image of FIG. 1A. More accurately, the hyperspectral image of FIG. IB can be understood to provide a hyperspectral representation that uses false colors to indicate the different spectral information corresponding to the different material compositions / plastic type used for each object in the image. The use of false colors to represent hyperspectral images is based on the fact that the additional hyperspectral information obtained using HSI techniques is, by definition, beyond the range of visual light - and is therefore beyond the range of human perception.

[0032] In the example of FIG. IB, each type of plastic and its chemical composition (e g., type of plastic material) can be identified and measured in the hyperspectral imaging data. For instance, the spectral properties of PET differ from the spectral properties of PP which differ from the spectral properties of PVC, etc. In other words, the spectral reflectance of different materials, in different physical conditions, etc. will be different across one or more spectral bands measured in the hyperspectral imaging data. Accordingly, the different types of plastic depicted in the hyperspectral image of FIG. IB can be identified and / or measured based on these differences in spectral reflectance / spectral properties. Here, objects made of PET plastic material are shown in red; objects made of PP plastic material are shown in yellow; objects made of PVC plastic material are shown in green; objects made of HDPE plastic material are shown in blue; and objects made of PS plastic material are shown in purple.

[0033] Notably, hyperspectral imaging can reveal information that cannot otherwise be determined using conventional imaging techniques (e.g., such as RGB / visual light imaging). For instance, the three plastic objects at the upper right of FIG. 1A appear to be the same under RGB / visual light imaging - all three objects have the same white color. However, the hyperspectralimage of FIG. IB reveals that these same three plastic objects at the upper right are different from one another - one object is made of an HDPE plastic material, while the remaining two objects are made of PET plastic material.

[0034] FIG. 2 provides another illustrative example of the additional information that can be extracted or otherwise determined from hyperspectral image data. At the left, FIG. 2 depicts an arrangement of fruits (plums, in this particular example) imaged using RGB / visual light imaging techniques. At the right, FIG. 2 depicts another portion of the same scene, as imaged using hyperspectral imaging techniques. In the RGB image portion on the left, the plums have approximately uniform visual appearances, and no particular plum is easily distinguishable from another. However, in the hyperspectral image portion on the right, the additional spectral information captured in the hyperspectral image data can be analyzed to reveal contaminants that are imperceptible in the RGB image. For instance, plums without surface contamination are represented in the hyperspectral image of FIG. 2 using a red overlay color. Plums with mold surface contamination are represented using a green overlay color, and are detected based on the spectral reflectance of a mold-covered surface being different from the spectral reflectance of a non-mold- covered surface. Additional contamination is revealed in the hyperspectral image of FIG. 2 in the form of plastic materials clinging to or intermingled with the plums - here, plastic materials are represented using a blue overlay color, and again can be detected based on the spectral reflectance of a plastic material / surface being different from the spectral reflectance of a non-plastic material or surface. The spectral reflectance and / or spectral difference information used to extract this additional information from hyperspectral images is contained in the hyperspectral image data itself, recalling that, as described previously above, hyperspectral image data can be captured to include spectral information over a plurality of spectral bands (e.g., wavelengths) for each spatial pixel of the hyperspectral image / each spatial location that is imaged using HSI techniques.

[0035] Various techniques and / or types of imaging devices can be used to capture or otherwise generate hyperspectral images. For example, spatial scanning can be used to acquire hyperspectral image data over time. Spatial scanners for hyperspectral imaging can include push broom scanners, whisk broom scanners, etc. In another example, spectral scanning can be used to successively image an area at different wavelengths which can be combined or composited into a hyperspectral image. Spectral scanners can also be referred to as band sequential scanners. In another example, snapshot hyperspectral imaging can be used to generate a hyperspectral image in a single captureprocess (e.g., a single snapshot). Snapshot hyperspectral imaging can be performed based on using a staring array (also referred to as a staring-plane array or a focal-plane array (FPA)).

[0036] As mentioned previously, hyperspectral imaging can be performed to obtain the spectrum for each pixel in an image of a scene. For instance, a five megapixel (MP) resolution may be 2,560 x 1,920 pixels - a 5MP hyperspectral image can include a captured spectrum for each of the 2,560 * 1,920 = 4,915,200 pixels included in the 5MP hyperspectral image. The spectral information of each pixel can be collected and / or processed over a same or similar range of wavelengths (e.g., a same or similar portion of the electromagnetic spectrum). For example, a hyperspectral image captured between UV and infrared wavelengths can include captured UV- infrared spectrum information for each constituent pixel included in the hyperspectral image, etc.

[0037] In some aspects, a hyperspectral image can comprise a plurality of individual “images” of an area, where each individual image corresponds to a different wavelength range of the electromagnetic spectrum. Each individual image can also be referred to as corresponding to a different spectral band. In some examples, the plurality of individual images, each corresponding to a different spectral band, can be combined to form a three-dimensional hyperspectral data cube (e.g., an array of values). The 3D hyperspectral data cube can have dimensions of (x, y, d), where x and represent the 2D spatial dimensions of each image and d represents the spectral dimension of the image. For instance, d can be indicative of the range of wavelengths included in the spectral bands of the 3D hyperspectral data cube and / or can be indicative of a quantity of constituent spectral bands used to generate the 3D hyperspectral data cube, etc. In some examples, hyperspectral image data can have a 2D spatial dimension of (x, y) pixels, and can have a channel dimensionality of d, where d is greater than three. In other words, hyperspectral image data can be captured or obtained to include a greater quantity of channels than a similar visual light image data. For example, an RGB visual light image has three different channels: an R (red) channel, a G (green) channel, and a B (blue) channel. Hyperspectral image data can include a plurality of channels that are captured in addition to RGB color channels (e.g., such that the hyperspectral image data includes R, G, and B color channels and additionally includes one or more channels corresponding to infrared, UV, etc., wavelengths). In another example, hyperspectral image data can include a plurality of channels that are each different from a conventional visual light imaging R, G, or B color channel. In some cases, a hyperspectral image and / or a hyperspectral sensor (e.g.,included in a hyperspectral imaging device) can be characterized by a spectral resolution. The spectral resolution can be indicative of the width of each spectral band that is captured.

[0038] The acquisition and processing of hyperspectral image data can also be referred to as imaging spectroscopy and / or spectral imaging, in which image data is captured within specific wavelength ranges across the electromagnetic spectrum. A hyperspectral imaging device can also be referred to as an imaging spectrometer. Imaging spectroscopy / spectral imaging can include multispectral imaging (e.g., in which light is measured in a relatively small quantity of spectral bands) and hyperspectral imaging (e.g., in which light is measured in a relatively large quantity of spectral bands). In some cases, multispectral imaging can be used to refer to spectral imaging in which light is measured in a relatively small quantity of spectral bands, such as three to 15, while hyperspectral imaging can be used to refer to spectral imaging in which light is measured in dozens (or hundreds) of different spectral bands.

[0039] In some cases, hyperspectral imaging can refer to a type of spectral imaging that uses continuous and contiguous ranges of wavelengths, and multispectral imaging can refer to a type of spectral imaging that uses a subset of target wavelengths at chosen locations of the electromagnetic spectrum. For example, hyperspectral imaging may be performed using a range of wavelengths from 400 - 1100 nm, in steps of 1 nm (e.g., a total of 700 different Inm spectral bands) whereas multispectral imaging may be performed using target wavelengths chosen in the range of 400 - 1100 nm in steps of 20+ nm. Accordingly, multispectral imaging bands are discrete and relatively narrow bands that may not correspond to the full spectrum of the scene or object being imaged. Because hyperspectral imaging bands are spectral bands over a continuous spectral range (e.g., are continuous bands), hyperspectral images can be seen to produce the spectra of all pixels in the scene or object being imaged (e.g., hyperspectral images are spectrally-resolved images).

[0040] The spectral data collected and used to generate a hyperspectral image can be analyzed to provide insight into the sources of radiation represented in the hyperspectral image (e.g., the sources of electromagnetic radiation measured within the different spectral bands). Hyperspectral imaging and spectral data analysis has various applications, including (in the context of terrestrial observations), vegetation identification, physical condition analysis, mineral identification, water quality assessment, etc.

[0041] As noted previously above, hyperspectral imaging (HSI) techniques can be understood to provide a combination of both spatial and spectral information of a scene or objects being imaged, such that, for each spatial pixel, a series of corresponding spectral information representing the chemical properties of the local point is obtained. For instance, a given hyperspectral image data can include a plurality of spatial pixels arranged in dimensions of (x, y) for a total pixel count of x*y pixels. The hyperspectral image data can further include a corresponding continuous spectral curve, with tens to hundreds of narrow bands, for each pixel of the x*y pixels. As such, both the dimensionality and the data size associated with hyperspectral images can be exponentially larger than that associated with conventional (e.g., RGB, visual light, etc.) imaging.

[0042] In many cases, the large amount of data and the high channel dimensionality of hyperspectral image data can have negative impacts upon (or even prevent) the classification or other analysis of hyperspectral images using conventional image processing pipelines, systems, and / or techniques. In other words, the huge quantity of data information per hyperspectral image (compared to an RGB or other visual light image) presents a first challenge when working with hyperspectral images. The large quantity of channels (e.g., each corresponding to a separate spectral band representing chemical and / or material properties of a local spatial point) for each spatial location / pixel presents another challenge when working with hyperspectral images, as a corresponding analog is not present in RGB or visual light images, or image processing techniques for RGB / visual light images. A still further challenge is presented by the often unknown and / or unquantified correlation(s) between the spectral information contained in the plurality of different channels of a given hyperspectral image.Cable Inspection and Analysis Based on Hyperspectral Imaging (HSI)

[0043] In one illustrative example, hyperspectral imaging can be utilized to analyze, inspect, and / or monitor cable systems. A cable system can include one cable, multiple cables, a plurality of cables, etc. Cable systems form a long-span structure between two objects subject to tensions typically used for pulling, fastening, attaching, carrying, lifting, and climbing. More specifically, cable systems are often used to transport or move weight or materials between two or more objects. Cables are often flexible structures that support the applied transverse loads by the tensile resistance developed in their members. Cables are used in suspension bridges, tension leg offshore platforms, transmission lines, and several other engineering applications. In some cases, the term“cable” can be used interchangeably to refer to ropes, cables, and similar members such as lifting slings, roundslings, etc. Cables can be manufactured from various metals such as steel (and various alloys thereof). Cables may also be manufactured from various artificial and / or manmade fibers. For example, cables manufactured from one or more types of synthetic fibers or synthetic filaments can also be referred to as “synthetic cables”.

[0044] As used herein, a “cable” that is associated with a hyperspectral image and / or hyperspectral imaging operations can be a synthetic fiber or filament cable (e.g., cables manufactured from one or more synthetic filament materials, comprising multiple strands of wound synthetic filaments, etc.). These include DYNEEMA (ultra-high-molecular weight polyethylene), SPECTRA (ultra-high-molecular weight polyethylene), TECHNORA (processed terephthaloyl chloride), TWARON (para-aramid), KEVLAR (para-aramid), VECTRAN (liquid crystal polymer), PBO (polybenzobisoxazole), carbon fiber, and glass fiber (among many others). Modem cables may also be made of older, lower-strength synthetic materials such as NYLON. In the case of high-strength synthetics, the individual filaments have a thickness that is less than that of human hair. The filaments are very strong in tension, but they are not very rigid. They also tend to have low surface friction. These facts make such synthetic filaments difficult to handle during the process of adding a termination and difficult to organize. Hybrid cable designs are also emerging in which traditional materials are combined with high-strength synthetic materials. Cables made from synthetic filaments have a wide variety of constructions. In many cases a protective jacket will be provided over the exterior of the synthetic filament. This jacket does not carry any significant tensile load, and it may therefore be made of a different material. Most larger cables are made as an organized grouping of smaller cables. The smaller cables are often referred to as “strands.” One example is a parallel core of synthetic filaments surrounded by a jacket of braided filaments. In other cases the cable may be braided throughout. In still other examples the cable construction may be: (1) an entirely parallel construction enclosed in a jacket made of different material (2) a helical “twist” construction, (3) a more complex construction of multiple helices, multiple braids, or some combination of helices and braids, or (4) a hybrid construction including metallic constituents.

[0045] The structural capabilities, environmental factors, and maintenance of the equipment often determine the effectiveness of a cable. Cables, like most equipment, are subject to gradual wear and tear that, with time, result in a more or less rapid loss of initial performance. For example,under normal use conditions of the cable, the characteristics of the cable will change over time. Due to a variety of factors that include the weight of the load the cable is subjected to, tension, or environmental effects, the cable can deteriorate or lose effectiveness in its structural and operational characteristics. A cable that is structurally compromised can create an increase in safety risks and the decrease the longevity of the equipment the cable is associated with.

[0046] In some examples, the inspection of various cables can occur manually or visually, where a member of the crew will physically handle and inspect the cabling system in an effort to detect or identify any structural compromises of one or more cables in the cabling system. These inspections are often instructed by manuals and / or procedures created by a manufacturer or operator of the cabling system, where instructions are provided for performing a thorough inspection. However, despite the available procedural approaches, the visual inspection of cables remains a highly subjective process that is subject to various sources of error.

[0047] Visual inspections can also miss (e.g., fail to identify) some types of damage that can occur to a cable. When exposure to damaging conditions does occur, damage to the cable can manifest in various ways, some of which make it difficult or even impossible to determine what has happened to the cable. For instance, the exposure of a cable (e.g., a synthetic fiber cable) to damaging or adverse environmental conditions may not be represented as wear or damage that can be detected in a manual inspection of the cable, based at least in part on the observation that a majority of the wear experienced by a cable occurs on the interior of a cable, while many types of existing cable inspection techniques are largely limited to examining only the exterior surface of the cable.

[0048] While it is possible to perform a manual, visual inspection of the interior of a cable, the process of doing so is often complex and / or time-consuming, as an inspector must physically separate the fibers or filaments of the cable in order to “open up” and expose the interior of the cable. Moreover, because the process of visually inspecting the interior of a cable is a manual one that requires physical manipulation of the cable in order to first expose the interior, such inspections of the cable interior are often performed at only a limited number of locations along the length of the cable (e.g., at a pre-determined interval such as every 50 or 100 feet, at a random selection of a pre-determined number of locations along the length of the cable, etc.). Accordingly, existing approaches for manual visual inspection of the cable interior may be insufficient, as wearor damage may be present in the interior of the cable at locations that were not selected for opening up and inspecting.

[0049] Further still, the manual visual inspection of the interior of a cable cannot be performed while the cable is in use or otherwise under tension. Accordingly, not only is manual visual inspection of the interior of a cable limited in its efficacy along the entirety of the length of the cable, it also requires a cable to be taken out of service before the inspection can be performed.

[0050] In other techniques, efforts have been made to automate the inspection of cables using external imaging systems that pass the length of a cable through the field of view of one or more fixed cameras. For instance, one or more sheaves may be used to run a cable through a known path, where at least a portion of the known path is surrounded by or otherwise covered by the field of view of one or more fixed cameras (e.g., a plurality of cameras arranged in a ring or circular fashion about the circumference of the cable as it is forced by the sheaves through the known / fixed path). While such approaches can be seen to automate the image collection underlying a cable inspection and analysis process, the visual light images provide only for an inspection of the exterior of the cable - and the majority of wear or damage experienced by a cable often presents in the interior of the cable. As such, automated cable inspection processes are often limited in their accuracy and / or ability to detect wear or damage in a timely fashion, based on the lack of actionable insight or data pertaining to the interior cable condition that can be inferred from purely external visual light images of the cable’s outer surface.

[0051] Errors that occur in the visual inspection of a cable can result in the inspection passing cables that have undetected damage or other significant problems. Often, the uncertainty associated with visual inspections of cables can result in a cable being retired too early (e.g., a conservative approach, which can present operational and financial challenges) or the cable being retired too late (e.g., an aggressive approach, which can pose a significant safety risk).

[0052] In some embodiments, the systems and techniques described herein can use hyperspectral imaging to obtain or otherwise determine information associated with the chemical constituents of various materials, objects, etc., that are depicted in a hyperspectral image. For instance, hyperspectral images can be analyzed to differentiate between substances with different material compositions, different conditions, different defects or abnormalities, etc., as will be described in greater depth below. In one illustrative example, the systems and techniques can be used to perform cable inspection and / or analysis based on hyperspectral imaging (e.g., cableinspection and / or analysis based on one or more hyperspectral images of a cable being inspected or analyzed). For instance, hyperspectral images of cables can be used to identify, detect, and / or predict the location or occurrence of damage or other types of wear experienced by a cable. Damage or other types of wear experienced by a cable can be detected based on analyzing the hyperspectral signature (e.g., spectral information) of the cable, where the analysis is performed at one or more locations along and / or within the cable.

[0053] For example, damaged or worn sections of cable may exhibit different material properties as compared to un-damaged or un-worn sections of the same or similar cables. Accordingly, it is contemplated that damaged / worn sections of cable may exhibit a different hyperspectral signature in one or more wavelengths / spectral bands as compared to the hyperspectral signature of undamaged / unworn sections of cable. In some embodiments, one or more machine learning models can be trained to detect the presence of damage or wear in hyperspectral images of cables. For instance, a trained machine learning model can receive as input one or more hyperspectral images of a particular cable, and can classify the hyperspectral images (or portions thereof) according to the presence of damage or wear detected in the imaged portion of the cable. In some aspects, different machine learning models can be implemented for different types of cables (e.g., cables can differ in physical size / geometry, construction technique, material(s), etc.). For instance, a first machine learning model can be trained to detect and / or predict the presence of failure and wear in synthetic fiber or filament cables, and a second machine learning model can be trained to detect and / or predict the presence of failure and wear in steel cables, etc. More generally, different types of cables can exhibit different timelines of failure and wear, e.g., subject to identical loading behavior and environmental conditions, a cable having a first material composition may wear at a different rate and / or in different locations, may fail at a different time and / or in different locations, may experience a different failure type or mode, etc., as compared to a cable having a second material composition.

[0054] As will be described in greater detail below, it is further contemplated herein that the systems and techniques for hyperspectral imaging (HSI)-based cable inspection and analysis can be implemented using various modalities. In some embodiments, HSI-based cable inspection and analysis can be implemented using a handheld device. For instance, the handheld device could be used to conduct fully automated cable inspection and analysis using one or more trained machine learning or artificial intelligence models to analyze hyperspectral image data captured using thehandheld device. Tn other examples, the handheld device could be used to augment a manual cable inspection and analysis process, wherein hyperspectral image data captured using the handheld device is analyzed to generate predictions or indications of areas of interest, areas of potential wear or failure, etc. In some cases, the handheld device implementation could augment manual cable inspections based on generating one or more suggested actions for a user (e.g., an individual conducting a manual inspection of a cable, while using the HSI handheld device). For instance, the handheld device can analyze HSI image data of a cable to identify areas of potential wear or damage and may suggest follow up actions such as capturing additional hyperspectral images from different angles, different distances, etc., to subsequently confirm or refute the presence of the potential / predicted wear or damage. In some embodiments, the HSI handheld device can provide a first pass solution for cable inspection and analysis, wherein the HSI handheld device captures and analyzes hyperspectral images of the cable to detect and localize one or more anomalies (e g., wear or failures) for a subsequent manual inspection to confirm or deny. In some embodiments, the HSI handheld device can be used as a second pass solution for cable inspection and analysis, wherein the HSI handheld device captures and analyzes hyperspectral images of portions of the cable where an earlier manual inspection has flagged wear, damage, failures, etc. In some embodiments, the HSI handheld device can be used in tandem with the conclusions of a manual inspection of the cable, such that the locations of wear, damage, or failure detected by the HSI handheld device analysis is considered in combination with those detected by a conventional manual inspection process.

[0055] In addition to the handheld device modality described above for implementing HSI- based cable inspection and analysis, in another illustrative example the systems and techniques can be used to implement automated HSI-based cable inspection and analysis as a fully external imaging system. For instance, one or more hyperspectral cameras or other hyperspectral imaging devices can be combined with existing approaches to automated vision systems used to conduct cable inspection and / or analysis. In some aspects, a fully automated cable inspection system can utilize one or more sheaves to force (e g., guide) a cable through a known path that is surrounded by one or more imaging devices. The imaging devices that surround the cable as it passes through the known path are used to capture image data of the cable, where the image data can be captured along a full or partial length of the cable. Each captured image can be tagged with its corresponding location on the cable, for example referenced to a fixed location on the cable or a fixed feature onthe cable (e.g., each image can be tagged with its distance / displacement from one of the distal ends of the cable, etc.). Each captured image can additionally be tagged with the corresponding time or date information of its capture, and can be stored for use as a historical example in future inspections or analyses of the same cable. In one illustrative example, the one or more imaging devices that surround the known path through which a cable is fed during an automated inspection can be provided as hyperspectral imaging devices. In some cases, the hyperspectral imaging devices can be the same as one another, and can be used to obtain hyperspectral image data from different angles but using the same spectral bands, wavelength range or sensitivity, etc. In some embodiments, at least a portion of the hyperspectral imaging devices that image the cable as it passes through a known path can be different from one another. For instance, in addition to capturing hyperspectral image data of the cable from different angles, the hyperspectral imaging devices can capture hyperspectral image data using different combinations and ranges of spectral bands, using different wavelength ranges, using different sensitivities, etc.

[0056] In some embodiments, the systems and techniques described herein can use hyperspectral imaging to perform improved cable analysis and / or inspection. For instance, hyperspectral imaging data of a cable can be captured and used to determine one or more characteristics of the cable (including a condition, wear, abnormality, flaw, etc., of the cable). In some aspects, the hyperspectral imaging-based analysis and inspection of cables and cable systems described herein can be based at least in part on one or more spectral properties of the constituent material(s) used to form the cable being inspected. For instance, different material types can be identified in hyperspectral imaging based on the location(s) of absorbance peaks or other spectral peak points. In some embodiments, a ratio of spectral peak points in different spectral bands can be used to distinguish between different material compositions and / or conditions of various ropes and rope systems imaged by a hyperspectral imaging system.

[0057] As noted previously, different material properties and characteristics of a cable can be associated with different hyperspectral imaging information. For instance, the spectral reflectance of a cable that is in a rolled state can be different than the spectral reflectance of the same cable when in an unrolled state. The spectral reflectance of a cable can also vary based on the color of the cable. For instance, a white cable, an orange cable, and a blue cable each may exhibit a different spectral reflectance (e.g., peak points at different wavelengths, different spectral reflectance magnitudes at the peak points and other wavelengths, etc.).

[0058] Notably, hyperspectral image data of a cable can be seen to provide a greater quantity of channel information, as compared to conventional RGB visual light image data (e.g., which comprises three channels of information). As mentioned previously, hyperspectral image data can include a plurality of channels of spectral information for each spatial location (e.g., each x,y pixel location) that is imaged, wherein each channel corresponds to a different range of wavelengths (e.g., a different spectral band). Accordingly, the increased channel information of hyperspectral images can be used by one or more machine learning and / or artificial intelligence models to learn certain wavelength sensitivities that correlate most strongly to performing the detection, prediction, classification, etc., of the condition or state of a cable. The condition or state of a cable, as determined by a machine learning or artificial intelligence model, can include, but is not limited to, the presence (or predicted current or future presence) of abnormalities such as wear, damage, failures, etc.

[0059] In some aspects, machine learning (ML) or artificial intelligence (Al) models and systems can be trained to perform HSLbased cable inspection and analysis using a plurality of training data inputs comprising annotated hyperspectral image data corresponding to cables that exhibit baseline characteristics (e g., no abnormalities) and annotated hyperspectral image data corresponding to cables that exhibit one or more instances of an abnormality (wear, damage, failure, etc.) that the ML or Al model is being trained to classify or detect. In some embodiments, the HSLbased training data can be obtained based on manual cable inspections that are performed using one or more HSI images, or manual cable inspections in which one or more HSI images are captured at the same time (e.g., even if not used to draw the conclusions of the manual cable inspection). The HSI images used to generate the training data can be captured using the HSL based handheld device described previously above for performed hyperspectraLaugmented manual inspections of cables, wherein the annotation(s) for the captured HSI image data comprise the conclusions or determinations made by the manual inspection.

[0060] In some embodiments, an ML or Al model trained to perform cable inspection and analysis based on hyperspectral imaging data can be implemented using the full channel complexity of the hyperspectral images. For instance, the ML or Al model can receive as input the full plurality of hyperspectral image channels (e.g., where the quantity of hyperspectral image channels is greater than three), and can analyze all of the different channel s / spectral bands to identify or recognize patterns indicative of abnormalities such as wear or damage in the cable. Inother embodiments, an ML or Al model can be trained to learn, over time, a subset of hyperspectral channels that are the most relevant or the most determinative in identifying abnormalities such as wear or damage in a particular type of cable. For instance, cables that have a first type of material composition may exhibit wear or damage that can be detected in a first wavelength range of the hyperspectral information. The wavelength range can be continuous or can be discontinuous. Cables that have a different, second type of material composition may exhibit wear or damage that can be detected in a different, second wavelength range of the hyperspectral information. Accordingly, a first ML or Al model can be trained to perform HS based cable inspection and analysis using only the corresponding first wavelength range of input hyperspectral information (or can be trained to give preferential or additional weight to the first wavelength range, while still also evaluating the full wavelength range). Similarly, a second ML or Al model can be trained to perform HSLbased cable inspection and analysis using only the corresponding second wavelength range of input hyperspectral information (or can be trained to give preferential or additional weight to the second wavelength range, while still also evaluating the full wavelength range).

[0061] In some embodiments, the ML or Al models can be trained based on a prior determination or identification of the wavelength ranges (e.g., particular hyperspectral image data channels) of relevance for a particular cable type and / or cable material composition. In another example, the ML or Al models can be trained to learn the most relevant wavelength ranges or particular hyperspectral image data channels for detecting abnormalities for a given cable type and / or material composition. For instance, augmented training or evaluation data can be generated from the full channel complexity hyperspectral image data in the training set, wherein the augmented training or evaluation data include random subsets of the full quantity of hyperspectral channels, until convergence is eventually reached towards the subset of most relevant channels for detecting abnormalities for a given cable type and material composition.

[0062] In some aspects, cable inspection and / or analysis that is performed based on one or more hyperspectral images of the cable can provide better surface penetration of the cable, as compared to conventional and existing means of cable inspection (e.g., which are largely visual inspections performed by humans, or are otherwise based on visible light spectrum images that lack surface penetration). For instance, hyperspectral image data may provide a surface penetration of at least, or approximately, 1-2 mm (e.g., as compared to a surface penetration of approximately 0mm for conventional, visual spectrum or RGB imaging of a cable). In many cases, a surface penetrationof l-2mm may be sufficient for the use case of identifying, detecting, or predicting the presence of one or more abnormality types present within the interior of a cable (e.g., worn, damaged, or failed strands / filaments / fibers, etc.).

[0063] In some embodiments, the improved or additional surface penetration that can be achieved using hyperspectral imaging of a cable can be utilized to perform automated hyperspectral imaging-based inspection and analysis of a cable without opening up the cable or removing tension from the cable. For instance, in the automated external HSI-based inspection modality previously described above, a cable can be passed through a hyperspectral imaging array (or the hyperspectral imaging array can be passed along the exterior of the cable, over the partial or full length of the cable). Where a conventional imaging approach that captures visual spectrum (e.g., RGB images) fails to achieve surface penetration and can provide cable inspection and analysis based only on pattern analysis or recognition in the exterior surface of the cable, the automated HSI-based implementation can provide surface penetration of l-2mm along the full length of the cable being inspected, thereby offering significantly improved accuracy of inspection (e.g., recalling that the majority of cable damage often occurs in the interior of the cable, and may be hidden from the exterior view).

[0064] In some embodiments, the improved or additional surface penetration of the presently disclosed HSI-based approach for cable inspection and analysis can be beneficial in the handheld HSI-based cable inspection modality that was also previously described above. For instance, in the handheld HSI-based cable inspection modality, it is contemplated that a user of the HSI handheld device (e.g., the user being a cable inspector) may open up a cable at one or more locations in order to obtain visibility of the interior of the cable. The improved or additional surface penetration of the HSI-based approach is beneficial in this modality as well, and can be seen to provide a more complete and thorough characterization of the portion of the cable interior that is exposed during the manual inspection. For instance, by penetrating below the portion of the cable interior that is exposed when the cable is opened up, the hyperspectral image data can be leveraged to significantly increase the total percentage of the cable interior that is imaged and analyzed. If the opening up of the cable exposes some surface area S of the cable interior, a purely manual or visual inspection will consider only this exposed interior surface area S when assessing the cable condition or looking for abnormalities. By comparison, the use of hyperspectral image data with asurface penetration of 2mm can increase the interior cable inspection to cover a volume of 5*2mm, thereby offering an improvement over the existing approaches.

[0065] In some examples, certain wavelengths of the hyperspectral image may provide visibility of interior layers or fibers of a cable being imaged, based on the interior layer(s) of the cable having different spectra than the exterior layer(s) of the cable. In other words, when the spectra of the interior layers of a cable are distinguishable or differentiable from the spectra of the exterior layers of a cable, hyperspectral imaging of the cable may provide at least partial surface penetration beyond the exterior layers of the cable. By providing at least partial visibility of the cable interior, cable inspections performed based on hyperspectral images of the cable may characterize properties of the interior layers of the cable in a non-destructive fashion (e.g., without requiring the physical separation or unwinding of the outer layers to reveal the inner layers, as would be conventionally required, and / or without requiring a sample of the cable to be cut out and the free ends of the cable at the cut location spliced back together).

[0066] In particular, the exterior or outer layers (e.g., fibers) of a cable often mask or hide problems with the cable that would otherwise remain undetected when subjecting the cable to a conventional visual or visible light inspection. In some embodiments, the systems and techniques described herein can automatically capture or generate hyperspectral images of cables based on one or more known or pre-determined properties of the cable. For instance, the hyperspectral imaging range, the quantity or width of the spectral bands used, etc., may be adjusted based on known material properties of the cable in order to achieve deeper surface penetration of the cable in the resulting hyperspectral image. In one illustrative example, one or more hyperspectral imaging parameters can be adjusted based on the material type of the outer layers or fdaments comprising the cable being imaged. In another example, one or more hyperspectral imaging parameters can be adjusted based on a thickness or density of the filaments that comprise the outer / exterior layer of the cable being imaged.

[0067] Hyperspectral imaging of cables can be used by the systems and techniques described herein to perform improved early detection of wear, damage, or failure (or improved early detection of potential / impending wear, damage, or failure) of a cable. Hyperspectral imaging of cables can also be used to detect various abnormalities that otherwise may go undetected in conventional approaches to cable inspection, whether conventional manual visual inspections conducted by individuals and / or in automated visual inspections conducted using visual spectrumcameras and images. Notably, hyperspectral imaging provides a plurality of spectral channels that capture spectral information which is indicative of material properties of the object (e.g., cable) being imaged by the HSI system. In particular, many of the abnormalities of interest in cable inspection and analysis (e.g., wear, damage, failure, etc.) manifest as changes to a material property of the cable. For instance, cable wear can be understood as a form of material fatigue, and as a material fatigues, its corresponding material properties may change as well. Hyperspectral images can reveal this change in material properties of a cable, which can be used to drive an inference or correlation to a matching type of abnormality of concern or of interest. The analysis can be performed based on comparing a current hyperspectral image data of a cable to past, historical examples of hyperspectral image data of the same cable or of similar cables. The analysis can additionally, or alternatively, be performed based on comparing a current hyperspectral image data of a cable to a pre-determined baseline example (or set of baseline examples) of hyperspectral image data corresponding to a same or similar type of cable at different stages of wear or fatigue.

[0068] In one illustrative example, the hyperspectral-based analysis of cable wear can be performed independent of a direct reference point of comparison. For instance, certain types of abnormalities or defects in a cable (e g., certain types of material fatigue, and thus, certain types of changed material properties) may have a corresponding hyperspectral “signature” that can be detected in the hyperspectral image data and used to identify the presence of the abnormality or defect in the current cable undergoing imaging. For example, many cables are constructed as organized groupings (and further sub-groupings) of many smaller cables. As mentioned previously above, the smaller cables can be referred to as strands. Wear, damage, and / or fatigue of a stranded cable often presents as a breakage of some percentage of the constituent strands that form the larger stranded cable. Overtime, some of the individual strands may become abraded, broken or snapped, stretched, cracked, etc. As a cable ages, there may be a series of acceptable thresholds of damaged strands within the larger overall construction. In other examples, there may be no acceptable threshold of damaged strands within the larger overall construction. In either case, it is desirable to detect and quantify the presence of any damaged strands within the plurality of strands that form a larger cable as a whole. According to aspects of the present disclosure, hyperspectral imaging of the cable can be used to identify the presence of damaged strands within a cable based on detecting a unique hyperspectral signature corresponding to the damaged strands. In particular, damage to a cable (or strands thereof) causes a change in the physical or material morphology of the cable,which often corresponds to a distinct and recognizable change in the corresponding hyperspectral information of the cable when imaged by an HSI device or camera.

[0069] For instance, the two broken ends of a strand that has snapped can present a significantly different hyperspectral signature as compared to the hyperspectral signature of a single, continuous strand (e.g., a non-broken or non-snapped strand). This may be based at least in part on the fact that the strand breaking or snapping corresponds to a massive release of mechanical stress, which can be sufficient to cause a pronounced change in the material properties, and therefore hyperspectral presentation or appearance, of the broken strand(s). This principle can be understood as being visually similar to the difference in appearance of a fiber optic strand viewed from the side vs. being viewed head-on - the exposed interior of the strand at the surface faces of the break appears differently than the exterior surface. By achieving a surface penetration of at least l-2mm below the exterior of a cable, the hyperspectral imaging described herein can be used to detect abnormalities such as the broken or damaged strands through a greater percentage of the total volume of the cable, and more notable, can detect broken strands in the location (e.g., interior of the cable) where they are significantly more likely to occur (e.g., significantly more likely to occur in the cable interior vs. the cable exterior).

[0070] Hyperspectral signatures, or differences in spectral properties of cable abnormalities, can extend beyond the example described above, without departing from the scope of the present disclosure. For instance, failure or wear can appear differently in cables made of different materials. In addition to broken strands, failure or wear of a cable can present as a “fuzzy” appearance of the cable and / or one or more constituent strands of the cable, can present as “glazing” of the cable and / or one or more constituent strands of the cable, etc. Some (or all) of the different types of failure modes or failure presentations of various cables can be associated with a learned hyperspectral signature for the failure mode or failure presentation, with the learned hyperspectral signature used to identify and detect the occurrence of that failure mode or failure presentation in other cables having a same or similar composition and design.

[0071] In some embodiments, the hyperspectral images can be obtained using one or more hyperspectral imaging devices that are associated with a cable analysis and inspection system. In some cases, a hyperspectral imaging device can be automatically controlled to obtain one or more hyperspectral images of a cable. In another example, a hyperspectral imaging device can be manually operated to obtain one or more hyperspectral images of a cable. For instance, thehyperspectral imaging device can be provided as a handheld imaging device that can be used by an individual tasked with performing a cable inspection. In some embodiments, hyperspectral images can be obtained from multiple different angles or perspectives of a cable. For instance, hyperspectral images can be obtained for one or more (or all) of a ‘front’ view of the cable, a ‘back’ view of the cable, a ‘left’ view of the cable, a ‘right’ view of the cable, etc. In some embodiments, hyperspectral images can be obtained for multiple portions or locations along the longitudinal length of the cable. In some examples, one or more hyperspectral images can be obtained for a majority of the longitudinal length of the cable. In another example, one or more hyperspectral images can be obtained for the full longitudinal length of the cable.

[0072] In some embodiments, the one or more hyperspectral images obtained for a cable can be automatically provided to or otherwise ingested by a cable inspection and analysis system. For instance, the cable inspection and analysis can be the same as or similar to the cable analysis system(s) described in commonly owned PCT Application No. PCT / US2023 / 061498, the contents of which are hereby incorporated by reference in their entirety and for all purposes.

[0073] In one illustrative example, some (or all) of the hyperspectral images obtained for a given cable being inspected can be automatically analyzed and / or inspected to determine a condition or state of the cable. For example, the automatic analysis can use the one or more hyperspectral images obtained for a given cable to determine a wear state of the cable, a remaining or predicted life of the cable, a defect or flaw present in the cable, etc. In some aspects, the automatic analysis can use the one or more hyperspectral images to identify potential or probable flaws or degradations of the cable, along with a corresponding location of each identified flaw. In some cases, the hyperspectral images can be used to automatically generate an inspection report or identified flaws within the cable. In another example, the hyperspectral images can be used to automatically identify regions of interest and / or potential (or probable) flaws and their corresponding location, and the relevant portion(s) of the hyperspectral image data for the given cable can subsequently be provided to a manual reviewer for further analysis and confirmation of the potential flaw.

[0074] In one illustrative example, the hyperspectral images obtained for a given cable under inspection can be analyzed and used to detect damage such as broken or frayed strands (e.g., filaments) on or within the cable. In some embodiments, broken strands / filaments of a cable may present a different spectral signature (e.g., in the hyperspectral imaging data of the cable) ascompared to non-broken strands / fil aments of the cable. For example, broken fdaments may present two loose ends (e.g., one on either end of the break) while non-broken filaments may present a continuous and relatively uniform body. Accordingly, spectral differences associated with hyperspectral imaging data of broken, frayed, and / or worn filaments (as compared to baseline spectra associated with non-broken filaments) can be used to identify or detect damage to a cable. In some embodiments, the systems and techniques can utilize a database of hyperspectral image data corresponding to various different combinations of cables (e.g., cable type / properties, material composition, deployed environment, etc.) and cable conditions (e.g., broken filaments, frayed or worn filaments, external damage, internal damage, abrasion damage, impact or compressive damage, etc.) can be analyzed against hyperspectral image data obtained for a given cable that is currently being inspected. In some cases, one or more similarities can be identified (e.g., based on the analysis) between the hyperspectral image data of the currently inspected cable and respective hyperspectral image data included in the database of historical examples. Based on the cable condition(s) depicted in the matching hyperspectral image data from the database, the systems and techniques can automatically generate an identification of detected flaws or damage present in the cable and / or can automatically generate a prediction or warning of potential flaws or damage present in the cable. In one illustrative example, one or more machine learning networks can be used to perform the automatic cable inspection and analysis based on hyperspectral imaging data of the cable, as described herein and above. For instance, the one or more machine learning networks can be trained based at least in part on the database of historical examples of hyperspectral imaging data corresponding to different combinations of cable type and condition type. In some aspects, each hyperspectral image included in the historical database can be labeled (e.g., annotated) with information indicative of the cable type and cable condition depicted in the hyperspectral images - the combination of a historical hyperspectral image data and the corresponding label(s) can comprise a training data pair. The one or more machine learning networks can be trained to perform various classification and / or detection tasks for inputs comprising hyperspectral imaging data of a cable, based on training the one or more machine learning networks using a plurality of labeled training data pairs generated based on the historical database, as described above.

[0075] As noted previously above, hyperspectral image data can be associated with a plurality of different spectral bands. In some embodiments, the systems and techniques can performinspection and analysis of a cable based on analyzing some (or all) of the spectral bands included in hyperspectral image data of the cable independently from one another. For example, flaws or damage to the cable can be detected based on analyzing a particular subset of the spectral bands where features or characteristics indicative of the presence of different types of cable flaws or damage may be present. In another illustrative example, the systems and techniques described herein can perform inspection and analysis of a cable based on analyzing some (or all) of the spectral bands included in hyperspectral image data of the cable in combination or aggregate. For example, the systems and techniques can analyze the stacked hyperspectral image (e.g., 3D hyperspectral data cube) comprising the same (x, y) spatial dimensions stacked over a plurality of different spectral bands d.

[0076] As noted previously, hyperspectral image data can be indicative of the wavelengths (and intensity) of light that are emitted by an imaged object. Accordingly, the systems and techniques can use hyperspectral image data to perform cable inspection and analysis for discrete wavelength ranges (e.g., the different spectral bands of the hyperspectral image) and / or for continuous wavelength ranges (e.g., some or all of the continuous hyperspectral imaging spectrum, which extends to wavelengths shorter and longer than those associated with the visible light spectrum).

[0077] Any of the steps, operations, functions, or processes described herein may be performed or implemented by a combination of hardware and software services or services, alone or in combination with other devices. In some embodiments, a service can be software that resides in memory of a client device and / or one or more servers of a content management system and perform one or more functions when a processor executes the software associated with the service. In some embodiments, a service is a program or a collection of programs that carry out a specific function. In some embodiments, a service can be considered a server. The memory can be a non-transitory computer-readable medium.

[0078] In some embodiments, the computer-readable storage devices, mediums, and memories can include a cable or wireless signal containing a bit stream and the like. However, when mentioned, non-transitory computer-readable storage media expressly exclude media such as energy, carrier signals, electromagnetic waves, and signals per se.

[0079] Methods according to the above-described examples can be implemented using computer-executable instructions that are stored or otherwise available from computer-readablemedia. Such instructions can comprise, for example, instructions and data which cause or otherwise configure a general-purpose computer, special purpose computer, or special purpose processing device to perform a certain function or group of functions. Portions of computer resources used can be accessible over a network. The executable computer instructions may be, for example, binaries, intermediate format instructions such as assembly language, firmware, or source code. Examples of computer-readable media that may be used to store instructions, information used, and / or information created during methods according to described examples include magnetic or optical disks, solid-state memory devices, flash memory, USB devices provided with non-volatile memory, networked storage devices, and so on.

[0080] Devices implementing methods according to these disclosures can comprise hardware, firmware and / or software, and can take any of a variety of form factors. Typical examples of such form factors include servers, laptops, smartphones, small form factor personal computers, personal digital assistants, and so on. The functionality described herein also can be embodied in peripherals or add-in cards. Such functionality can also be implemented on a circuit board among different chips or different processes executing in a single device, by way of further example.

[0081] The instructions, media for conveying such instructions, computing resources for executing them, and other structures for supporting such computing resources are means for providing the functions described in these disclosures.

[0082] Illustrative aspects of the disclosure include:

[0083] Aspect 1. A method comprising: obtaining one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyzing the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

[0084] Aspect 2. The method of Aspect 1, wherein: each hyperspectral image of the one or more hyperspectral images is associated with a hyperspectral wavelength range that is wider than a respective wavelength range associated with a visible light spectrum.

[0085] Aspect 3. The method of Aspect 2, wherein the plurality of different spectral bands are continuous spectral bands of the hyperspectral wavelength range.

[0086] Aspect 4. The method of any of Aspects 1 to 3, wherein each hyperspectral image is obtained for a different location along a longitudinal length of the cable.

[0087] Aspect 5. The method of any of Aspects 1 to 4, wherein each hyperspectral image is obtained for a same location along a longitudinal length of the cable.

[0088] Aspect 6. The method of any of Aspects 1 to 5, wherein analyzing the one or more hyperspectral images comprises analyzing the respective spectral information for the plurality of different spectral bands in combination.

[0089] Aspect 7. The method of any of Aspects 1 to 6, wherein analyzing the one or more hyperspectral images comprises analyzing at least a portion of the respective spectral information for the plurality of different spectral bands independently from a remaining portion of the respective spectral information for the plurality of different spectral bands.

[0090] Aspect 8. The method of any of Aspects 1 to 7, wherein: a first state of wear of the cable or a first type of damage to the cable is associated with a first pattern of spectral information in a particular spectral band.

[0091] Aspect 9. The method of Aspect 8, further comprising analyzing the one or more hyperspectral images to detect the first state of wear of the cable or the first type of damage to the cable.

[0092] Aspect 10. The method of Aspect 9, wherein the analyzing is based on comparing the respective spectral information for the particular spectral band to the first pattern of spectral information associated with the first state of wear of the cable or the first type of damage to the cable.

[0093] Aspect 11. An apparatus, the apparatus comprising: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor being configured to: obtain one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyze the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

[0094] Aspect 12. The apparatus of Aspect 11, wherein: each hyperspectral image of the one or more hyperspectral images is associated with a hyperspectral wavelength range that is wider than a respective wavelength range associated with a visible light spectrum.

[0095] Aspect 13. The apparatus of Aspect 12, wherein the plurality of different spectral bands are continuous spectral bands of the hyperspectral wavelength range.

[0096] Aspect 14. The apparatus of any of Aspects 11 to 13, wherein each hyperspectral image is obtained for a different location along a longitudinal length of the cable.

[0097] Aspect 15. The apparatus of any of Aspects 11 to 14, wherein each hyperspectral image is obtained for a same location along a longitudinal length of the cable.

[0098] Aspect 16. The apparatus of any of Aspects 11 to 15, wherein, to analyze the one or more hyperspectral images, the at least one processor is configured to analyze the respective spectral information for the plurality of different spectral bands in combination.

[0099] Aspect 17. The apparatus of any of Aspects 11 to 16, wherein, to analyze the one or more hyperspectral images, the at least one processor is configured to analyze at least a portion of the respective spectral information for the plurality of different spectral bands independently from a remaining portion of the respective spectral information for the plurality of different spectral bands.

[0100] Aspect 18. The apparatus of any of Aspects 11 to 17, wherein a first state of wear of the cable or a first type of damage to the cable is associated with a first pattern of spectral information in a particular spectral band.

[0101] Aspect 19. The apparatus of Aspect 18, wherein the at least one processor is further configured to analyze the one or more hyperspectral images to detect the first state of wear of the cable or the first type of damage to the cable, the analyzing based on comparing the respective spectral information for the particular spectral band to the first pattern of spectral information associated with the first state of wear of the cable or the first type of damage to the cable.

[0102] Aspect 20. A non-transitory computer-readable medium having stored thereon instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising: obtaining one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyzing the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

[0103] Aspect 21. The non-transitory computer-readable medium of Aspect 20, wherein: each hyperspectral image of the one or more hyperspectral images is associated with a hyperspectral wavelength range that is wider than a respective wavelength range associated with a visible light spectrum.

[0104] Aspect 22. The non-transitory computer-readable medium of Aspect 21, wherein the plurality of different spectral bands are continuous spectral bands of the hyperspectral wavelength range.

[0105] Aspect 23. The non-transitory computer-readable medium of any of Aspects 20 to 22, wherein each hyperspectral image is obtained for a different location along a longitudinal length of the cable.

[0106] Aspect 24. The non-transitory computer-readable medium of any of Aspects 20 to 23, wherein each hyperspectral image is obtained for a same location along a longitudinal length of the cable.

[0107] Aspect 25. The non-transitory computer-readable medium of any of Aspects 20 to 24, wherein analyzing the one or more hyperspectral images comprises analyzing the respective spectral information for the plurality of different spectral bands in combination.

[0108] Aspect 26. The non-transitory computer-readable medium of any of Aspects 20 to 25, wherein analyzing the one or more hyperspectral images comprises analyzing at least a portion of the respective spectral information for the plurality of different spectral bands independently from a remaining portion of the respective spectral information for the plurality of different spectral bands.

[0109] Aspect 27. The non-transitory computer-readable medium of any of Aspects 20 to 26, wherein: a first state of wear of the cable or a first type of damage to the cable is associated with a first pattern of spectral information in a particular spectral band.

[0110] Aspect 28. The non-transitory computer-readable medium of Aspect 27, further comprising analyzing the one or more hyperspectral images to detect the first state of wear of the cable or the first type of damage to the cable.

[0111] Aspect 29. The non-transitory computer-readable medium of Aspect 28, wherein the analyzing is based on comparing the respective spectral information for the particular spectral band to the first pattern of spectral information associated with the first state of wear of the cable or the first type of damage to the cable.

[0112] Aspect 30. An apparatus comprising means for performing any of the operations of Aspects 1 to 10.

[0113] Aspect 31. An apparatus comprising means for performing any of the operations of Aspects 11 to 19.

[0114] Aspect 32. An apparatus comprising means for performing any of the operations of Aspects 20 to 29.

Claims

CLAIMSWhat is claimed is:

1. A method comprising: obtaining one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyzing the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

2. The method of claim 1, wherein: each hyperspectral image of the one or more hyperspectral images is associated with a hyperspectral wavelength range that is wider than a respective wavelength range associated with a visible light spectrum.

3. The method of claim 2, wherein the plurality of different spectral bands are continuous spectral bands of the hyperspectral wavelength range.

4. The method of claim 1, wherein each hyperspectral image is obtained for a different location along a longitudinal length of the cable.

5. The method of claim 1, wherein each hyperspectral image is obtained for a same location along a longitudinal length of the cable.

6. The method of claim 1, wherein analyzing the one or more hyperspectral images comprises analyzing the respective spectral information for the plurality of different spectral bands in combination.

7. The method of claim 1, wherein analyzing the one or more hyperspectral images comprises analyzing at least a portion of the respective spectral information for the plurality of different spectral bands independently from a remaining portion of the respective spectral information for the plurality of different spectral bands.

8. The method of claim 1, wherein: a first state of wear of the cable or a first type of damage to the cable is associated with a first pattern of spectral information in a particular spectral band.

9. The method of claim 8, further comprising analyzing the one or more hyperspectral images to detect the first state of wear of the cable or the first type of damage to the cable.

10. The method of claim 9, wherein the analyzing is based on comparing the respective spectral information for the particular spectral band to the first pattern of spectral information associated with the first state of wear of the cable or the first type of damage to the cable.

11. An apparatus, the apparatus comprising: at least one memory; and at least one processor coupled to the at least one memory, the at least one processor being configured to: obtain one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyze the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

12. The apparatus of claim 11, wherein: each hyperspectral image of the one or more hyperspectral images is associated with a hyperspectral wavelength range that is wider than a respective wavelength range associated with a visible light spectrum.

13. The apparatus of claim 12, wherein the plurality of different spectral bands are continuous spectral bands of the hyperspectral wavelength range.

14. The apparatus of claim 11 , wherein each hyperspectral image is obtained for a different location along a longitudinal length of the cable.

15. The apparatus of claim 11 , wherein each hyperspectral image is obtained for a same location along a longitudinal length of the cable.

16. The apparatus of claim 11, wherein, to analyze the one or more hyperspectral images, the at least one processor is configured to analyze the respective spectral information for the plurality of different spectral bands in combination.

17. The apparatus of claim 11, wherein, to analyze the one or more hyperspectral images, the at least one processor is configured to analyze at least a portion of the respective spectral information for the plurality of different spectral bands independently from a remaining portion of the respective spectral information for the plurality of different spectral bands.

18. The apparatus of claim 11, wherein a first state of wear of the cable or a first type of damage to the cable is associated with a first pattern of spectral information in a particular spectral band.

19. The apparatus of claim 18, wherein the at least one processor is further configured to analyze the one or more hyperspectral images to detect the first state of wear of the cable or the first type of damage to the cable, the analyzing based on comparing the respective spectral information for the particular spectral band to the first pattern of spectral information associated with the first state of wear of the cable or the first type of damage to the cable.

20. A non-transitory computer-readable medium having stored thereon instructions which, when executed by one or more processors, cause the one or more processors to perform operations comprising: obtaining one or more hyperspectral images associated with a cable, each hyperspectral image including respective spectral information for a plurality of different spectral bands; and analyzing the one or more hyperspectral images to determine information indicative of one or more of a state of wear of the cable or damage to the cable.

Images