Estimating cleave quality of hollow core fibre

Abstract

An apparatus comprising a processor and a memory storing instructions that, when executed by the processor, perform a method for estimating cleave quality of a cleaved hollow core optical fibre, is described. The method comprises receiving an image of an end face of the fibre, and analysing pixel data of the image to determine cleave quality of the fibre by: determining at least one feature of the image, the feature representing a characteristic of the end face of the fibre, and using a model to map the at least one feature to an indication of cleave quality. The method further comprises outputting the indication of cleave quality. A method for creating a model mapping at least one feature of an image of an end face of a cleaved hollow core optical fibre to an indication of cleave quality of the fibre is disclosed. The indication of cleave quality comprises an indication of at least one of cleave angle and cleave profile. The method comprises at least one of training a machine-learning model using training inputs, defining at least one rule mapping at least one feature of an end face image of a cleaved hollow core optical test fibre and defining a lookup table by associating at least one feature of an end face image of a cleaved hollow core optical test fibre.

Description

Docket No. 13768-4558a_502046-PCT01ESTIMATING CLEAVE QUALITY OF HOLLOW CORE FIBREBACKGROUND

[0001] Splicing is typically used to join optical fibres, and there are various splicing methods that may be used, including fusion splicing and mechanical splicing. Prior to splicing, an optical fibre is cleaved, resulting in an end face of the fibre having a cleave angle. Splicing involves aligning the fibres end-to-end and fixing them in the aligned position. Fusion splicing, for example, joins the two fibres by heating the end region in order to soften the glass from which the fibres are made. By pressing the ends together, the softened glass is made to fuse so that the fibres are permanently connected when the glass cools and hardens. Mechanical splicing, in contrast, does not permanently join the fibres together but instead uses a mechanical arrangement to maintain their aligned position and hold the fibre ends together.

[0002] The join formed by splicing two fibres together is referred to as a splice. The quality of the splice is an important factor in enabling low loss optical propagation for light travelling from one fibre to the other. Accurate alignment of structural features within the two fibres so as to reduce structural discontinuities at the splice contributes to low loss.

[0003] Conventional solid core optical fibres, comprising an annular cladding surrounding a single circular core, are relatively simple to align for splicing. The structures have continuous rotational symmetry' in transverse cross-section so that transverse alignment of the fibre ends to match the positions of the longitudinal axes of the fibres necessarily aligns the cores and the cladding. However, hollow core fibres such as antiresonant hollow core fibres have a complex internal structure that lacks continuous rotational symmetry- and multi-core solid core fibres also lack continuous rotational symmetry-. Misalignment of this internal structure and / or of the individual cores increases the optical loss that occurs when light travels from one fibre to the other across the splice.

[0004] The embodiments described below are not limited to implementations which solve any or all of the disadvantages of know n fibre cleaving, splicing and analysis apparatus. SUMMARY

[0005] The following presents a simplified summary of the disclosure in order to provide a basic understanding to the reader. This summary is not intended to identify key features or essential features of the claimed subject matter nor is it intended to be used to limit the scope of the claimed subject matter. Its sole purpose is to present a selection ofDocket No. 13768-4558a_502046-PCT01 concepts disclosed herein in a simplified form as a prelude to the more detailed description that is presented later.

[0006] An apparatus comprising a processor and a memory7storing instructions that, when executed by the processor, perform a method for estimating cleave quality of a cleaved hollow core optical fibre, is described. The method comprises receiving an image of an end face of the fibre, and analysing pixel data of the image to determine cleave quality of the fibre by: determining at least one feature of the image, the feature representing a characteristic of the end face of the fibre, and using a model to map the at least one feature to an indication of cleave quality. The method further comprises outputting the indication of cleave quality.

[0007] Many of the attendant features will be more readily appreciated as the same becomes better understood by reference to the following detailed description considered in connection with the accompanying drawings.DESCRIPTION OF THE DRAWINGS

[0008] The present description will be better understood from the following detailed description read in light of the accompanying drawings, wherein:FIGS. 1-3 illustrate transverse cross-sectional views of three different examples of antiresonant hollow core optical fibres;FIG. 4 illustrates a transverse cross-sectional view of a cleaved antiresonant hollow core optical fibre;FIG. 5A illustrates a schematic diagram of an approach to determining a cleave angle of a fibre using a side-view camera;FIG. 5B illustrates a chart showing variation in determined cleave angle using the apparatus of FIG. 5 A;FIGS. 6A and 6B illustrate a flow diagram of a method according to the herein disclosed technology;FIG. 7 illustrates a schematic diagram of a system according to the herein disclosed technology;FIG. 8 illustrates a schematic diagram of a system according to the herein disclosed technology, the system associated with a cleaver;FIG. 9 illustrates a transverse cross-sectional view of a cleaved antiresonant hollow core optical fibre that is analysed according to the herein disclosed technology, and an exemplary chart illustrating the analysis;FIG. 10 illustrates a chart showing a correlation between angular tear width and cleave profile for cleaved hollow core fibres;Docket No. 13768-4558a_502046-PCT01FIG. 11 illustrates a flow diagram of a method according to the herein disclosed technology;FIG. 12A illustrates a transverse cross-sectional view of a cleaved antiresonant hollow core optical fibre with damage that is analysed according to the herein disclosed technology;FIG. 12B illustrates a flow diagram of a method including a quality control check according to the herein disclosed technology;FIG. 13 illustrates a flow diagram of a method for creating a model according to the herein disclosed technology; andFIG. 14 illustrates an exemplary computing-based device in which embodiments of a cleave quality estimation method according to the herein disclosed technology are implemented.Like reference numerals are used to designate like parts in the accompanying drawings. DETAILED DESCRIPTION

[0009] The detailed description provided below in connection with the appended drawings is intended as a description of the present examples and is not intended to represent the only forms in which the present examples are constructed or utilized. The description sets forth the functions of the examples and the sequence of operations for constructing and operating the examples. However, the same or equivalent functions and sequences may be accomplished by different examples.

[0010] FIGS. 1-3 illustrate transverse cross-sectional views of three different examples of antiresonant hollow core optical fibres (ARFs) 100, 200. 300. Light is guided in these fibres by an antiresonant optical effect. Each of the fibres 100, 200, 300 comprises a tubular outer cladding (or jacket) 102, a structured, inner, cladding (microstructure) comprising a plurality7of tubular cladding capillaries 104, 204, 304 and a hollow core 106. The outer cladding 102 has a glass thickness that is typically much larger than that of the cladding capillaries 104, 204, 304. In the first example, shown in FIG. 1, the structured, inner, cladding comprises five capillaries 104 of the same cross-sectional size and shape, which are arranged inside the outer cladding 102 in a single ring so that the longitudinal axes of each cladding capillary 104 and of the outer cladding 102 are substantially parallel. Each cladding capillary 104 is in contact with (e.g. bonded to) the inner surface of the outer cladding 102 at an azimuthal location 108, such that the cladding capillaries 1 4 are usually evenly spaced around the inner circumference of the outer cladding 102, and are also spaced apart from each other by gaps 110 (i.e. such that there is no contact between neighbouringDocket No. 13768-4558a_502046-PCT01 capillaries). In some designs of ARF, the cladding capillaries 104 may be positioned in contact with each other (in other words, not spaced apart as in FIG. 1), but spacing to eliminate this contact improves the fibre’s optical performance. The gaps 110 remove nodes that arise at the contact points between adjacent capillaries and which tend to cause undesirable spectral resonances that result in high transmission losses. Accordingly, fibres with spaced-apart cladding capillaries may be referred to as “nodeless antiresonant hollow core fibres”.

[0011] The arrangement of the cladding capillaries 104 in a ring around the inside of the tubular outer cladding 102 creates a central space, cavity, or void within the fibre, also with its longitudinal axis parallel to those of the outer cladding 102 and the cladding capillaries 104, which is the fibre’s hollow core 106. The hollow core 106 is bounded by the inwardly facing parts of the outer surfaces of the cladding capillaries 104. This is the core boundary, and the material (glass or polymer, for example) of the capillary walls that make up this boundary contributes to the antiresonance optical guidance effect or mechanism. The antiresonance optical guidance effect is provided by reflections of light in the hollow core by the core boundary, and further reflections of light in the capillary wall by a capillary boundary between a wall of the capillary and an interior space, cavity or void encompassed by the capillary. The cladding capillaries 104 have a thickness, t, at the core boundary which defines the wavelengths for which antiresonant optical guiding occurs in the ARF.

[0012] In the second example, shown in FIG. 2, each primary cladding capillary 104 has a secondary', smaller capillary 204 nested inside it, bonded to the inner surface of the primary cladding capillary 104, in this example, but not necessarily, at the same azimuthal location 108 as the point of bonding between the primary cladding capillary 104 and the outer cladding 102. These additional smaller capillaries 204 can reduce the optical loss. ARF designs of this type, with secondary capillaries, may be referred to as “nested antiresonant nodeless fibres” (NANFs) (TM).

[0013] The third example, shown in FIG. 3, has two smaller cladding capillaries 204, 304 nested inside each cladding capillary 104. As with the example shown in FIG. 2, each of the smaller capillaries 204. 304 is bonded to the inner surface of the immediately larger capillary at an azimuthal location that is in some cases the point of bonding between the primary cladding capillary 104 and the outer cladding 102. In this example, the smaller capillary 204 may be referred to as the secondary cladding capillary' and the smallest capillary 304 may be referred to as the tertiary' cladding capillary'. The tertiary' cladding capillary' 304 is bonded to the inner surface of the secondary cladding capillary 204 and the secondaryDocket No. 13768-4558a_502046-PCT01 cladding capillary 204 is bonded to the inner surface of the primary cladding capillary 104. ARF designs of this type, with secondary and tertiary cladding capillaries may be referred to as ‘‘double-nested antiresonant nodeless fibres” (DNANFs). In yet further examples (not shown in the drawings) there may be a different configuration of cladding capillaries. For example, there may be smaller further capillaries, within the tertiary capillary 304 to provide further levels of nesting and / or there may be a plurality of secondary cladding capillaries within each primary cladding capillary, each secondary cladding capillary being bonded to the inner surface of the primary cladding capillary at a different azimuthal location and / or each primary cladding capillary may have an internal structure (e g. one or more dividing walls).

[0014] All of the examples show n in FIGs. 1-3 comprise five primary cladding capillaries 104 and hence have five-fold rotational symmetry7. Furthermore, all the cladding capillaries are circular in cross-section. In other examples, there may be a different number of pnmary cladding capillaries surrounding the core (e.g. four, six, seven, eight, nine or ten) and / or the cladding capillaries may not be of circular cross-section. Additionally, whilst in the examples of FIGs. 1-3, all the primary7cladding capillaries 104 are of the same size and shape, in other examples, the primary cladding capillaries w ithin the outer cladding 102 may not all be the same size and / or shape. Whilst ARFs are described and illustrated, the present technology further relates to hollow core fibres, such as those w ith or w ithout microstructure present interior to an outer cladding.

[0015] Prior to splicing two lengths of ARF together, the lengths of ARF to be spliced are produced by cleaving a length of ARF. Cleaving refers to performing a controlled break of a fibre, and results in an end face of the ARF, or more broadly, of a hollow core fibre, with a cleave profile.

[0016] The topology of the end-face of a cleaved hollow core fibre is very different from that of a cleaved all-solid fibre. When using a tension-based cleaver, a blade hits a tensioned fibre, and a crack will propagate transversely across the fibre from the point of impact of the blade. Since the crack cannot propagate into the central void of the hollow-core fibre, it is divided and travels around the void. When the cracks from both sides of the central void of the hollow core fibre meet, the fibre does not break smoothly and a tear is created on the end face of the hollow-core fibre. In some cases the tear is a step in the end-face. In some cases, a highest and lowest point of the end face surface are either side of the step. Other cleaver ty pes, such as score and bend cleavers, in various examples result in a different shapeDocket No. 13768-4558a_502046-PCT01 of the end face of the hollow core fibre compared to a tension-based cleaver, and in some cases the end face surface of a cleaved hollow core fibre does not comprise a tear.

[0017] In a solid fibre, in contrast to a hollow core fibre, the crack from cleaving can propagate through the fibre unobstructed. Therefore, a tear is not present. In some cases, the highest and lowest points of the end-face surface are opposite each other on the end face. The direction of the normal of the (flat) fibre end face can vary from the fibre axis, enabling a cleave angle to be defined. A cleave angle is defined herein in some cases as a difference between a completely flat surface of an end face of the fibre which is perpendicular to the longitudinal axis (which corresponds to the optical axis) of the fibre (a cleave angle of zero degrees, i.e. between cleaved end-face surface normal and fibre axis) and an angle of a surface of the end face of the fibre after cleaving.

[0018] Where a step is not present in the end-face of a fibre, such as when a solid fibre is cleaved, a cleave angle is simply defined as mentioned above, i.e. being a difference between a cleaved end-face surface normal and fibre axis, where the cleaved end-face surface normal is simply derived as for a substantially flat surface the normal is substantially at the same angle from the central axis of the fibre when the normal is evaluated anywhere on the end face surface. In other cases, such as when a step is present for example across a tear in an end face of a hollow core fibre, the end-face surface normal is difficult to determine given that the cleaved end face is not flat. In various examples, a normal at a single point of the cleaved end face surface is determined, or an average normal across a plurality of points of the cleaved end face surface is determined, to define the cleave angle associated with the cleaved end-face surface.

[0019] However, in other cases, a cleave profile (or an indication of a cleave profile) for a cleaved fibre is instead determined, which represents i.e. comprises, indicates and / or refers to, in various cases, an indication of cleave angle, for example across a cleaved end face. A cleave profile indicates a shape of a cleaved end face surface, and in some cases additionally comprises an indication of a step height of a step in the cleaved end face surface. The cleave profile in some cases comprises an indication of a cleave angle determined based on a normal to the cleaved end face at a single point on the cleaved end face surface, an indication of cleave angles each determined based on a normal at different points on the cleaved end face surface, and / or an indication of an average cleave angle based on an average normal to the cleaved end face across a plurality7of points of the end face surface. In some cases, a cleave profile has an associated numerical value, for example being an indicated average cleave angle (or in other cases being any value defined based on a determination ofDocket No. 13768-4558a_502046-PCT01 an indication of cleave angle such as via one of the above mentioned methods). In other cases, a cleave profile does not have an associated numerical value and instead is a label. In various examples, the indication of cleave profile is whether the cleave profile is acceptable or not for splicing in that the associated fibre with the cleave profile, when spliced with another fibre with an acceptable cleave profile, would be suitable for a desired application, i.e. having a suitable optical propagation loss across the splice.

[0020] Cleaving in some cases is performed by scoring a fibre and then applying tension to the fibre or flexing the fibre. The presence of the hollow core and microstructure in an ARF, as shown in the examples in FIGs. 1-3. and as described above, affects the way that the cleave propagates through the fibre.

[0021] Splicing HCFs requires consideration of the cleave profile of an end face of an HCF that is to be spliced, as this affects the alignment between microstructures of the fibres. If an indicated cleave angle is too large, and / or if a step height of a step in a cleaved end-face is too high, issues in some cases arise during splicing and the splice will in some cases have an optical propagation loss that is too high for an application of the spliced HCF, and therefore the fibre with the unacceptable cleave angle is for example re-cleaved or discarded.

[0022] The acceptable loss of a splice may be application dependent and hence the acceptable cleave profile of the fibres being spliced may also be application dependent. In various examples, an indicated cleave angle that is substantially zero degrees and / or an indicated step height of substantially 0 i.e. a flat end face surface, is acceptable, and in other cases a cleave angle of zero to one degree is acceptable. In some cases, an indicated cleave angle of zero to 1.5 degrees is acceptable. What is considered an acceptable and / or unacceptable cleave profile may be predefined for a particular application.

[0023] Determination of a cleave profile, in some cases being or comprising an indication of cleave angle, of a fibre is time consuming, especially in an approach wherein a cleave angle is determined in a splicer after the ARFs to be spliced have been inserted, and where multiple splices are performed at each joint of a fibre. In this approach, a cleave angle is determined (for one or more of the inserted ARFs) immediately prior to splicing (i.e. determined by a system associated with splicing apparatus of the splicer), and the splicing process is consequently delayed by an evaluation of the cleave angle. Additionally, as the ARFs to be spliced are inserted into the splicer prior to determining a cleave angle, in response to the angle being deemed unacceptable, the ARF with the unacceptable cleave angle is then in some cases removed from the splicing apparatus of the splicer and replacedDocket No. 13768-4558a_502046-PCT01 with a different ARF, which must then have its cleave angle evaluated, which further delays the splicing process.

[0024] It should be appreciated that in various examples not only does the cleave profile affect a quality of a splice, but that a quality of a cleave more broadly affects the quality of the splice. The quality of the cleave in some cases is affected by a cleave profile, debris present on the fibre and / or damage to the cleaved fibre, and refers to a suitability of the cleaved fibre for splicing i.e. splicing with another fibre, in a way that ideally results in a high-quality splice i.e. one with low optical propagation loss.

[0025] It should further be appreciated that, though splicing and cleaving is described with reference to ARFs, they relate also to hollow core fibres. In the case of a first hollow core fibre comprising microstructure, this microstructure is ideally aligned with microstructure of a second hollow core fibre with which the first fibre is spliced. In the case of hollow core fibres not comprising microstructure, low damage to the fibre and alignment with a second fibre during splicing is still desired for a low optical propagation loss over the splice.

[0026] FIG. 4 illustrates a transverse cross-sectional view- of a cleaved antiresonant hollow core optical fibre, and shows some features of a cleaved ARF. The central core 406 comprises the microstructure of the ARF 400, in some cases comprising the plurality of tubular cladding capillaries 104, 204, 304 illustrated in FIG. 3. Cleaver impact site 408 is a structure formed when a blade of a cleaver contacts the outer cladding 410 of the ARF 400 during scoring, which enables the controlled break of the cleaving process. Mechanical waves due to the cleave travel from the impact site 408 in two general directions around the core 406 of the ARF 400, through the outer cladding 410, and meet at a side of the outer cladding 410 that is opposite to the impact site 408. At this opposite side, the w aves recombine, resulting in a tear 402 in the outer cladding 410. Additionally, a triangular structure 404 is located close to the tear 402. the triangular structure associated with the tear 402 and further characterising it. The shape of the tear 402 (including the substantially triangular structure 404) varies in terms of visual characteristics (including for example shape, size, visibility, position) based on characteristics of a cleave, in some cases indicated by a cleave profile, in some cases comprising a cleave angle, of the ARF 400. Additionally, a relative height of elements of the microstructure of the core 406 and of points within the outer cladding 410 of the cleaved end face of the ARF 400 vary based on a cleave profile of the ARF 400. In some cases, the height of the outer cladding 410 around either side of the tear 402 varies based on characteristics of a cleave indicated by a cleave profile, in some cases being a cleave angle, ofDocket No. 13768-4558a_502046-PCT01 the ARF 400. The height of microstructure elements such as the cladding capillaries within the hollow core 406 shares in various examples substantially the same height as the outer cladding 410 at substantially same distances along the path of a mechanical wave from the cleave impact site 408 during cleaving, as both the microstructure elements within the hollowcore 406 and the outer cladding 410 are affected by the same wave from the cleave impact site 408 during cleaving.

[0027] Though an ARF 400 is illustrated, in various examples a hollow core fibre comprises outer cladding 410 and a cleaved hollow core fibre comprises an impact site 408, a tear 402, and a triangular structure 404 in the outer cladding 410. A hollow core fibre in some cases also comprises microstructure elements within a core 406 of the hollow core fibre.

[0028] FIG. 5A illustrates a schematic diagram 500 of an approach to determining a cleave quality of a fibre, wherein the cleave quality comprises or consists of a cleave profile of the fibre, and wherein the cleave profile is determined using a side-view camera. Such an approach is for example implemented in a splicer. An end face surface 504A of a hollow core fibre, in some cases an ARF502A has a cleave profile (as HCF 502A has been cleaved). Camera 506 is used to image a side-view of HCF 502A including the cleaved end face surface 504A (from a side-view), and provides the image to a processing system 508 that determines an indication of a cleave angle of the end face surface 504 A (and therefore in some cases a cleave profile) using the side-view image taken by the camera 506. In various examples, a second side-view image is taken using the camera 506 of a second hollow- core fibre 502B. the second image including the cleaved surface 504B (from a side-view), and the second image being provided to the processing system 508 for cleave angle determination. Processing system 508 in some cases analyses a received image of a fibre and determines a slope of the end face surface 504 A, 504B depicted in the image in order to determine a cleave angle.

[0029] However, a cleave profile, which in some cases comprises an indication of a cleave angle and is determined via this method depends on a rotation of the fibre being imaged, as illustrated by FIG. 5B, which illustrates a chart showing variation in determined cleave angle using the apparatus of FIG. 5 A. This chart is merely exemplary, but shows that the determined cleave angle varies from ~0.2 to -1.6 degrees as the fibre being imaged is rotated. Where an acceptable cleave angle for splicing is 0 to 1 degree, the range from 0.2 to 1.6 degrees is large, and is inconsistent between acceptable or non-acceptable cleave angles, which can affect decisions as to whether to splice or re-cleave, for example. The dotted lines mark full 360 degree rotations of the fibre. For hollow core fibres, this in some cases isDocket No. 13768-4558a_502046-PCT01 because a step in surface height across a tear in the imaged cleaved end face surface causes a determined cleave angle, and therefore a determined cleave profile in some cases, to depend on a viewing angle. It is clear, therefore, that using a side view image to determine a cleave angle and / or in some cases a cleave profile of a cleaved fibre provides an inconsistent cleave angle that depends on the rotation of the fibre, and a cleave angle and therefore cleave profile measurement is in some cases inaccurate.

[0030] The herein disclosed technology provides a more accurate way of estimating a cleave quality comprising an indication of cleave profile, which in some cases comprises an indication of cleave angle, of a HCF, by analysing an end face image of the HCF rather than a side-view image, which avoids the determined cleave angle depending on a rotation of the fibre as in an approach using side view images. As a cleave angle is in various examples difficult to determine given a non-flat cleaved end face surface, a cleave profile is instead determined, which comprises in some cases an indication of cleave angle in the sense that it comprises an indication of a cleave angle determined as an angular difference between a normal to the cleaved end face surface and the central axis of the cleaved fibre, wherein the normal to the end face is an average normal across a plurality of points on the surface, or a normal at a single point on the surface. In some cases, the cleave profile comprises a plurality of indications of cleave angle, each being an indication of cleave angle at different or across different points of the cleaved end face surface. In some cases, an indication of cleave quality, the indication of cleave quality comprising a cleave profile or an indication of a cleave profile, is determined or estimated by the herein disclosed technology.

[0031] As defined herein an indication of cleave quality is any indication, such as at least one of: a range, probability of a range, probability of a value, value, level, whether a quality is above or below a threshold quality’, and / or a label, associated with a cleave quality. In some cases, a cleave quality is indicated by and / or comprises an indication of cleave profile, in some cases the indication of cleave profile being an indication of cleave angle.

[0032] FIGS. 6A and 6B illustrate a flow diagram of a method according to the herein disclosed technology. An end face image 600 of a cleaved HCF, in some cases being a nondepth image and a non-interferometer image, i.e. such that the end face image 600 comprises no explicit information regarding a pixel distance from the camera or relative distance of two pixels, is received by an analysis system 602, the analysis system comprising a model 604, and the analysis system 602 processes the received end face image 600 to estimate cleave quality’ of the cleaved HCF, in some cases cleave quality’ 606 comprising a cleave profile 607, by determining an indication of cleave quality using a model. The indication of cleaveDocket No. 13768-4558a_502046-PCT01 quality, in some cases comprising an indication of cleave profile, is output by the analysis system 602. Outputting the indication of cleave quality refers in various examples to providing the indication of cleave quality to another entity, displaying the indication of cleave quality on a screen, or providing the indication of cleave quality to or using for displaying the indication of cleave quality any other method or entity’, including providing the indication of cleave quality to a database or otherwise storing the indication of cleave quality. In some cases, the database or storage storing the indication of cleave quality is accessed by an entity such as a cleave and / or splicer to perform or decide whether to perform an action such as cleaving and / or splicing associated with the imaged fibre for which the indication of cleave quality is determined.

[0033] In method 608, performed for example by analysis system 602, an end face image depicting a HCF is received 610, in some cases the end face image being of a cleaved end face of the HCF. and pixel data of the end face image is analysed to determine cleave quality by determining at least one feature of the image representing a characteristic of an end face of the HCF and using a model to map the at least one feature to an indication of cleave quality612. In this way, a cleave quality of the HCF is determined more accurately than in alternate approaches using a side-view image, especially where the indication of cleave quality comprises an indication of a cleave angle of the HCF. An end face image as defined herein refers to an image of an end face of an entity, in the present context the entity being a fibre. A feature of an image refers to a feature associated with pixel values of the image, and thus a feature of an image representing a characteristic of a HCF refers, for example, to a width in pixels of a tear in the HCF depicted in the image (therefore being associated with pixel values, and representing a tear width as a characteristic of an HCF), an angular width (i.e. an angle at the camera subtended by the tear), and / or a width in microns of a tear in the HCF depicted in the image (therefore being associated with pixel values, as a micron width is associated with a pixel width in the image i.e. a larger micron width would result in a larger pixel width).

[0034] Method 608 further comprises outputting the indication of cleave quality, such as by providing the indication of cleave quality to a splicer and / or a cleaver 614, for example for determining whether to perform an action such as splicing and / or cleaving associated with the fibre. Additionally or alternatively, method 608 further comprises initiating an action, for example cleaving or splicing, at the cleaver or splicer, respectively. Initiating an action in some cases refers to the provision of the indication of cleave quality, such as where such provision causes the action to occur, for example where a splicer is expecting to receive anDocket No. 13768-4558a_502046-PCT01 indication of cleave quality to determine based on the indication whether to initiate splicing, or for example where a splicer is configured to determine based on the indication whether to initiate splicing in response to receiving the indication of cleave quality. These situations are merely exemplary, and it should be appreciated that any other initiation of an action is performed in various cases.

[0035] Pixel data refers to data associated with pixels of an image, for example values each associated with a pixel defining an intensity of the pixel (represented in some cases by a greyscale image). By analysing pixel data of an end face image without requiring a splicer or cleaver to determine a cleave quality, the disclosed technology is enabled to be implemented independently of a splicer and / or a cleaver, enabling deployment in the field and reducing a delay involved in splicing. For example, a cleave quality comprising in some cases a cleave profile (in some cases comprising an indication of cleave angle) for a first HCF is determined in parallel with splicing of a second HCF with a third HCF, prior to preparing the first HCF for splicing, or otherwise prior to insertion of the first HCF into a splicer, and then, in response to the first HCF being determined to have an acceptable cleave quality, for example using the indication of cleave quality as mentioned above, the first HCF is inserted into the splicer for splicing. If the first HCF is determined to not have an acceptable cleave quality, the first HCF is in some cases discarded or re-cleaved without entering the splicer, further preventing a delay associated with removing the first HCF from the splicer, for example in order to discard or re-cleave the first HCF. In this way, the disclosed technology enables a more efficient cleaving and / or splicing system, saving in some cases multiple minutes of time during splicing, whilst ensuring a high quality splice. Though cleave quality is mentioned above, it should be appreciated that in some cases this comprises or consists of a cleave profile, and / or in some cases the indication of cleave quality as mentioned above comprises or consists of an indication of cleave profile.

[0036] In various examples, the received end face image 600 by the analysis system 602 (corresponding to the end face image of block 610), is a single image. In some cases, the received end face image 600 is a non-depth image, for example an image recording an intensity of light received by a camera.

[0037] In various examples, the model 604, corresponding to the model of block 612, is predefined, and in some cases comprises at least one of: a machine-learning model that is trained to output an indication of cleave quality given at least one feature of an image representing a characteristic of an end face of a fibre depicted in the image as an input, at least one rule mapping the at least one feature to an indication of cleave quality, and a lookupDocket No. 13768-4558a_502046-PCT01 table associating the at least one feature with an indication of cleave quality. The machinelearning model is trained in various examples in a supervised or semi-supervised way. In various examples, the model 604 comprises both a machine-learning model and at least one rule, in some cases configured to determine different factors, which are then combined to determine an indication of cleave quality’. For example, the machine-learning model is used to map to an indication of cleave angle and / or cleave profile, and the at least one rule is used to map to an indication of damage to the imaged fibre, where the two mapped results are combined to determine an indication of cleave quality. This example is merely illustrative, and in some cases any combination of model components are used, for any purpose associated with determining an indication of cleave quality.

[0038] In some cases, the model is initialised to process image features and map them to an indication of cleave quality according to the herein disclosed technology7. In some cases, the machine-learning model is trained using training inputs, each comprising at least one feature representing a charactenstic of an end face of a test fibre (for example a HCF, in some cases being an ARF, where a test fibre refers to a fibre used to construct the model i.e. to train the machine-learning model in cases where the model comprises the machine-learning model) determined from an end face image of the test fibre, and corresponding ground truth cleave profiles, each determined from an end face image, taken using a depth camera, of a same test fibre as a corresponding training input is associated with, and / or corresponding ground truth cleave angles, each determined from a side view image (in some cases a nondepth image) of a same test fibre as a corresponding training input is associated with. The training data comprises data associated with a plurality of test fibres and with a plurality of cleave profiles and / or angles. The test fibres are cleaved HCFs. In some cases, a ground truth cleave profile is determined by measuring a cleave angle using a measurement apparatus such as by using a depth camera for example being an interferometer, and / or by using a plurality7of high precision imaging stages to determine a cleave angle from an end-face of a test fibre. Using a plurality of high precision imaging stages comprises using a camera to focus on a portion of the end face of a fibre, and obtaining a pixel intensity7for the portion of the end face of the fibre. The camera is then used to focus on another portion of the end face of the fibre by moving the camera towards or away the end face of the fibre, and / or the end face of the fibre towards or away from the camera, to obtain the same pixel intensity. In this way, a relative movement of the stage with respect to various portions of the end face of the fibre (each associated w ith a pixel) at different positions on the end face of the fibre is obtained, which enables relative heights and / or a slope of the end face to be obtained and therefore anDocket No. 13768-4558a_502046-PCT01 indication of cleave angle and / or cleave quality to be obtained. In some cases, training data comprises at least two training inputs for a same test fibre but where an associated cleave angle to each of the at least two training inputs is determined by determining an indication of cleave angle from a side view image at a rotation of the same test fibre that is different for each of the at least two training inputs for the same test fibre. Alternatively, the training data comprises an indication of whether the cleave profile and / or angle is acceptable or not, instead of a ground truth cleave profile and / or angle (e.g. a binary label indicating a pass or a fail).

[0039] In an example, the model comprises at least one rule mapping the at least one feature to a cleave profile and / or angle, and the model is created by at least defining the at least one rule using a feature representing a characteristic of an end face of a test fibre (for example a HCF) determined from an end face image of the test fibre and an indication of at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre. For example, such a rule could include ‘Rule 1 : If tear width is 3 microns, cleave angle is 1 degree (i.e. cleave profile comprises an indication of a cleave angle of 1 degree)’, ‘Rule 2: If tear width is 1 micron, go to rule 3’, ‘Rule 3: If tear width is 3 pixels and standard deviation of pixel intensities of pixels depicting microstructure of fibre is less than 0.3. cleave angle is 0.2 degrees’ or any other rules. It should be appreciated that, whilst the provided examples are in natural language, they are in various examples implemented using code, variable names and / or by any other method. In some cases, the at least one rule comprises at least two interrelated rules. The feature representing the characteristic of the test fibre in the context of the at least one rule refers to a feature of an end face image of the test fibre, the feature representing a characteristic of an end face image of the test fibre.

[0040] In various examples, the at least one rule is defined automatically by determining the at least one feature of the rule and the cleave profile as mentioned, is defined manually, or is defined by any mixture of the two (for example, at least one rule being automatically determined and at least one rule being manually determined). The at least one feature of the rule is determined from an end face image of a fibre in a same w ay as the disclosed technology determines at least one feature of an end face image of a fibre, and a mapped cleave angle of the at least one rule is determined from at least one side view image of the fibre, in some cases being determined by calculating an average of cleave angles determined from numerous side view images at various rotations of the imaged fibre.Docket No. 13768-4558a_502046-PCT01

[0041] In some cases, the model comprises a lookup table associating the at least one feature with at least one of: a cleave profile; and a cleave angle, and the lookup table is created by associating at least one feature representing a characteristic of an end face of a test fibre determined from an end face image of the test fibre with at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre. The associating is performed automatically, manually or through a mixture of the two, as mentioned above, where an automatic determination in some cases comprises associating a plurality of features of an image depicting an end face view of a cleaved HCF, the features each representing a characteristic of an end face of the cleaved HCF and being determined from an end face image of the cleaved HCF, with a cleave profile and / or angle. In some cases this association is then performed in the same way but for a different test fibre and / or different cleave profile and / or angle, to construct a lookup table. To map determined features to an indication of cleave quality, for example comprising a cleave profile, which in some cases comprises an indication of cleave angle, a search is then performed for the determined features, in various examples being a vector search to determine a closest set of features in the lookup table to the determined features (where an embedding of the determined features is compared to embeddings of the features of the lookup table), and a corresponding cleave profile and / or angle associated with the closest features in the lookup table is retrieved. In some cases, the association of the lookup table is performed by determining at least one feature of an end face image and an associated cleave profile and / or angle using at least one side view image as mentioned above with respect to the model being at least one rule.

[0042] Though determining a cleave profile is mentioned herein, and is associated with determining cleave quality, in various examples determining a cleave quality and in some cases cleave profile comprises or refers to determining an indication of cleave profile, an indication of cleave quality, and / or whether a cleave quality and in some cases cleave profile (of an HCF) is acceptable or not, which refers to a cleave quality or profile that is suitable for splicing. In some cases, an acceptable cleave profile is one that indicates a cleave angle of substantially zero degrees, or is zero to one degree. As such, in the machine learning model, at least one rule, or lookup table as mentioned above, the ground truth cleave profiles, the mapped cleave profiles or the associated cleave profiles respectively are in various examples an indication as to whether a cleave profile is acceptable or not (referring to whether the cleave profile is associated with cleave quality of a fibre that is suitable for splicing). For example, such an indication is ‘pass’ or ‘fail’, ‘acceptable’ or ‘not acceptable’Docket No. 13768-4558a_502046-PCT01 or any other indication. Alternatively, in various examples, the mapped indication of cleave quality, in some cases comprising an indication of cleave profile using the model is a value of a cleave profile (in some cases comprising a value of cleave angle), and the method 608 further comprises determining whether the mapped indication of cleave profile (in some cases being a cleave angle itself, which is an indication of cleave quality) is acceptable or not (for example by determining whether it is within a predefined range of cleave profiles or indicates a cleave angle within a predefined range of cleave angles). In a same way, in various examples method 608 further comprises determining whether the mapped indication of cleave quality indicates a quality above a threshold. In some cases, the model is used to map a feature of the image to an indication of cleave quality which is whether the quality is above a threshold, for example ‘pass’, ‘fail’ or the other indications as mentioned above.

[0043] In some cases, the method 608 further comprises, in response to determining the cleave quality to be acceptable, or in response to the indication of cleave quality indicating a uality above a threshold quality, initiating splicing of the fibre depicted in the received end-face image (via block 610) at a splicer. In some cases, the method 608 further comprises, in response to determining the cleave quality to not be acceptable (i.e. to be unacceptable), or in response to the indication of cleave quality indicating a quality below a threshold quality, initiating re-cleaving of the fibre depicted in the received end-face image (via block 610) at a cleaver, or discarding the fibre depicted in the received end-face image (via block 610).

[0044] Exemplary implementations of a system according to the disclosed technology will now be described.

[0045] FIG. 7 illustrates a schematic diagram of a system according to the herein disclosed technology7. Apparatus 700 comprises analysis system 712, which is configured to receive an end face image of aHCF. such as an ARF 702A. 702B as mentioned above. In various examples, apparatus 700 comprises a camera 710 configured to image an end face 706 A, 706B of an ARF 702A, 702B. In some cases the camera 710 is configured to image the end face 706A, 706B directly without a lighting element being used, and in other cases a lighting element is used to illuminate at least a portion the end face 706A, 706B imaged with the camera 710. In some examples, such as when the apparatus 700 is comprised in or comprises a splicer, the camera 710 is configured to image the end face 706 A, 706B of the ARF 702A, 702B by7the apparatus 700 being configured such that the imaged ARF 702A, 702B has light coupled into it by a lighting element 704A, 704B, and the apparatus 700 comprising a mirror 708 or other reflective element that directs light from the imaged ARFDocket No. 13768-4558a_502046-PCT01702 A, 702B to the camera 710. In other examples, a lighting element is used to emit light from the camera such as by directing light from the lighting element through an optical component such as a lens of the camera, or is used to emit light from the direction of the camera relative to the end face.

[0046] The mirror 708 and in some cases the camera 710 are in various examples temporarily inserted into a gap between two ARFs 702A, 702B prior to splicing the two ARFs 702A, 702B together, an image is taken using the camera 710, and the camera 710 and / or mirror 708 are then removed from the gap between the two ARFs 702A, 702B. The image taken by the camera is provided to the analysis system 712, which implements for example the method 608 of FIG. 6B. It will be appreciated that in the orientation shown in FIG. 7, the mirror 708 enables imaging of only one of the two ARFs 702A. To image the second ARF 702B, the mirror 708 may be rotated or a second mirror may instead be inserted into the gap. In another example, the mirror 708 may be reflective on both sides and a second camera may be provided to enable an image of each ARF to be captured using the same mirror and substantially simultaneously. Alternatively, mirror 708 is in some cases a split mirror, enabling both end faces i.e. end faces of ARF 702A and 702B respectively, to be imaged by a single camera 710 at a same time.

[0047] Though FIG. 7 illustrates two ARFs 702A, 702B, in some cases only a single ARF 702A, 702B is provided to apparatus 700, and / or is imaged by the camera 710, the image being provided to the analysis system 712. In other cases, a plurality of ARFs 702A, 702B are imaged by the camera 710, the images being provided to the analysis system 712. Though ARFs are mentioned in this context, this is merely exemplary’, and HCFs are instead used in various examples.

[0048] FIG. 8 illustrates a schematic diagram of a system according to the herein disclosed technology, the system associated with a cleaver. Cleavers typically operate by scoring the surface of a fibre at the desired cleave location and then applying tension until the fibre breaks along stress lines created by the scoring. The fibre may be placed under a small amount of tension prior to scoring and the tension may be increased after scoring. Alternatively, a first tension is in some examples applied to the fibre prior to scoring and this first tension remains constant throughout scoring. The tension applied to the fibre is in various examples 100 gram-force, 200 gram-force, 300 gram-force or any other tension, and in some cases is determined to be suitable for a passable cleave based on characteristics of the fibre. The resultant end face is ideally without any cracks or chips, or any chips that areDocket No. 13768-4558a_502046-PCT01 present are not close to the core but instead are just at the outer edge of the cladding (e.g. in the form of a cleaver impact site 408 as shown in FIG. 4 and described above).

[0049] Apparatus 800 comprises cleaving apparatus 806, comprising any suitable apparatus for cleaving a fibre, for example a blade for scoring a fibre in some cases inserted into apparatus 800. Apparatus 800 further comprises two clamps 802A and 802B configured to clamp an inserted ARF (where the position of the longitudinal axis of the ARF is shown by the dotted line 804), and a tensioning mechanism 814. The tensioning mechanism 814 is connected to one or both of the clamps 802A, 802B and is arranged to apply tension to the ARF when held by the two clamps. The cleaving apparatus 806 is moveable between a first position (as illustrated) where it is distant from the inserted ARF and a second position where it is in contact with the outer edge of the cladding of the ARF between the two clamps (i.e. the cleaving apparatus 806 is moveable along the dotted line with the arrow immediately to the right of the cleaving apparatus 806). In various examples, the cleaver of apparatus 800 comprises further or alternative components for cleaving the ARF.

[0050] The apparatus 800 further comprises analysis system 812, which implements the disclosed technology, for example method 608 of FIG. 6B. The analysis system 812 is configured to receive an image of a cleaved end face of an ARF inserted into the apparatus 800 once the ARF is cleaved. In an example, the apparatus 800 comprises a camera 810 configured to directly image the cleaved end face of the ARF after cleaving, and in another example comprises a camera 810 and a mirror 808 or other reflective surface that enables end face imaging in the same way as the camera 710 and mirror 708 of FIG. 7. As such, the apparatus 800 is configured to, after cleaving, position the camera 810 (and in some cases the mirror 808) so as to enable imaging of the cleaved end face of the ARF (for example by following the path illustrated by the dotted line with arrowhead immediately to the left of the camera 810). After imaging, the camera 810 and mirror 808 are configured to return to their original positions so as to enable cleaving of another ARF.

[0051] Instead of positioning the camera 810 and / or mirror 808, in various examples the apparatus 800 positions the cleaved ARF so as to have its cleaved end face imaged by the camera 810.

[0052] The image taken by the camera 810 is provided to the analysis system 812, corresponding to the analysis system 712 and analysis system 602 of FIGS. 7 and 6A respectively.

[0053] In some cases, one or both of the cleaved end-faces of a cleaved ARF are imaged and the images are provided to the analysis system 812. In other cases, one cleavedDocket No. 13768-4558a_502046-PCT01 end-face of the cleaved ARF is imaged, and the other cleaved end-face is not imaged and is discarded as it is comprised in a portion of the cleaved ARF that is discarded. Cleaving a single ARF results in two portions of the single ARF, each portion being a cleaved ARF with a cleaved end face at an end of the cleaved ARF where the cleaving apparatus 806 contacted.

[0054] By including the analysis system 812 within a cleaver, this simplifies the overall cleaving and splicing operation and thereby reduces the time taken. This may be particularly important for field operations as a cable may contain large numbers of fibres that require splicing (e.g. following a cable break). Furthermore, as the position of the cleaved face is known precisely within the cleaver and is consistent between cleaving operations, the disclosed technology reduces time taken to focus the camera arrangement. In some examples, if the analysis system 812 determines that the angle of the cleaved end face is not acceptable, the fibre may be automatically repositioned within the apparatus 800 and re-cleaved.

[0055] Though FIGS. 7 and 8 illustrate exemplary apparatuses implementing the disclosed technology, in various examples an apparatus implementing the disclosed technology is independent of a splicer and a cleaver, in some cases the apparatus comprising an analysis system corresponding to systems 712 and 812, a camera configured to image a cleaved end face of an ARF, and optionally, a holder configured to receive the ARF. In other cases, the apparatus implementing the disclosed technology comprises a cleaver and / or a splicer. Though the above description is illustrated with respect to ARFs, it should be appreciated that in various examples HCFs are instead used.

[0056] By estimating cleave quality which in some cases comprises estimating a cleave profile, using an end face image of a cleaved HCF. a more accurate cleave profile is determined as compared to an approach using a side view image in which a cleave angle is determined. In some cases, an increased height of outer cladding around a tear in the outer cladding, relative to areas of the outer cladding further away from the tear, can obstruct a view of areas further away from the tear when a side view image is taken. Especially, in the case of a step being present across the tear, as will be elaborated upon below, a step with a height above a threshold in some cases causes a view of a side of the cladding or a side of the tear, such as towards a lower surface of the step, to be obscured or unable to be imaged from a side view image taken at a specific orientation relative to the fibre, such as where the higher surface of the step is closer to the camera. By using an end face image, this issue is countered, further improving an accuracy of the cleave profile (which in some cases comprises an indication of cleave angle) determination. Additional performance and efficiency improvements are obtained by estimating the cleave quality more accurately in aDocket No. 13768-4558a_502046-PCT01 splicer (especially if the estimation of the cleave quality occurs prior to seating the HCF ready for splicing), in a cleaver (as this reduces delays in the splicer and enables re-cleaving if the cleave angle is unacceptable), or independently (enabling parallel cleaving / analysis / splicing without delays in the splicing and cleaving steps which prevent cleaving and / or splicing of fibres that are not having their cleave quality determined).

[0057] The at least one feature of the received image by the analysis system 812 will now be elaborated upon, each of the at least one feature representing a characteristic of an end face of an imaged fibre and the at least one feature used with a model to map the at least one feature to an indication of cleave quality according to the herein disclosed technology.

[0058] FIG. 9 illustrates a transverse cross-sectional view of a cleaved antiresonant hollow core optical fibre that is analysed according to the herein disclosed technology, and an exemplary chart illustrating the analysis.

[0059] The ARF 900 comprises a tear and impact region associated with a cleaving process, and dotted lines 902A-902E illustrate portions of the microstructure of the ARF 900 over which analysis is performed, in one example the portions of the microstructure being used to determine a feature of the image that is mapped to a cleave profile according to the herein disclosed technology. The dotted lines 902A-902E are located such that they cross only a single, outermost tubular cladding capillary of the microstructure.

[0060] Chart 904 shows five corresponding pixel intensity line profiles 906A-906E corresponding to pixel intensities across the corresponding dotted lines 902A-902E, and provides an example of a feature of an image of an end face of the ARF 900 that represents a characteristic of the end face of the ARF.

[0061] Though pixel intensity line profiles 906A-906E are illustrated across lines 902A-902E respectively, it should be appreciated that in various examples lines 902A-902E are located additionally or alternatively such that they each cross multiple cladding capillaries, such that they each cross a single cladding capillary that is not an outermost cladding capillary i.e. are across a secondary and / or tertiary capillary corresponding to elements 204 and 304 of FIG. 3, and / or such that they each cross a different but corresponding capillary i.e. each cross different primary capillaries, different secondary capillaries, or different tertiary capillaries. By obtaining pixel intensity line profiles over different but corresponding capillaries, the cleave profile estimation is made more robust in that the estimation does not rely solely on a primary capillary. Additionally, by comparing pixel intensity line profiles across multiple capillaries associated with a same primary' capillary, for example secondary and tertiary capillaries, an indication of a quality of a cleaveDocket No. 13768-4558a_502046-PCT01 can be obtained, given that a good quality cleave will have all capillaries associated with a same primary capillary at a same level. In some cases, values such as pixel line intensities across different but corresponding capillaries are averaged. In additional cases, pixel line intensities across different lines across a same capillary are averaged, which improves accuracy, reduces errors and / or counters the effect of any artefacts and / or anomalies that may affect the pixel line intensity measurement.

[0062] More particularly, chart 904 shows that profiles 906A and 906B have higher peak pixel intensities than profiles 906C, 906D and 906E, and that profile 906E is broader than profile 906D which is broader than profiles 906 A, 906B and 906C.

[0063] The line profiles themselves are a feature of an image of the end face of the ARF 900 (being related to pixel intensities) that represents a characteristic of the end face of the ARF. This is because a peak pixel intensity is related to a distance of the element of the fibre over which the line profile is constructed, in this case the outermost tubular cladding capillaries of the core, from the camera. A higher peak pixel intensity, in some cases where variables impacting the pixel intensity aside from a distance from the camera of the element of the fibre over which the line profile is constructed are controlled or known, implies that the corresponding cladding capillary is closer to the camera than a cladding capillary with a corresponding pixel intensity profile that has a lower peak pixel intensity. Such variables impacting the pixel intensity aside from a distance from the camera of the element of the fibre over which the line profile is constructed include a focus i.e. focal plane distance of the camera relative to the element of the fibre over which the line profile is constructed and / or an illumination of the imaged end face. The distance of a cladding capillary from the camera is a characteristic of the fibre. In this way, a higher relative peak pixel intensity over a first line comprising pixels of a first portion of the end face of the fibre, relative to a peak pixel intensity over a second line comprising pixels of a second portion of the end face of the fibre, implies that the first portion of the end face of the fibre is closer to the camera than the second portion of the end face of the fibre, where the focal plane of the camera is closer to the second portion than the first portion of the end face of the fibre.

[0064] How ever, a peak pixel intensity of a pixel line intensity7profile is also related to a distance of a portion of the fibre that the line intensity profile is across, from the focal plane of the camera. Where the portion is further from the focal plane, the line profile around the peak intensity is broader, and the peak is in some cases lower than if the same portion was closer to the focal plane. Additionally, there are in some cases, such as where the portion is sufficiently far from the focal plane of the camera, one or more smaller peaks than a primaryDocket No. 13768-4558a_502046-PCT01 highest peak located either side of the primary peak in the line profile, due to blurring of the image taken by the camera.

[0065] In this way, a higher relative peak pixel intensity over a first line comprising pixels of a first portion of the end face of the fibre, relative to a peak pixel intensity over a second line comprising pixels of a second portion of the end face of the fibre, implies that the first portion of the end face of the fibre is some combination of closer to the camera than the second portion of the end face of the fibre or closer to the focal plane of the camera than the second portion.

[0066] To counter this, in various examples, the distance of the focal plane of the camera from a respective element is controlled, for example by refocusing the camera when different pixels are imaged for construction of a different line profile, or by focusing the camera such that a focal plane is consistently closest to a furthest or nearest portion of the imaged fibre, for example a furthest or nearest capillary. The nearest or furthest portion of the imaged fibre is for example determined by analysing peak pixel intensities of a line profile across a plurality’ of portions of the imaged fibre after focusing the camera on each of the plurality of portions, and then determining the nearest or closest portion by the brightest or faintest peak intensity respectively. Alternatively, the camera is focused such that its focal plane is located closer than all portions of the end face of the imaged fibre or is located further than all portions of the end face of the imaged fibre, from the camera. Alternatively or additionally, a model takes into account a shape of the line profile, for example secondary peaks due to blurring, in mapping a feature of an image such as a peak pixel line profile intensity to an indication of cleave quality, such that more complete factors are taken into account in the mapping to an indication of cleave quality.

[0067] Additionally, the relative peak pixel intensity of a first pixel intensity profile 906A-906E to at least one other pixel intensity profile 906A-906E provides an indication of how the ends of the cladding capillaries 902A-902E are sloped in terms of a surface comprising the ends of the capillaries. For example, the chart 904 indicates that capillaries 902A and 902B corresponding to profiles 906A and 906B are closer to the focal plane of the imaging setup than capillaries 902C-902E corresponding to profiles 906C-906E, sequentially. This indicates that the capillaries slope away from the camera (assuming a focal plane is controlled as mentioned above), from 902A to 902E, which is a characteristic of the end face of ARF 900.

[0068] The relative height of at least one capillary relative to a reference capillary, indicating a slope of a surface comprising ends of a plurality of the capillaries, determined forDocket No. 13768-4558a_502046-PCT01 example by computing pixel intensities for pixels of each capillary relative to a pixel intensity for pixels of a reference capillary, indicates a cleave angle in terms of direction and / or degree, and a model is in some cases used to map this slope to an estimated cleave angle and therefore an estimated cleave profile. In some cases, multiple slopes are determined across different sets of points on the end face surface, and the multiple slopes are used to define a cleave profile as they each indicate a cleave angle determined across different sets of points.

[0069] Additionally, or alternatively, a broader pixel intensity profile 906A-906E indicates that the pixel intensity profile is of a portion of the imaged fibre that is further away from the camera, in some cases where additional factors relating to the distance of a portion of the fibre from the camera are known or controlled as mentioned above, relative to a narrower pixel intensity profile 906A-906E, as an element of the ARF 900 that is further from the camera is further away from the focal plane of the camera. Where the camera is focused with a set focal plane on a point closer to the camera than all the outer cladding capillaries of the ARF 900, or is focused on a closest capillary of the ARF 900. being further away from the focal plane of the camera indicates being further away from the camera. As such, in various examples, a method according to the herein disclosed technology comprises using a model to map at least on feature of an image of an ARF 900 and further at least one parameter of a camera that took the image, to an indication of cleave quality, in some cases comprising an indication of cleave profile .

[0070] The at least one parameter of the camera in various examples comprises a focal length, location of a focal plane, resolution, aperture size, or any other parameter used in combination with a feature of a taken image of an end face of an ARF 900 to indicate a cleave quality, in some cases comprising or consisting of an indication of a cleave profile.

[0071] Whilst the above description relates to relative pixel intensities in terms of breadth of a profile or peak intensity, in some cases a single intensity value or single breadth value is used as a feature of the image, and is for example used in combination with another feature (in some cases a relative feature) when using a model to map at least one feature to a cleave profile. Illustratively, a peak pixel intensity' of a pixel intensity' profile across a cladding capillary may be used in combination with a relative pixel intensity of a plurality of pixels at different points on the image of ARF 900 corresponding to the outer cladding ‘jacket’ of the ARF 900.

[0072] Moreover, though pixel intensity' profiles across tubular cladding capillaries have been described, in various examples a standard deviation, average, or breadth of pixel intensities of an image are determined as features of the image used to map to a cleaveDocket No. 13768-4558a_502046-PCT01 profile. The pixel intensities in these metrics are in some cases of pixels in the image depicting a first and second portion of the fibre, a portion of the fibre being for example at least one pixel depicting outer cladding of the ARF 900, at least one pixel depicting one or more capillaries in the ARF 900, at least one pixel depicting a microstructure element of the ARF 900, at least one pixel depicting a part of a tear in the ARF 900 or any other portion of the fibre.

[0073] In various examples, a pixel intensity profile itself (in some examples a relative pixel intensity profile of pixels i.e. a profile depicting a first relative to a profile depicting a different portion of the ARF 900) is used as a feature of an image that is mapped to an indication of cleave quality’, in some cases comprising an indication of cleave profile.

[0074] A radial line pixel intensity’ profile, in one example illustrated by radial line 903, of an image depicting ARF 900 is further used in some cases as a feature of the image representing a characteristic of an end face of the ARF 900. A radial line pixel intensity profile is a pixel intensity profile across a line originating from a point on an image and located in a direction defined by an angle. Radial line 903 illustrates an example radial line originating at a centre of the ARF 900 depicted in an image, in some cases a pixel intensity profile being determined across the radial line 903. In various examples, the radial line pixel intensity profile starts at a pixel depicting a centre of the ARF 900. In other examples, the radial line pixel intensity’ profile starts at a pixel depicting an inner border of the outer cladding ‘jacket’ of the ARF 900, or at any other pixel. Additionally, some examples radial line 903 ends at a pixel outside of the imaged ARF 900. and a pixel outside of the ARF 900 is for example used to determine a responsivity of camera pixels, determine an indication of a noise of an image, and / or to define a ‘dark’ pixel i.e. to set a threshold pixel intensity. A radial line pixel intensity’ profile in some cases indicates pixel intensities across at least a portion of the outer cladding ‘jacket’ of the ARF 900. In this way, for example, a change in distance from the camera (indicated for example by pixel intensities) of points along a line across the outer cladding ‘jacket’ indicates a slope of the outer cladding ‘jacket’ and therefore a cleave angle and therefore a cleave profile. In some cases, a first radial line pixel intensity of an image of the ARF 900 at a first angle relative to a second radial line pixel intensity’ of the image at a second angle is used.

[0075] In various examples, a radial line pixel intensity is determined, which is a total intensity across a radial line. A first radial line pixel intensity is in some cases determined relative to a second radial line pixel intensity at a different angle to the first radial line pixel intensity (and in some cases both radial line pixel intensities being from a centre of theDocket No. 13768-4558a_502046-PCT01 imaged fibre), where a lower radial line pixel intensity at a first angle compared to a second angle (and where the radial line pixel intensity is determined from the same point and with a same line length, in some cases to the edge of the imaged fibre or the edge of the image) means that the imaged fibre towards the first angle is on average further away from the camera than the imaged fibre towards the second angle, providing an indication of the cleave profile of the fibre. In some cases, a radial line pixel intensity or radial line pixel intensity profile is determined for angles from zero to 360 degrees, or otherwise located across the entire range of angles, in some cases in equal angular increments. In some examples, a radial line pixel intensity or profile is determined for a line across a capillary, or a relative radial line pixel intensity or profile is determined for a first line across a capillary and a second line across a capillary, for example such lines being across a same capillary and each radial line having a different angle relative to a same point such as a centre of the same capillary. A radial line is in various examples used to determine a width or relative width of a pixel intensity profile as mentioned above with respect to a line profile.

[0076] In various examples, a line pixel intensity and / or a line pixel intensity profile across at least a portion of a closed loop of pixels of an image, in various examples being pixels depicting at least a portion of an ARF 900, such as a same capillary, are further used as a feature of the image representing a characteristic of the ARF 900. In some examples a relative line pixel intensity or profile, such as a relative circumferential line pixel intensity and / or relative circumferential line pixel intensity profile, are used as a feature of the image representing a characteristic of the ARF 900. Circular line 905, defined by a radial distance from a centre of the ARF 900 depicted in an image or from any other point, illustrates an exemplary closed loop of pixels, at least a portion across which a line pixel intensity and / or profile are determined in some cases. A line pixel intensity is a total intensity' across at least a portion of the closed loop line, for example line 905, and in various examples is a relative pixel intensity i.e. is compared to a line pixel intensity of at least a portion of a different closed loop line, i.e. including at least one different pixel.

[0077] In addition or alternative to the above-mentioned features, at least one visual characteristic of a tear in the cladding, in some cases the outer cladding, of the ARF 900 is determined. Such a visual characteristic in some cases includes a size, shape, thickness, step height or other characteristic. In some cases, such as when a cleave quality of the fibre is particularly low, such as when a cleave angle is particularly high, a noticeable step is present in the outer cladding of a hollow core fibre at the tear. As such, the end face surface of the outer cladding is in some cases discontinuous. The height of this step i.e. the distance of theDocket No. 13768-4558a_502046-PCT01 higher surface one side of the step boundary (i.e. the tear) from the lower surface on the other side of the step boundary, is referred to as a step height. Though step height is mentioned herein, this in some cases refers to an indication of a step height, such as a label, value, or range of values. Additionally, the tear includes in some cases a triangular structure which has its shape, size, thickness or other characteristic determined. The visual characteristic is determined in some cases to be in terms of pixel distance, for example a ‘width of 5 pixels’ or ‘surface area of 10 pixels’, or in terms of actual distance, for example a ‘width of 10 microns’ or ‘surface area of 2 square-microns’. In some cases a pixel distance is mapped to an actual distance using one or more camera parameters.

[0078] In other cases, the visual characteristic is determined to be a category, for example a ‘category A tear’, ‘zig-zag tear’ and / or ‘triangular structure present’.

[0079] The visual characteristic of the tear indicates cleave profile, and it arises because of the cleaving process. A thinner and smaller tear arises with a higher quality cleave and / or a lower cleave angle (e.g. a flatter surface), whereas a lower quality cleave and / or a higher cleave angle results in a wider and larger tear. Additionally, the triangular structure associated with the tear arises due to a combination of waves in the cladding originating from the cleaving blade during cleaving.

[0080] A visual characteristic of the tear is in various examples determined using a machine-learning model trained to process an input image i.e. pixel data of an input image, and output a visual characteristic of the tear. Training data in some cases comprises a plurality of input end-face images of cleaved ARFs. with a tear, and with various cleave profiles (e.g. at various cleave angles), and ground truth visual characteristics such as labels indicating ‘size of 3 pixels’, 'category A tear’ or other labels.

[0081] Alternatively, a machine-learning model is trained to process an input image of an ARF 900 and output boundaries of a tear or otherwise recognise a location of the tear, and then the disclosed technology compnses determining a visual characteristic of the tear using the determined boundaries or location. It should be noted that alternative methods of determining such visual characteristics as mentioned are in various examples used.

[0082] FIG. 10 illustrates a chart 1000 showing a correlation between an angular tear width on the y-axis and cleave profile, in this case being an indication of a cleave angle (i.e. the cleave profile has as associated value which is an indication of a cleave angle), of an imaged fibre on the x-axis, where the cleave profile is an indication of cleave quality. The dotted line shows a linear relationship between cleave profile (indicating cleave angle) and angular tear width derived from the points shown on the chart 1000. Each point on the chartDocket No. 13768-4558a_502046-PCT011000 represents an imaged cleaved hollow core fibre with an associated angular tear width and cleave profile indicating a cleave angle. As angular tear width increases, this is correlated with an increase in indicated cleave angle and therefore a lower-quality cleave. As such, by determining a visual characteristic of the tear, it is enabled that the visual characteristic be mapped to an indication of cleave quality using a model. FIG. 10 is merely exemplary, and similar correlations are present with respect to the features and indications of cleave quality as mentioned herein. Angular tear width refers in some cases to an angular distance across a length of the tear that is perpendicular to a length along a longest axis of the tear.

[0083] In addition to the above-mentioned features with respect to cleave angle, the mentioned techniques including pixel intensities, line profiles, and breadth of line profiles, and of visual characteristics of the tear, are in some cases used to determine a feature of an image that is mapped using a model to an indication of cleave quality that is not an indication of cleave profile. For example, a pixel intensity line profile across a capillary may have a drop in intensity close to the peak where cleaving has caused damage to the capillary, and this is a characteristic of an end face of an ARF that is represented by the line profile of an image of the ARF end face. As another example, a radial line pixel intensity profile with multiple peaks that are not associated with a capillary or the outer cladding of an imaged ARF indicates debris or other material on the end face, which can further indicate cleave quality. Any other feature is in various examples determined and mapped using a model to an indication of cleave quality.

[0084] Moreover, though an ARF is illustrated, it should be understood that in various examples the described techniques relate to a HCF, where capillaries relating to an ARF are instead microstructure elements of the HCF where present, which may include capillaries if these are present in the HCF, or where the techniques relating to capillaries of an ARF are not performed but where at least one of those techniques relating to structures also present in a HCF. such as the outer cladding, a tear in the outer cladding, potential debris and damage, are performed.

[0085] FIG. 11 illustrates a flow diagram of a method according to the herein disclosed technology. Method 1100 comprises receiving an end face image 1102 of a HCF, and analysing pixel data of the image to determine cleave quality of the HCF of the image by: determining at least one feature of the image representing a characteristic of and end face of the fibre 1104, using a model to map the at least one feature to an indication of cleave quality 1120 of the fibre, and outputting the indication of cleave quality' 1121.Docket No. 13768-4558a_502046-PCT01

[0086] The at least one feature of the image referred to in block 1104 comprises at least one of: a visual characteristic such as a width of a depicted tear in cladding of the imaged fibre i.e. ARF 1106, a step height of outer cladding of the imaged fibre across a tear 1107, a pixel intensity profile of pixels depicting a first microstructure element of the imaged fibre relative to a pixel intensity profile of pixels depicting a second microstructure element of the imaged fibre 1 108, a standard deviation of pixel intensities of pixels of the image depicting a first microstructure element of the imaged fibre relative to a standard deviation of pixel intensities of pixels of the image depicting a second microstructure element of the imaged fibre 1110. an average intensity of pixels depicting a first microstructure element of the imaged fibre relative to an average intensity of pixels of the image depicting a second microstructure element of the imaged fibre 1112, a relative intensity of a first pixel (or first pixels) depicting a first portion of the imaged fibre to a second pixel (or second pixels) depicting a second portion of the imaged fibre 1114, a radial line pixel intensity profile from a centre of the imaged fibre 11 16, a first radial line pixel intensity of pixels of the image relative to a second radial line pixel intensity of pixels of the image at a different angle to the first radial line pixel intensity, from a centre of the imaged fibre 1118, and a line pixel intensity profile across at least a portion of a closed loop of pixels of the image and / or a line pixel intensity across at least a portion of a first closed loop of pixels of the image relative to a line pixel intensity across at least a portion of a second closed loop of pixels of the image 1119. Additionally, method 1100 further comprises, in some cases, determining a relative position of a blade contact point relative to a tear in the outer cladding depicted in the image of the hollow core fibre, where this relative position is a feature of the end face image and is used alongside the at least one feature, to, using a model, map the relative position and the at least one feature to an indication of cleave quality. This relative position is in some cases an angular separation between the blade contact point and the tear. The relative position provides an indication of how waves from the cleave impact site propagated during cleaving, and in some cases improves an accuracy of an estimation of cleave quality as disclosed herein.

[0087] It should be appreciated that the cleave quality is in some cases determined for a HCF that does not have a tear in a cleaved end face, and in these cases a visual characteristic such as a width of a depicted tear in cladding of the imaged fibre i.e. ARF 1 106, and / or a step height of outer cladding of the imaged fibre across a tear 1107 are not comprised in the at least one feature of the image referred to in block 1104.Docket No. 13768-4558a_502046-PCT01

[0088] Method 1100 is a more specific method of method 608 of FIG. 6B, and in various examples comprises the alternatives or additions as mentioned herein with respect to FIG. 6B and the disclosed technology throughout.

[0089] An exemplary model for mapping at least one image feature representing a characteristic of an imaged fibre will now be provided. Though this exemplary model is in the form of rules in natural language, it should be appreciated that in various examples the model is in the form of a machine learning model, a lookup table, or at least one rule in natural language, machine code, or any other format. Additionally, the values provided are in some cases ranges, and in various examples are any other values instead of or in addition to those provided. Though determining cleave profile is illustrated, the model relates in some cases to a mapping to an indication of cleave quality, which encompasses an indication of cleave profile, and the rules in some cases include any features such as damage, debris, and / or visual characteristics of a tear.

[0090] Rule 1 : If tear thickness is 8um, cleave profile indicating cleave angle is within range of 0.0 to 0.6 degrees.

[0091] Rule 2: If line pixel intensity over an arc of pixels depicting outer cladding of the imaged fibre has a slope of -0.25 per degree, plus or minus 0.04 degrees, cleave profile indicating cleave angle is within range of 0.2 to 0.5 degrees.

[0092] Rule 3: If a difference between a peak of a pixel line intensity7profile across a capillary with a highest peak pixel line intensity of the capillaries and a peak of a pixel line intensity profile across a capillary with a lowest peak pixel line intensity is smaller than 0. 1 times the peak pixel line intensity of the capillary with the lowest peak pixel line intensity, cleave profile indicating cleave angle is within a range of 0.1 to 0.3 degrees.

[0093] Rule 4: If estimated cleave profile indicates angle above 1.0 degree, indicate FAILED cleave profile. Otherwise indicate PASSED cleave profile.

[0094] Combining ranges, and in some cases probabilities, of a plurality of rules, in some cases all rules, provides a resultant estimated cleave profile. In the case of the exemplary model above, and assuming each of the conditions of Rules 1-3 is met, the estimated cleave profile indicates an angle of 0.2 - 0.3 degrees. As this range is lower than the rule 4 threshold of 1.0 degree, an indication of a cleave profile is generated, and therefore an indication of cleave quality.

[0095] Further exemplary components of a model illustrated as natural language rules are:Docket No. 13768-4558a_502046-PCT01

[0096] Rule 1: If a difference between a pixel radial line intensity at a first angle and from a centre of an imaged ARF and a pixel radial line intensity at a second angle and from the centre of the imaged ARF is greater than 60% of the pixel radial line intensity at the first angle, indicate FAILED cleave profile.

[0097] Rule 2: If width at widest point of tear is smaller than 5.5 microns, go to ‘Rule 3’.

[0098] Rule 3: If a difference between, for any combination of first and second primary capillaries, a pixel line intensity over a first primary capillary’ and a pixel line intensity over a second primary capillary is greater than 105, indicate FAILED cleave profile.

[0099] In addition to the above-mentioned features, the disclosed technology in various examples comprises determining an indication of end face quality and in some cases a quality control check stage.

[0100] FIG. 12A illustrates a portion of a transverse cross-sectional view of a cleaved antiresonant hollow core optical fibre with damage and / or contamination that is analysed according to the herein disclosed technology.

[0101] Cleaved ARF 1200 comprises microstructure capillaries 1206 visible in an end face view as illustrated and as described thus far, but additionally comprises surface damage 1202 and visible liquid drops from the cleaving process. Additionally, the hollow core of the ARF 1200 has debris 1208 located within the hollow core. The capillary’ structure 1207 is damaged, with the capillary not having a continuous, smooth structure, and there is debris 1209 in the outer cladding close to a location where an inner boundary of the outer cladding of the ARF 1200 has been breached such that there is no longer a smooth boundary’ between the outer cladding and the inner core of the ARF 1200. Though damage is illustrated on a single capillary 1207, in various examples at least one capillary depicted in an end face image is damaged or destroyed due to a cleave.

[0102] Additionally, in some cases, as noted above, there is a step having a step height of the outer cladding of a hollow core fibre across a tear in the outer cladding. If the step height is larger than a threshold height, an open channel is in some cases created between the inner core of the HCF and the outside of the HCF, for example after splicing the HCF (i.e. splicing the HCF with another fibre). This open channel in some cases negatively affects optical properties of the splice such as an optical propagation loss by allowing ingress of debris, liquid and / or gasses via the channel.Docket No. 13768-4558a_502046-PCT01

[0103] Ideally, to reduce unwanted increases in loss over a splice, an end face is minimally surface damaged, microstructure such as capillaries remains intact, the end face does not have liquid or other contamination present, and a hollow core does not have debris within it, which would scatter light transmitted via the hollow core. Additionally, a round shape of the fibre itself is in some cases desired, and / or a step height across a tear is minimal.

[0104] As such, determining an indication of end face quality and in some cases a quality7control check is in various examples implemented according to the herein disclosed technology.

[0105] FIG. 12B illustrates a flow diagram of a method including a quality control check according to the herein disclosed technology. Method 1210 comprises receiving an end face image 1212 of a cleaved HCF, and analysing pixel data of the end face image to determine cleave quality of the HCF by determining at least one feature of the image representing a characteristic of an face of the HCF and using a model to map the at least one feature to an indication of cleave quality 1214. This method corresponds to method 1100 of FIG. 11 and method 608 of FIG. 6B, with the addition in method 1210 of a quality control check 1218. The quality control check in various examples is performed prior to analysis of the pixel data to determine a cleave quality 1214, after analysis of the pixel data to determine a cleave quality 1214, or during analysis of the pixel data to determine a cleave quality 1214.

[0106] In some cases, the quality control check 1218 determines at least one feature of the image representing a characteristic of an end face of the imaged fibre, where the model maps the at least one feature determined by the quality control check 1218. in some cases among at least one other determined feature, to an indication of cleave quality. In this way, blocks 1218 and 1214 arejointly performed.

[0107] In some cases, including the quality control check 1218 refers to including it as part of the model as described herein, i.e. where blocks 1218 and 1214 arejointly performed. In these cases, the indication of cleave quality comprises an indication of an end face cleave quality determined by analysing pixel data of the image and using the model to map at least one end face quality characteristic to the indication of end face cleave quality7, which in some examples is combined with an indication of cleave profile to determine an indication of cleave quality, such as by jointly mapping, using the model as described herein to an indication of cleave quality, the at least one end face quality7characteristic and at least one feature usable to map, using the model as described herein, the at least one feature to an indication of cleave profile.Docket No. 13768-4558a_502046-PCT01

[0108] Additionally or alternatively, in some cases the quality control check 1218 is used as part of determining an indication of cleave profile and therefore an indication of cleave quality. In these cases, the indication of cleave quality' comprises an indication of cleave profile, and using the model to map the at least one feature to the indication of cleave quality comprises using the model to map at least one end face quality characteristic, in some cases jointly with at least one feature usable to map, using the model as described herein, the at least one feature to an indication of cleave profile, to the indication of cleave profile and therefore the indication of cleave quality .

[0109] Additionally or alternatively, the quality control check 1218 is used as a prescreen to decide whether or not to perform block 1214, i.e. a cleave profile determination or more broadly a cleave quality determination. For example, where the quality control check 1218 fails, a fibre is in some cases discarded or re-cleaved, which saves determining an unacceptable cleave profile. Also, a damaged or contaminated fibre may make determining an indication of cleave quality, for example an indication of cleave profile, more inaccurate. Where a quality control check 1218 passes, a fibre is spliced or has its cleave quality estimation determined. In these cases, method 1210 further comprises analysing pixel data of the image (received in block 1212) to determine an indication of end face quality of the fibre prior to the analysing pixel data of the image to determine cleave quality of the fibre, the end face quality comprising at least one end face quality' characteristic, wherein the analysing pixel data of the image to determine cleave quality of the fibre (i.e. block 1214) is performed in response to the indication of end face quality indicating an end face quality above a threshold quality.[001 1 0] In some examples, wherein the method 1210 comprises analysing pixel data of the image to determine an indication of end face quality of the fibre prior to the analysing pixel data of the image to determine cleave quality of the fibre (i.e. comprises block 1218 prior to block 1214). the method further comprises in response to the determined indication of end face quality indicating a quality’ below a threshold quality, performing one of: initiating discarding of the fibre, and initiating re-cleaving of the fibre.[001 1 1 ] Method 1210 further comprises outputting the indication of cleave quality1216, such as by providing the indication of cleave quality to, and / or initiating an action at, an apparatus such as a cleaver and / or a splicer.[001 1 2] An end face quality characteristic as referred to herein comprises at least one of: a level of damage to a fibre, a level of damage to a cleaved end face of a fibre, a level of debris present on an end face of a fibre, a level of liquid present on an end face of a fibre, aDocket No. 13768-4558a_502046-PCT01 shape of a fibre, a shape of an outer surface such as an outer surface of outer cladding of a fibre, for example whether the outer surface of the cladding is substantially round or circular or whether the outer surface is continuous or smooth, a level of surface damage, a level of blade contact point damage, whether tear damage is unacceptable, whether a mirror region is undamaged or a level of mirror region damage, whether liquid or other contamination is present on the end face of the imaged fibre, whether a mist region is undamaged or a level of mist region damage, whether microstructure elements are undamaged or a level of microstructure damage, a step height of outer cladding across a tear and / or whether debris is present in a core of an imaged fibre. A level of a feature in this context refers in some cases to an indication of the feature.[001 1 3] In some cases, a bad cleave results in damage to the microstructure of aHCF such as an ARF, such as breaks, glass debris, and / or leaning capillaries (i.e. capillaries which move from their constructed positions) which significantly increases losses in the fibre. Additionally, the outer cladding 'jacket’ of a HCF is in some cases broken, has multiple tears, has an uneven surface or other damages are present from a bad cleave.[001 1 4] In some cases, the quality control check 1218 is performed using a machine learning model trained to process an input image of a HCF and identify the above-mentioned end face quality characteristics. Alternatively, the machine learning model is trained to process an input image of a HCF and identify boundaries or an amount of surface damage elements, debris elements or other fibre characteristics, and then the above mentioned end face quality characteristics are determined using the information output by the machine learning model.[001 1 5] Though machine-learning models are mentioned throughout in reference to examples of how features of an image and / or a fibre are determined, it should be appreciated that alternative methods are in various examples used, such as the application of a set of a rules defining for example sequences of pixel intensities that correspond to a specific feature of the image and fibre, or the application of image filters based on pixel values to determine the features.[001 1 6] Where the quality control check 1218 is used to determine a factor for mapping using a model to a cleave profile, this is because surface damage or the other above mentioned determined factors, for example, are in some cases indicative of a cleave profile, especially a cleave profile indicating a high cleave angle such as 2, 3, 5, 10 degrees or any other cleave angle that would result, when the cleaved HCF is spliced, in a splice unsuitable for an application of the spliced HCF.Docket No. 13768-4558a_502046-PCT01[001 1 7] Where reference is made to a threshold quality, such a threshold quality is in various examples predetermined, and refers to a quality indicating a fibre suitable for splicing (i.e. splicing with another fibre) in the sense of a resulting splice having an acceptable loss over the splice when light is directed through the spliced fibre.[001 1 8] By including a quality control check, the disclosed technology further enables a low loss and robust splice joint.[001 1 9] Though the description thus far has focused on using a model to map at least one feature of an end face image to an indication of cleave quality’, in some cases being or comprising an indication of cleave profile, where the model in various examples is a machine-learning model, at least one rule, or a lookup table, or any combination of those, the model used in the methods disclosed herein in an example is created as illustrated in FIG. 13. [001 20] Method 1300 comprises receiving end face images and, optionally, side view images 1302, in some cases a single end face and / or side view image, the images depicting at least one HCF. The method 1300 further comprises determining at least one feature of a received end face image 1304, in some cases in the same way as described herein according to the disclosed technology, for example as in method 608 of FIG. 6B. Additionally, method 1300 comprises determining a ground truth indication of cleave angle or cleave profile 1305 of an imaged HCF (where ground truth refers herein to a ‘true’ value i.e. a ’target’ value of a model, and is not limited to machine-learning model contexts), such as by using the methods as described herein with respect to analy sing pixel data of an image, by receiving the ground truth indication of cleave angle or cleave profile, by analysing an image taken by a depth camera of an end face of a HCF, or by any other method. Determining a ground truth indication of cleave angle comprises determining a ground truth indication of cleave angle from a side view image 1306, and determining a ground truth indication of cleave profile comprises determining an indication of cleave profile from an end face image of the imaged HCF taken using a depth camera 1306. In some cases, an indication of a cleave profile and / or angle is received by the method 1300 and the determination does not occur. In various examples, instead of the determination of at least one feature of an end face image, the at least one feature is received by the method 1300. In various examples, the method 1300 is repeated with further end face and / or side view images until training or creation of the model is complete, such as after an elapsed time, after a number of images are processed, after an accuracy is achieved when testing the model by comparing a mapped indication of cleave quality from a feature of an end face image, to an associated ground truth indication of cleaveDocket No. 13768-4558a_502046-PCT01 quality comprising an indication of cleave angle and / or cleave profile, or until any other condition is met.[001 21 ] Finally, method 1300 comprises training a machine-learning model using training inputs each comprising a determined at least one feature and a ground truth indication of cleave angle and / or profile 1308, defining at least one rule mapping a determined at least one feature to an indication of cleave angle and / or profile 1310, and / or defining a lookup table by associating a determined at least one feature with an indication of cleave angle and / or profile. Though the model is in some examples trained or defined to associate or generate an indication of cleave angle and / or profile, in some examples the model is trained or defined to associate or generate an indication of a cleave quality, including an indication of cleave profile and / or cleave angle, the indication of cleave profile and in some cases the indication of cleave angle comprising an indication of an estimated range of cleave angles and / or a probability of a cleave angle or range of cleave angles, for example by using the methods described herein in relation to creating a model with indications of cleave angle and / or profile, and instead or additionally performing creating the model with indications of cleave quality'.[001 22] Optionally, method 1300 further comprises initiating using the model to map at least one feature of an image of an end face of a cleave hollow core optical fibre, the feature representing a characteristic of the end face of the fibre, to an indication of cleave quality' and initiating outputting the indication of cleave quality'. As such, method 1300 is in some cases used to create i.e. define the model used to map a feature to an indication of cleave quality in the methods as described herein.[001 23] In various examples, training the machine learning model 1308 comprises using training inputs each comprising at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, and the training inputs each further comprising at least one of: a ground truth indication of cleave angle, determined from a side view image of a same test fibre as the training input is associated with; and a ground truth indication of cleave profile, determined from an end face image taken using a depth camera, of a same test fibre as the training input is associated with.[001 24] In various examples, defining the at least one rule 1310 comprises defining at least one rule mapping at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, to at least one of: an indication of cleave angle determined from a side viewDocket No. 13768-4558a_502046-PCT01 image of the test fibre; and an indication of cleave profile determined from an end face image, taken using a depth camera, of the test fibre.[001 25] In various examples, defining the lookup table 1312 comprises defining a lookup table by associating at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, with an indication of at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre.[001 26] In some cases, the indication of cleave angle determined from a side view image of a test fibre is determined to be an average of cleave angles determined from a plurality of side view images of the test fibre, the plurality of side view images being of the test fibre with at least two different, relative to a camera that produced the side view images (i.e. such that the camera images a same test fibre with different rotation angles about a central axis), rotation angles about an axis central to a core of the test fibre. In other cases, the indication of cleave angle and / or cleave profile is instead of being determined by the method 1300, received by the method 1300, for example having been determined elsewhere, such as in a splicer, and / or by using a depth camera image such as an interferometer image.[001 27] It should be noted that method 1300 is in various examples performed by apparatus comprising or comprised in a splicer, a cleaver, or that is independent of a splicer and a cleaver.[001 28] Overall, the herein disclosed analysis of an image of an end face of a cleaved hollow core optical fibre to determine an indication of cleave quality by determining at least one feature of the image, the feature representing a characteristic of an end face of the fibre, and using a model to map the at least one feature to an indication of cleave quality, operates in an unconventional manner to achieve a method of estimating cleave quality with a higher accuracy and that is more robust than in alternative approaches, whilst enabling a more efficient splicing process. By taking into account end face features rather than determining a cleave quality via side view images, this results in a more accurate estimation of a cleave quality.[001 29] Where reference is made herein to ARFs, it should be noted that the disclosed technology may alternatively be used with other types of hollow core fibre (HCF). Where capillaries are referred to throughout, these in some cases refer instead to elements of a microstructure of a HCF, such as structures used to create a core of the HCF, structures present in a core of the HCF, nested structures or any other microstructure. The techniques asDocket No. 13768-4558a_502046-PCT01 described herein in relation to cladding apply also to HCFs. Additionally, it will be evident that the techniques described herein including those relating to pixel intensities, line profile widths, and apparatus for cleaving and splicing apply not only to ARFs but also to HCFs. Where a technique relates to a capillary, in a HCF this relates instead to a microstructure element of the HCF such as one located in a core of the HCF.[001 30] Additionally, the herein described techniques, including relative line pixel intensities or profiles, are in various examples performed whilst maintaining a focus i.e. a position of a focal plane, of a camera used to image a HCF, in some cases being an ARF. In other examples, a focus of the camera is adjusted, for example automatically, and line pixel intensities or profiles determined for multiple focuses, for example in two different images. In this way, more information and more accurate information is enabled to be gleaned regarding the cleave quality7of the imaged fibre at least because more information about the cleave profile is able to be determined.[001 31 ] FIG. 14 illustrates an exemplary computing-based device in which embodiments of a cleave quality estimation method according to the herein disclosed technology are implemented. Exemplary7computing-based device 1400 is implemented as any form of a computing and / or electronic device, and comprises one or more processors 1402 which are microprocessors, controllers or any other suitable type of processors for processing computer executable instructions to control the operation of the device in order to perform the herein disclosed methods including of FIGS. 6B, 11 and 12B.[001 32] In some examples, for example where a system on a chip architecture is used, the processors 1402 include one or more fixed function blocks (also referred to as accelerators) which implement a part of the method of FIGS. 6B, 11 and 12B in hardware (rather than software or firmware). Platform software comprising an operating system or any other suitable platform software is in some cases provided at the computing-based device 1400 to enable application software such as image analysis software 1406 to be executed on the device.[001 33] Image analysis software 1406 implements the methods of FIGS. 6B, 11 and 12B, and the functionality described herein. Optionally, memory 1404 comprises model creation software that implements the method of FIG. 13. in addition or alternative to the image analysis software 1406 which implements the methods of FIGS. 6B, 11 and 12B.[001 34] The computer executable instructions are provided using any computer- readable media that is accessible by computing based device 1400. Computer-readable media includes, for example, computer storage media such as memory 1404 andDocket No. 13768-4558a_502046-PCT01 communications media. Computer storage media, such as memory 1404, includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or the like. Computer storage media includes, but is not limited to, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM), electronic erasable programmable read only memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), digital versatile disks (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that is used to store information for access by a computing device. In contrast, communication media embody computer readable instructions, data structures, program modules, or the like in a modulated data signal, such as a carrier wave, or other transport mechanism. As defined herein, computer storage media does not include communication media. Therefore, a computer storage medium should not be interpreted to be a propagating signal per se. Although the computer storage media (memory 1404) is shown within the computing-based device 1400 it will be appreciated that the storage is, in some examples, distributed or located remotely and accessed via a network or other communication link (e.g. using communication interface 1414).[001 35] Memory 1404 optionally comprises cleave quality store 1408 and / or image store 1410, which store determined estimated cleave quality and received end face images respectively, for the image analysis software 1406.[001 36] The computing-based device 1400 also comprises an input / output interface 1412 arranged to in some cases output display information to a display device which may be separate from or integral to the computing-based device 1400. The display information may provide a graphical user interface. The input / output interface 1412 is also arranged to receive and process input from one or more devices, such as a user input device (e.g. a mouse, keyboard, camera, microphone or other sensor), a cleaver, a camera, and / or splicer.[001 37] The input / output interface 1412 outputs data to devices other than the display device in some examples, e.g. a splicer and / or cleaver.[001 38] As will be appreciated from the above description, computing-based device 1400 in various examples comprises or is comprised in a cleaver, a splicer, or is independent of a cleave and a splicer.[001 39] Alternatively or in addition to the other examples described herein, examples include any combination of the following:Docket No. 13768-4558a_502046-PCT01[001 40] Clause A. An apparatus comprising: a processor; and a memor\' storing instructions that, when executed by the processor, perform a method for estimating cleave quality of a cleaved hollow core optical fibre, the method comprising: receiving an image of an end face of the fibre; and analysing pixel data of the image to determine cleave quality of the fibre, by: determining at least one feature of the image, the feature representing a characteristic of the end face of the fibre; and using a model to map the at least one feature to an indication of cleave quality; and outputting the indication of cleave quality.[001 41 ] In this way. an apparatus that estimates cleave quality in an efficient and accurate manner, using an end face image is provided. Costly depth camera equipment is not required, and the apparatus is capable of integration in a cleaving and / or splicing process or independently of such processes.[001 42] The cleave quality may comprise a cleave profile, and wherein the indication of cleave quality comprises an indication of a cleave profile. In this way, a physical measurement of the end face of the hollow core fibre is determined in an accurate and efficient manner, where the determination of cleave profile is accurate, efficient, and does not require costly and complex depth imaging apparatus.[001 43] Clause B. The apparatus of Clause A, wherein outputting the indication of cleave quality comprises providing the indication of cleave quality to at least one of: a cleaver, and a splicer, for determining whether to perform an action associated with the fibre. In this way, the determination of whether to perform an action is based on accurate and efficiently obtained information, improving the process of performing the action in the cleaver and / or splicer.[001 44] Clause C. The apparatus of any preceding Clause, wherein the method further comprises at least one of: initiating splicing of the fibre at a splicer in response to the indication of cleave quality indicating a quality above a threshold; and in response to the indication of cleave quality indicating a quality above a threshold, initiating one of: re-cleaving of the fibre at a cleaver, and initiating discarding the fibre.Docket No. 13768-4558a_502046-PCT01[001 45] In this way, an accurate and efficiently obtained indication of cleave quality further improves an efficiency of a splicing or cleaving process, as splicing is attempted where the splice is accurately and efficiently predicted to have a quality' above a threshold, therefore reducing instances of splicing that unexpectedly result in a poor quality splice. Additionally, initiating re-cleaving or discarding of the fibre improves an efficiency of a splicing / cleaving process by preventing poor quality splices.[001 46] Clause D. The apparatus of any preceding Clause, the apparatus further comprising a camera, and the method further comprising imaging an end face of the fibre directly, using the camera, wherein receiving an image of an end face of the fibre comprises receiving an image from the camera. In this way, an expensive and complex imaging apparatus is not required, and the apparatus is easily capable of integration into a splicer / cleaving pipeline.[001 47] Clause E. The apparatus of any preceding Clause, the apparatus one of: further comprising a cleaver, being independent of a splicer and a cleaver. In this way, the apparatus is integrated into an existing cleaver and provides efficiency improvements by reducing instances of an unsuitable fibre being inserted into a splicer and only then being determined to be unsuitable and having to be removed or resulting in a low quality splice. Being independent of a splicer and a cleaver enables further efficiency improvements by enabling estimating cleave quality' in parallel with splicing and / or cleaving of other fibres.[001 48] Clause F. A computer implemented method for estimating cleave quality of a cleaved hollow core optical fibre, the method comprising: receiving an image of an end face of the fibre; and analysing pixel data of the image to determine cleave quality of the fibre, by: determining at least one feature of the image, the feature representing a characteristic of the end face of the fibre; and using a model to map the at least one feature to an indication of cleave quality; and outputting the indication of cleave quality.[001 49] In this way. a method that estimates cleave quality in an efficient and accurate manner, using an end face image is provided. Costly depth camera equipment is not required, and the method is capable of integration in a cleaving and / or splicing process or independently of such processes.Docket No. 13768-4558a_502046-PCT01[001 50] Clause G. The method of Clause F, wherein cleave quality comprises a cleave profile, and wherein the indication of cleave quality comprises an indication of a cleave profile. In this way, a physical measurement of the end face of the hollow core fibre is determined in an accurate and efficient manner, where the determination of cleave profile is accurate, efficient, and does not require costly and complex depth imaging apparatus.[001 51 ] Clause H. The method of any of Clauses F to G, wherein the image is a single image. In this way, a cleave quality is determined from only a single image, therefore providing an efficient method of determining cleave quality.[001 52] Clause I. The method of any of Clauses F to H, wherein outputting the indication of cleave quality comprises providing the indication of cleave quality to at least one of: a cleaver, and a splicer, for determining whether to perform an action associated with the fibre. In this way, the determination of whether to perform an action is based on accurate and efficiently obtained information, improving the process of performing the action in the cleaver and / or splicer.[001 53] Clause J. The method of any of Clauses F to I, further comprising at least one of: initiating splicing of the fibre at a splicer in response to the indication of cleave quality indicating a quality above a threshold; and in response to the indication of cleave quality indicating a quality above a threshold, initiating one of: re-cleaving of the fibre at a cleaver, and initiating discarding the fibre.[001 54] In this way, an accurate and efficiently obtained indication of cleave quality further improves an efficiency of a splicing or cleaving process, as splicing is attempted where the splice is accurately and efficiently predicted to have a quality above a threshold, therefore reducing instances of splicing that unexpectedly result in a poor quality splice. Additionally, initiating re-cleaving or discarding of the fibre improves an efficiency of a splicing / cleaving process by preventing poor quality splices.[001 55] Clause K. The method of any of Clauses F to J, wherein the at least one feature comprises at least one of: a visual characteristic of a tear in cladding of the fibre; a step height of cladding of the fibre across a tear in cladding of the fibre;Docket No. 13768-4558a_502046-PCT01 a pixel intensity profile of pixels of the image depicting a first microstructure element of the fibre relative to a pixel intensity profile of pixels of the image depicting a second microstructure element of the fibre; a standard deviation of pixel intensities of pixels of the image depicting a first microstructure element of the fibre relative a standard deviation of pixel intensities of pixels of the image depicting a second microstructure element of the fibre; an average intensity of pixels of the image depicting a first microstructure element of the fibre relative to an average intensity of pixels of the image depicting a second microstructure element of the fibre; a relative intensity of a first pixel depicting a first portion of the fibre to a second pixel depicting a second portion of the fibre; a radial line pixel intensity profile of the image; a first radial line pixel intensity of the image at a first angle relative to a second radial line pixel intensity of the image at a second angle; a line pixel intensity profile across at least a portion of a closed loop of pixels of the image; and a line pixel intensity across at least a portion of a first closed loop of pixels of the image relative to a line pixel intensity across at least a portion of a second closed loop of pixels of the image.[001 56] In this way, particularly useful features are provided which provide a comprehensive, accurate and efficient way of mapping an indication of cleave quality based on an end face image of a cleaved hollow core fibre.[001 57] Clause L. The method of any of Clauses F to K, wherein using the model to map the at least one feature to the indication of cleave quality comprises using the model to map the at least one feature and at least one parameter of a camera that took the image of the end face of the fibre, to an indication of cleave quality. In this way, an accuracy of the mapping to an indication of cleave quality is improved, as the parameter of the camera is likely to impact the pixel analysis performed.[001 58] Clause M. The method of any of Clauses F to L, wherein the model comprises at least one of: a machine-learning model that is trained to output an indication of cleave quality given the at least one feature as an input; at least one rule mapping the at least one feature to an indication of cleave quality; andDocket No. 13768-4558a_502046-PCT01 a lookup table associating the at least one feature with an indication of cleave quality. [001 59] In this way, specific examples of model components are provided that enable the mapping of a feature of an end face image to an indication of cleave quality.

[0160] Clause N. The method of Clause M, wherein the indication of cleave quality comprises an indication of at least one of: cleave profile and cleave angle, and wherein at least one of: the model comprises a machine-learning model and the machine-learning model is trained using training inputs each comprising at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, and the training inputs each further comprising at least one of: a ground truth indication of cleave profile, determined from an end face image, taken using a depth camera, of a same test fibre as the training input is associated with; and a ground truth indication of cleave angle determined from a side view image of a same test fibre as the training input is associated w ith; the model comprises at least one rule mapping the at least one feature to an indication of a cleave profile and the model is created at least by defining the at least one rule to map at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, to an indication of at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre; and the model comprises a lookup table associating the at least one feature with an indication of a cleave profile and the lookup table is created by associating at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, with an indication of at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken by a depth camera, of the test fibre.

[0161] In this w ay, a specific w ay of constructing the model used to map an image feature to an indication of cleave quality is provided. This specific way provides an accurateDocket No. 13768-4558a_502046-PCT01 way of determining an indication of cleave profile and therefore an indication of cleave quality.[001 62] Clause O. The method of any of Clauses F to N, wherein at least one of: the indication of cleave quality comprises an indication of an end face quality, the indication of end face quality determined by analysing pixel data of the image and using the model to map at least one of: a level of damage to the fibre, a level of damage to a cleaved end face of the fibre, a level of debris present in a core of the fibre, an indication of step height of cladding of the fibre across a tear in cladding of the fibre, a level of debris present on the end face of the fibre, a shape of an outer surface of the fibre, and a level of liquid present on the end face of the fibre, to the indication of end face cleave quality: the indication of cleave quality comprises an indication of cleave profile, and wherein using the model to map the at least one feature to the indication of cleave quality comprises using the model to map at least one of: a level of damage to the fibre, a level of damage to a cleaved end face of the fibre, a level of debris present in a core of the fibre, an indication of step height of cladding of the fibre across a tear in cladding of the fibre, a level of debris present on the end face of the fibre, a shape of an outer surface of the fibre, and a level of liquid present on the end face of the fibre, to the indication of cleave profile and therefore the indication of cleave quality; and the method further comprises analysing pixel data of the image to determine an indication of end face quality of the fibre prior to the analysing pixel data of the image to determine cleave quality of the fibre, the end face quality comprising at least one of: a level of damage to the fibre, a level of damage to a cleaved end face of the fibre, an indication of step height of cladding of the fibre across a tear in cladding of the fibre, a level of debris present in a core of the fibre, a level of debris present on the end face of the fibre, a shape of an outer surface of the fibre, and a level of liquid present on the end face of the fibre, wherein the analysing pixel data of the image to determine cleave quality of the fibre is performed in response to the indication of end face quality indicating an end face quality above a threshold quality.[001 63] In this way, an accurate and efficient way of estimating a cleave quality is provided, by taking into account specific characteristics of the end face fibre that provide a comprehensive, accurate and efficient way of mapping such characteristics to an indication of cleave quality. Additionally, where the method comprises analysing pixel data of the image to determine an indication of end face quality of the fibre prior to the analysing pixel data of the image to determine cleave quality of the fibre, this reduces instances of performingDocket No. 13768-4558a_502046-PCT01 redundant analysis such as to determine an indication of cleave profile, where a cleave quality is below a threshold determined in a ‘pre-check’. As such, further efficiency improvements are provided.[001 64] Clause P. The method of Clause O, wherein the method comprises analysing pixel data of the image to determine an indication of end face quality of the fibre prior to the analysing pixel data of the image to determine cleave quality of the fibre, and wherein the method further comprises: in response to the indication of end face quality indicating a quality below a threshold quality, performing one of: initiating discarding of the fibre, and initiating re-cleaving of the fibre.[001 65] In this way, should a pre-check of cleave quality prior to the primary analysis of pixel data indicate a low quality cleave, initiating re-cleaving or discarding reduces instances of poor quality splices and in the case of re-cleaving provides a chance of improving a cleave quality, where the 'second chance’ is determined in an efficient and accurate manner.[001 66] Clause Q. A computer readable medium storing instructions that, when executed by a computer, perform the method of any of Clauses F to P.[001 67] Clause R. A method for creating a model for mapping at least one feature of an image of an end face of a cleaved hollow core optical fibre to an indication of cleave quality of the fibre, the indication of cleave quality' comprising an indication of at least one of: cleave angle and cleave profile, the method comprising at least one of:1) training a machine-learning model, the training comprising using training inputs each comprising at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, and the training inputs each further comprising at least one of: a ground truth indication of cleave angle, determined from a side view image of a same test fibre as the training input is associated with; and a ground truth indication of cleave profile, determined from an end face image, taken using a depth camera, of a same test fibre as the training input is associated with;2) defining at least one rule mapping at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, to an indication of at least one of: a cleave angle determined from a side view image of the test fibre; andDocket No. 13768-4558a_502046-PCT01 a cleave profile determined from an end face image, taken using a depth camera, of the test fibre; and3) defining a lookup table by associating at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, with an indication of at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre.[001 68] In this way. a method of constructing a model capable of being used in the methods and by the apparatus of prior Clauses is provided. This model enables the mapping of a feature of an end face image to an indication of cleave quality in an accurate and efficient way.[001 69] Clause S. The method of Clause R. wherein the indication of cleave angle determined from a side view image of a test fibre is determined to be an average of cleave angles determined from a plurality of side view images of the test fibre, the plurality of side view images being of the test fibre with at least two different, relative to a camera that produced the side view images, rotation angles about an axis central to a core of the test fibre. In this way, an accurate way of determining a ground truth indication of cleave profile and therefore of cleave quality is provided.[001 70] Clause T. The method of any of Clauses R and S, further comprising: initiating using the model to map at least one feature of an image of an end face of a cleaved hollow core optical fibre, the feature representing a characteristic of the end face, to an indication of cleave quality and outputting the indication of cleave quality.[001 71 ] In this way, the model with its associated advantages is initiated to be used such that a cleave quality that is accurately and efficiently determined from an end face image of a cleaved hollow core fibre is output.[001 72] The term ‘computer’ or ‘computing-based device’ is used herein to refer to any device with processing capability7such that it executes instructions. Those skilled in the art will realize that such processing capabilities are incorporated into many different devices and therefore the terms ‘computer’ and ‘computing-based device’ each include personal computers (PCs), servers, mobile telephones (including smart phones), tablet computers, set- top boxes, media players, games consoles, personal digital assistants, wearable computers, and many other devices.Docket No. 13768-4558a_502046-PCT01[001 73] The methods described herein are performed, in some examples, by software in machine readable form on a tangible storage medium e.g. in the form of a computer program comprising computer program code means adapted to perform all the operations of one or more of the methods described herein when the program is run on a computer and where the computer program may be embodied on a computer readable medium. The software is suitable for execution on a parallel processor or a serial processor such that the method operations may be carried out in any suitable order, or simultaneously.[001 74] Those skilled in the art will realize that storage devices utilized to store program instructions are optionally distributed across a network. For example, a remote computer is able to store an example of the process described as software. A local or terminal computer is able to access the remote computer and download a part or all of the softw are to run the program. Alternatively, the local computer may dow nload pieces of the softw are as needed, or execute some software instructions at the local terminal and some at the remote computer (or computer network). Those skilled in the art will also realize that by utilizing conventional techniques know n to those skilled in the art that all, or a portion of the software instructions may be carried out by a dedicated circuit, such as a digital signal processor (DSP), programmable logic array, or the like.[001 75] Any range or device value given herein may be extended or altered without losing the effect sought, as will be apparent to the skilled person.[001 76] Although the subject matter has been described in language specific to structural features and / or methodological acts, it is to be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are disclosed as example forms of implementing the claims.[001 77] It will be understood that the benefits and advantages described above may relate to one embodiment or may relate to several embodiments. The embodiments are not limited to those that solve any or all of the stated problems or those that have any or all of the stated benefits and advantages. It will further be understood that reference to ‘an’ item refers to one or more of those items.[001 78] The operations of the methods described herein may be carried out in any suitable order, or simultaneously where appropriate. Additionally, individual blocks may be deleted from any of the methods without departing from the scope of the subject matter described herein. Aspects of any of the examples described above may be combined withDocket No. 13768-4558a_502046-PCT01 aspects of any of the other examples described to form further examples without losing the effect sought.[001 79] The term ‘comprising’ is used herein to mean including the method blocks or elements identified, but that such blocks or elements do not comprise an exclusive list and a method or apparatus may contain additional blocks or elements.[001 80] It will be understood that the above description is given by way of example only and that various modifications may be made by those skilled in the art. The above specification, examples and data provide a complete description of the structure and use of exemplary embodiments. Although various embodiments have been described above with a certain degree of particularity, or with reference to one or more individual embodiments, those skilled in the art could make numerous alterations to the disclosed embodiments without departing from the scope of this specification.

Claims

Docket No. 13768-4558a_502046-PCT01CLAIMSWhat is claimed is:

1. An apparatus (1400) comprising: a processor (1402); and a memory (1404) storing instructions that, when executed by the processor, perform a method (608, 1100) for estimating cleave quality of a cleaved hollow core optical fibre, the method comprising: receiving an image of an end face of the fibre (610, 1102); and analysing pixel data of the image to determine cleave quality of the fibre, by: determining at least one feature of the image, the feature representing a characteristic of the end face of the fibre (612, 1104); and using a model to map the at least one feature to an indication of cleave quality (612. 1120); and outputting the indication of cleave quality (1121, 1216).

2. The apparatus of claim 1 , wherein outputting the indication of cleave quality comprises providing the indication of cleave quality to at least one of: a cleaver, and a splicer, for determining whether to perform an action associated with the fibre.

3. The apparatus of any preceding claim, wherein the method further comprises at least one of: initiating splicing of the fibre at a splicer in response to the indication of cleave quality indicating a quality above a threshold; and in response to the indication of cleave quality indicating a quality above a threshold, initiating one of: re-cleaving of the fibre at a cleaver, and initiating discarding the fibre.

4. The apparatus of any preceding claim, the apparatus further comprising a camera, and the method further comprising imaging an end face of the fibre directly, using the camera, wherein receiving an image of an end face of the fibre comprises receiving an image from the camera.

5. The apparatus of any preceding claim, the apparatus one of: further comprising a cleaver, being independent of a splicer and a cleaver.

6. A computer implemented method (608, 1100) for estimating cleave quality of a cleaved hollow core optical fibre, the method comprising: receiving an image of an end face of the fibre (610, 1102); andDocket No. 13768-4558a_502046-PCT01 analysing pixel data of the image to determine cleave quality of the fibre, by: determining at least one feature of the image, the feature representing a characteristic of the end face of the fibre (612, 1104); and using a model to map the at least one feature to an indication of cleave quality (612, 1120); and outputting the indication of cleave quality (1121, 1216).

7. The method of claim 6, wherein cleave quality comprises a cleave profile, and wherein the indication of cleave quality comprises an indication of a cleave profile.

8. The method of claim any of claims 6 and 7, wherein the image is a single image.

9. The method of any of claims 6 to 8, wherein outputting the indication of cleave quality comprises providing the indication of cleave quality to at least one of: a cleaver, and a splicer, for determining whether to perform an action associated with the fibre.

10. The method of any of claims 6 to 9, further comprising at least one of: initiating splicing of the fibre at a splicer in response to the indication of cleave quality indicating a quality above a threshold; and in response to the indication of cleave quality indicating a quality above a threshold, initiating one of: re-cleaving of the fibre at a cleaver, and initiating discarding the fibre.

11. The method of any of claims 6 to 10, wherein the at least one feature comprises at least one of: a visual characteristic of a tear in cladding of the fibre; a step height of cladding of the fibre across a tear in cladding of the fibre; a pixel intensity' profile of pixels of the image depicting a first microstructure element of the fibre relative to a pixel intensity profile of pixels of the image depicting a second microstructure element of the fibre; a standard deviation of pixel intensities of pixels of the image depicting a first microstructure element of the fibre relative a standard deviation of pixel intensities of pixels of the image depicting a second microstructure element of the fibre; an average intensity of pixels of the image depicting a first microstructure element of the fibre relative to an average intensity of pixels of the image depicting a second microstructure element of the fibre; a relative intensity of a first pixel depicting a first portion of the fibre to a second pixel depicting a second portion of the fibre;Docket No. 13768-4558a_502046-PCT01 a radial line pixel intensity profile of the image; a first radial line pixel intensity of the image at a first angle relative to a second radial line pixel intensity of the image at a second angle; a line pixel intensity profile across at least a portion of a closed loop of pixels of the image; and a line pixel intensity across at least a portion of a first closed loop of pixels of the image relative to a line pixel intensity across at least a portion of a second closed loop of pixels of the image.

12. The method of any of claims 6 to 11, wherein using the model to map the at least one feature to the indication of cleave quality comprises using the model to map the at least one feature and at least one parameter of a camera that took the image of the end face of the fibre, to an indication of cleave quality.

13. The method of any of claims 6 to 12, wherein the model comprises at least one of: a machine-learning model that is trained to output an indication of cleave quality given the at least one feature as an input; at least one rule mapping the at least one feature to an indication of cleave quality7; and a lookup table associating the at least one feature with an indication of cleave quality.

14. The method of claim 13, wherein the indication of cleave quality comprises an indication of at least one of: cleave profile and cleave angle, and wherein at least one of: the model comprises a machine-learning model and the machine-learning model is trained using training inputs each comprising at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, and the training inputs each further comprising at least one of: a ground truth indication of cleave profile, determined from an end face image, taken using a depth camera, of a same test fibre as the training input is associated with; and a ground truth indication of cleave angle determined from a side view image of a same test fibre as the training input is associated with; the model comprises at least one rule mapping the at least one feature to an indication of a cleave profile and the model is created at least by defining the at leastDocket No. 13768-4558a_502046-PCT01 one rule to map at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, to an indication of at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre; and the model comprises a lookup table associating the at least one feature with an indication of a cleave profile and the lookup table is created by associating at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a charactenstic of an end face of the test fibre, with an indication of at least one of: a cleave angle determined from a side view image of the test fibre; and a cleave profile determined from an end face image, taken by a depth camera, of the test fibre.

15. The method of any of claims 6 to 14, wherein at least one of: the indication of cleave quality comprises an indication of an end face quality, the indication of end face quality determined by analysing pixel data of the image and using the model to map at least one of: a level of damage to the fibre, a level of damage to a cleaved end face of the fibre, a level of debns present in a core of the fibre, an indication of step height of cladding of the fibre across a tear in cladding of the fibre, a level of debris present on the end face of the fibre, a shape of an outer surface of the fibre, and a level of liquid present on the end face of the fibre, to the indication of end face cleave quality; the indication of cleave quality comprises an indication of cleave profile, and wherein using the model to map the at least one feature to the indication of cleave quality comprises using the model to map at least one of: a level of damage to the fibre, a level of damage to a cleaved end face of the fibre, a level of debris present in a core of the fibre, an indication of step height of cladding of the fibre across a tear in cladding of the fibre, a level of debris present on the end face of the fibre, a shape of an outer surface of the fibre, and a level of liquid present on the end face of the fibre, to the indication of cleave profile and therefore the indication of cleave quality: and the method further comprises analysing pixel data of the image to determine an indication of end face quality of the fibre prior to the analysing pixel data of the image to determine cleave quality of the fibre, the end face quality comprising at leastDocket No. 13768-4558a_502046-PCT01 one of: a level of damage to the fibre, a level of damage to a cleaved end face of the fibre, an indication of step height of cladding of the fibre across a tear in cladding of the fibre, a level of debris present in a core of the fibre, a level of debris present on the end face of the fibre, a shape of an outer surface of the fibre, and a level of liquid present on the end face of the fibre, wherein the analysing pixel data of the image to determine cleave quality of the fibre is performed in response to the indication of end face quality indicating an end face quality above a threshold quality'.

16. The method of claim 15, wherein the method comprises analysing pixel data of the image to determine an indication of end face quality of the fibre prior to the analysing pixel data of the image to determine cleave quality of the fibre, and wherein the method further comprises: in response to the indication of end face quality indicating a quality below a threshold quality, performing one of: initiating discarding of the fibre, and initiating re-cleaving of the fibre.

17. A computer readable medium storing instructions that, when executed by a computer, perform the method of any of claims 6 to 16.

18. A method (1300) for creating a model for mapping at least one feature of an image of an end face of a cleaved hollow core optical fibre to an indication of cleave quality of the fibre, the indication of cleave quality’ comprising an indication of at least one of: cleave angle and cleave profile, the method comprising at least one of:1) training a machine-learning model (1308), the training comprising using training inputs each comprising at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, and the training inputs each further comprising at least one of: a ground truth indication of cleave angle, determined from a side view image of a same test fibre as the training input is associated wi th; and a ground truth indication of cleave profile, determined from an end face image taken using a depth camera, of a same test fibre as the training input is associated with;2) defining at least one rule (1310) mapping at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one featureDocket No. 13768-4558a_502046-PCT01 representing a characteristic of an end face of the test fibre, to an indication of at least one of: a cleave angle determined from a side view image of the test fibre, and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre; and3) defining a lookup table (1312) by associating at least one feature of an end face image of a cleaved hollow core optical test fibre, each of the at least one feature representing a characteristic of an end face of the test fibre, with an indication of at least one of: a cleave angle determined from a side view image of the test fibre, and a cleave profile determined from an end face image, taken using a depth camera, of the test fibre.

19. The method of claim 18, wherein the indication of cleave angle determined from a side view image of a test fibre is determined to be an average of cleave angles determined from a plurality of side view images of the test fibre, the plurality of side view images being of the test fibre with at least two different, relative to a camera that produced the side view images, rotation angles about an axis central to a core of the test fibre.

20. The method of any of claims 18 to 19, further comprising: initiating using the model to map at least one feature of an image of an end face of a cleaved hollow core optical fibre, the feature representing a characteristic of the end face, to an indication of cleave quality and outputting the indication of cleave quality.

Images