Optical fiber alignment method, optical fiber connection body manufacturing method, optical fiber alignment device, and optical fiber fusion splicer

Abstract

The purpose of the present invention is to provide an optical fiber alignment method, an optical fiber connector manufacturing method, an optical fiber alignment device, and an optical fiber fusion splicing machine with which it is possible to suitably perform alignment in the circumferential direction.　This alignment device (200) comprises: imaging units (105A), (105B) that perform one round's worth of capturing side images in the circumferential direction of a pair of optical fibers (10A), (10B) at a plurality of focus positions; a feature value calculation unit (112) that, for each focus position, performs one round's worth of calculating feature values obtained by digitizing features of the side images for the respective optical fibers (10A), (10B); an asymmetry calculation unit (113) that, for each focus position, calculates the asymmetry of a cross-correlation between the one round's worth of feature values of the respective optical fibers (10A), (10B); a focus position selection unit (114) that selects specific focus positions among focus positions having a prescribed asymmetry or higher; and a rotation alignment unit that performs alignment in the circumferential direction of the pair of optical fibers (10A) (10B) on the basis of the one round's worth of side images of the respective optical fibers (10A), (10B) at the selected focus positions.

Description

[Technical Field]

[0001] The present invention relates to an optical fiber aligning method, an optical fiber connection assembly manufacturing method, an optical fiber aligning device, and an optical fiber fusion splicer. [Background technology]

[0002] In some cases, a pair of optical fibers are spliced ​​together to increase the length for long-distance optical transmission, and such splicing is also performed in multicore fibers. Fusion splicing using a fusion splicer is one method for splicing optical fibers together. When fusion splicing multicore fibers together, it is necessary to splice the cores of the respective multicore fibers together. Therefore, at least one of a pair of multicore fibers, whose end faces face each other while their central axes are aligned, is rotated in the circumferential direction to align the multicore fiber in the rotational direction. As a method for aligning such a multicore fiber, for example, the method described in Patent Document 1 below is known. In the method for aligning a multicore fiber described in Patent Document 1, the multicore fiber is rotated 0.1 degrees around its axis, and an image viewed from the outer peripheral surface of the multicore fiber is obtained for each 0.1-degree rotation. Thereafter, the rotation angle of the multicore fiber is determined by machine learning based on the obtained image, and the multicore fiber is aligned, or a correlation coefficient is determined and the multicore fiber is aligned at the rotation angle at which the correlation coefficient is maximized.

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2019-159017 Summary of the Invention

[0004] In a multi-core fiber, cores arranged at the outermost periphery may be arranged at equal intervals on a circle centered on the center of the cladding. However, the positions of the cores may be slightly misaligned. Even when the positions of the cores are slightly misaligned, when the multi-core fiber is imaged from the side as in the method described in Patent Document 1, approximately the same image is obtained for each of a pair of multi-core fibers at each predetermined angle. In this case, even when attempting to splice predetermined cores in each multi-core fiber, it is difficult to determine which image to select for alignment from multiple approximately identical images of each multi-core fiber. Therefore, it is difficult to determine the appropriate combination of opposing cores in each multi-core fiber for fusion splicing. For this reason, there is a demand for appropriate circumferential alignment. Furthermore, even in a single-core fiber, there is a demand for appropriate circumferential alignment when the core is eccentric from the center of the cladding.

[0005] Therefore, an object of the present invention is to provide an optical fiber alignment method, an optical fiber connection body manufacturing method, an optical fiber alignment device, and an optical fiber fusion splicer that can properly perform circumferential alignment.

[0006] In order to achieve the above object, aspect 1 of the present invention is an optical fiber alignment method comprising: an imaging step of capturing side images of a pair of optical fibers in a circumferential direction for one revolution at a plurality of focus positions; a feature calculation step of calculating, for each focus position, feature amounts that quantify the characteristics of the side images for one revolution of each of the optical fibers; an asymmetry calculation step of calculating, for each focus position, the asymmetry between the feature amounts for one revolution of each of the optical fibers; a focus position selection step of selecting a specific focus position from the focus positions having a predetermined asymmetry amount greater than the smallest asymmetry amount; and a rotational alignment step of circumferentially aligning the pair of optical fibers based on the side images for one revolution of each of the optical fibers at the selected focus positions.

[0007] In such an alignment method, the specific focus position selected is a position where the asymmetry of the feature quantities for one revolution based on the side image of one revolution of each of the pair of optical fibers is equal to or greater than a predetermined asymmetry, which is greater than the smallest asymmetry. Therefore, the side image captured at the selected focus position clearly shows the structural differences between the optical fibers than a side image captured at a focus position smaller than the predetermined asymmetry. In this way, a focus position where the structural differences between the optical fibers are clear is selected, and alignment is performed using a side image where the structural differences are clear, so that the optical fiber alignment method of the present invention can properly perform circumferential alignment.

[0008] A second aspect of the present invention is the method for aligning an optical fiber according to the first aspect, wherein the asymmetry calculation step includes a cross-correlation calculation step and a difference calculation step, and when the feature quantities for one revolution consist of n-times repetition patterns that are similar to each other two or more times, the cross-correlation calculation step changes the relative angle of each of the optical fibers in the circumferential direction for each of the focus positions, and calculates the cross-correlation between the feature quantities for one revolution of each of the optical fibers at each relative angle, and the difference calculation step calculates, for each of the focus positions, differences between multiple peak values ​​among the n-th largest peaks of the cross-correlation, and determines the asymmetry based on the differences.

[0009] When the feature quantities for one revolution consist of n-times or more repeating patterns, each optical fiber has a refractive index distribution similar to each other with n-times rotational symmetry along the circumferential direction around the central axis of the cladding. Examples of such optical fibers include a multicore fiber in which multiple cores are arranged with n-times rotational symmetry along the circumference around the central axis of the cladding, and an optical fiber having stress-applying portions arranged to sandwich a core arranged at the center of the cladding. When aligning such optical fibers, if the relative angle of the optical fibers is changed and the cross-correlation between the feature quantities for one revolution of the side images of a pair of optical fibers is calculated, the same number of large peaks as the multiple repeating patterns are calculated. In other words, n large peaks are calculated. These large peaks are due to the influence of the refractive index distribution forming each pattern. Therefore, the deviation between the first n-th large peaks, which is the same number as the multiple repeating patterns, in the cross-correlation indicates a deviation in the refractive index distribution forming each repeating pattern. Therefore, the asymmetry can be easily determined by calculating the difference between multiple peak values ​​among the n peaks due to the influence of the refractive index distribution and calculating the asymmetry based on this difference. Note that when calculating the asymmetry, the calculated difference may be used as the asymmetry itself.

[0010] A third aspect of the present invention is the optical fiber aligning method according to the second aspect, wherein in the difference calculation step, the difference is calculated based on a standard deviation or variance of the plurality of peak values.

[0011] A fourth aspect of the present invention is the optical fiber aligning method according to the second aspect, characterized in that in the difference calculation step, the difference is calculated as a ratio or difference between two peak values ​​among the nth largest peak values.

[0012] A fifth aspect of the present invention is the optical fiber aligning method according to any one of the first to fourth aspects, characterized in that in the focus position selection step, the focus position having the asymmetry of 1+1.96σ or more is selected, where σ is the standard deviation of all the asymmetry degrees.

[0013] By selecting such a focus position, it is possible to perform proper alignment with a statistical probability of 95% or more.

[0014] A sixth aspect of the present invention is the optical fiber aligning method according to any one of the first to fourth aspects, wherein the focus position selecting step selects the focus position at which the degree of asymmetry is greatest.

[0015] By selecting such a focus position, it is possible to perform proper alignment with the highest probability.

[0016] A seventh aspect of the present invention is a method for manufacturing an optical fiber connection body, characterized in that it comprises a fusion splicing step of aligning a pair of optical fibers by the optical fiber aligning method of any one of aspects 1 to 6, and then fusion splicing the pair of optical fibers.

[0017] According to this method for manufacturing an optical fiber connection assembly, an optical fiber connection assembly can be obtained in which a pair of optical fibers are properly aligned in the circumferential direction.

[0018] In addition, in order to solve the above-mentioned problems, aspect 8 of the present invention is an optical fiber alignment device characterized by comprising: an imaging unit that images side images of a pair of optical fibers in the circumferential direction for one revolution at multiple focus positions; a feature calculation unit that calculates, for each focus position, feature amounts that quantify the features of the side images for one revolution of each of the optical fibers; an asymmetry calculation unit that calculates, for each focus position, the asymmetry between the feature amounts for one revolution of each of the optical fibers; a focus position selection unit that selects a specific focus position from the focus positions that have a predetermined asymmetry amount greater than the smallest asymmetry amount; and a rotary alignment unit that circumferentially aligns the pair of optical fibers based on the side images for one revolution of each of the optical fibers at the selected focus position.

[0019] According to this optical fiber alignment device, similarly to the first embodiment, it is possible to appropriately perform alignment in the circumferential direction.

[0020] A ninth aspect of the present invention is the optical fiber aligning device of the eighth aspect, wherein the asymmetry calculation unit includes a cross-correlation calculation unit and a difference calculation unit, and when the feature quantities for one revolution consist of n-times repetition patterns that are similar to each other two or more times, the cross-correlation calculation unit changes the relative angle of each of the optical fibers in the circumferential direction for each of the focus positions and calculates the cross-correlation between the feature quantities for one revolution of each of the optical fibers at each relative angle, and the difference calculation unit calculates, for each of the focus positions, differences between multiple peak values ​​among the n-th largest peaks of the cross-correlation, and determines the asymmetry based on the differences.

[0021] According to such an optical fiber aligning device, the degree of asymmetry can be easily determined in the same manner as in the second embodiment.

[0022] A tenth aspect of the present invention is the optical fiber aligning device according to the ninth aspect, wherein the difference calculation unit calculates the difference based on the standard deviation or variance of the plurality of peak values.

[0023] An eleventh aspect of the present invention is the optical fiber aligning device of the ninth aspect, wherein the difference calculation unit calculates the difference as a ratio or difference between two peak values ​​among the n-th largest peak values.

[0024] A twelfth aspect of the present invention is the optical fiber alignment device of any one of aspects 8 to 11, characterized in that the focus position selection unit selects the focus position where the asymmetry is 1+1.96σ or more, where σ is the standard deviation of all the asymmetry.

[0025] In this case, similarly to the fifth embodiment, the alignment can be performed appropriately with a statistical probability of 95% or more.

[0026] A thirteenth aspect of the present invention is the optical fiber aligning device according to any one of the eighth to eleventh aspects, wherein the focus position selection unit selects the focus position at which the degree of asymmetry is maximum.

[0027] In this case, similarly to the sixth embodiment, proper alignment can be performed with the highest probability.

[0028] A fourteenth aspect of the present invention is an optical fiber fusion splicer comprising an optical fiber aligning device according to any one of aspects eight to thirteen, and a fusion splicing section for fusing a pair of the optical fibers aligned by the aligning device.

[0029] Such an optical fiber fusion splicer can provide an optical fiber splice in which a pair of optical fibers are properly aligned in the circumferential direction.

[0030] As described above, according to the present invention, there can be provided an optical fiber alignment method capable of properly performing circumferential alignment, a method for manufacturing an optical fiber connection body using the alignment method, an optical fiber alignment device capable of properly performing circumferential alignment, and an optical fiber fusion splicer using the alignment device. [Brief explanation of the drawings]

[0031] [Figure 1] 1 is a side view showing an outline of an optical fiber connector according to an embodiment of the present invention. [Figure 2] FIG. 2 is a cross-sectional view of the optical fiber shown in FIG. [Figure 3] 1 is a diagram conceptually illustrating an example of the configuration of a fusion splicer according to an embodiment of the present invention. [Figure 4] FIG. 10 is a diagram showing a profile of feature quantities for one revolution of a pair of optical fibers when the focus position of the imaging unit is 0.71. [Figure 5] FIG. 10 is a diagram showing a profile of feature quantities for one revolution of a pair of optical fibers when the focus position of the imaging unit is 0.56. [Figure 6]FIG. 10 is a diagram showing a profile of the relationship between the relative angle of a pair of optical fibers at a focus position of 0.71 of the imaging unit and the cross-correlation between the feature amounts of one revolution of the pair of optical fibers. [Figure 7] FIG. 10 is a diagram showing a profile of the relationship between the relative angle of a pair of optical fibers at a focus position of 0.56 of the imaging unit and the cross-correlation between the feature amounts of one revolution of the pair of optical fibers. [Figure 8] FIG. 10 is a diagram illustrating the relationship between the focus position and the degree of asymmetry. [Figure 9] 1 is a flowchart showing steps in a method for manufacturing an optical fiber connection assembly. [Figure 10] FIG. 10 is a diagram showing the relationship between the focus position and the ratio of peak values ​​obtained by combining two peaks. [Figure 11] FIG. 10 is a diagram showing the relationship between the focus position and the difference in peak value between two peak combinations. [Figure 12] FIG. 10 is a diagram showing the relationship between the focus position and the standard deviation of all peak values. DETAILED DESCRIPTION OF THE INVENTION

[0032] Hereinafter, embodiments for carrying out the optical fiber aligning method, optical fiber connection assembly manufacturing method, optical fiber aligning device, and optical fiber fusion splicer according to the present invention will be described with reference to the accompanying drawings. The embodiments exemplified below are intended to facilitate understanding of the present invention and are not intended to limit the present invention. The present invention can be modified and improved from the following embodiments without departing from the spirit of the present invention. Furthermore, in this specification, the dimensions of each component may be exaggerated to facilitate understanding.

[0033] FIG. 1 is a side view showing an outline of an optical fiber connection body according to an embodiment. In this embodiment, an example will be described in which the optical fiber is a multicore fiber. As shown in FIG. 1, the optical fiber connection body 1 includes an optical fiber 10A located on one side and an optical fiber 10B located on the other side, and includes a splice portion 1F in which one end of the optical fiber 10A and one end of the optical fiber 10B are fused to each other. The optical fibers 10A and 10B have substantially the same configuration. Therefore, the configurations of the optical fibers 10A and 10B will be described using a diagram of the optical fiber 10A.

[0034] Fig. 2 is a cross-sectional view of the optical fiber 10A shown in Fig. 1. As shown in Fig. 2, the optical fiber 10A of this embodiment includes a plurality of cores 11, a cladding 12, and a coating layer 13 that coats the cladding 12.

[0035] 1, in each of the optical fibers 10A and 10B, the coating layer 13 is peeled off over a certain distance from one end, which becomes the connection portion 1F, to expose the cladding 12. The coating layer 13 is made of, for example, an ultraviolet-curable resin.

[0036] In the optical fiber 10A of this embodiment, the cores 11 are arranged at approximately equal intervals on a circle centered on the central axis C of the cladding 12. In this embodiment, four cores 11 are arranged at equal intervals. Each core 11 is formed to have approximately the same diameter and approximately the same refractive index, and may propagate only light in the fundamental mode, or may propagate light in several higher-order modes in addition to light in the fundamental mode. The refractive index of each core 11 is higher than the refractive index of the cladding 12.

[0037] In the optical fiber connection body 1 of this embodiment, one end of the optical fibers 10A and 10B is fused to each other in a state where the central axes C of the claddings 12 coincide with each other and the relative positions in the rotational direction are aligned so that the cores 11 of the optical fibers 10A and 10B are optically coupled to each other. Therefore, as shown in Fig. 1, the core 11 of the optical fiber 10A and the core 11 of the optical fiber 10B are fused to each other separately.

[0038] Next, a fusion splicer for the optical fibers 10A and 10B that can manufacture such an optical fiber connection body 1 will be described.

[0039] FIG. 3 is a conceptual diagram illustrating the configuration of a fusion splicer 100 according to this embodiment. As shown in FIG. 3, the fusion splicer 100 primarily includes an aligning device 200 for aligning optical fibers 10A and 10B and a fusion splicing unit 101. The aligning device 200 primarily includes rotating units 102A and 102B, imaging units 105A and 105B, a processing unit 110, a memory 120, and an input unit 130. The processing unit 110 primarily includes an image processing unit 111, a feature calculation unit 112, an asymmetry calculation unit 113, a focus position selection unit 114, and a control unit 115. In this embodiment, the asymmetry calculation unit 113 includes a cross-correlation calculation unit 113A and a difference calculation unit 113B. FIG. 3 illustrates an example in which the various units in the processing unit 110 are connected via a bus line.

[0040] The rotating unit 102A holds the optical fiber 10A so that it can rotate about a central axis C, and the rotating unit 102B holds the optical fiber 10B so that it can rotate about the central axis C. The rotating units 102A and 102B are configured to be movable in a direction perpendicular to the direction of the central axis C, and can align the central axes C of the optical fibers 10A and 10B so that one end face of the optical fibers 10A and 10B face each other. The rotating units 102A and 102B are each rotated by, for example, a stepping motor or the like and can be stopped at a desired rotation angle. The rotating units 102A and 102B are electrically connected to the processing unit 110 and can be rotated at the above-mentioned rotation angle based on a signal from the control unit 115 of the processing unit 110.

[0041] Fusion splicing unit 101 fuses an end of optical fiber 10A held by rotating unit 102A to an end of optical fiber 10B held by rotating unit 102B. Fusion splicing unit 101 includes, for example, a pair of discharge electrodes facing each other across the ends of optical fibers 10A and 10B, and fusion splices optical fibers 10A and 10B by heating them with discharge from the discharge electrodes. Fusion splicing unit 101 is electrically connected to processing unit 110, and the timing and intensity of discharge are adjusted by signals from control unit 115 of processing unit 110.

[0042] The imaging unit 105A is disposed substantially opposite the side surface of one end of the optical fiber 10A and can capture a side image of the optical fiber 10A from a direction perpendicular to the longitudinal direction of the optical fiber 10A. The imaging unit 105B is disposed substantially opposite the side surface of one end of the optical fiber 10B and can capture a side image of the optical fiber 10B from a direction perpendicular to the longitudinal direction of the optical fiber 10B. As described above, the coating layer 13 is stripped from one end of each of the optical fibers 10A and 10B. Therefore, the imaging unit 105A can capture an image of the side surface of the cladding 12 of the optical fiber 10A and a portion of the cores 11 visible through the cladding 12. The imaging unit 105B can capture an image of the side surface of the cladding 12 of the optical fiber 10B and at least a portion of the cores 11 visible through the cladding 12. Each of the imaging units 105A and 105B is electrically connected to the processing unit 110. The imaging units 105A and 105B can capture images at any timing based on a signal from the control unit 115 of the processing unit 110. For example, an image can be captured every time the rotating units 102A and 102B rotate the optical fibers 10A and 10B by a desired rotation angle. This desired rotation angle is, for example, 0.1 degrees. The imaging units 105A and 105B input the captured images to the image processing unit 111 of the processing unit 110.

[0043] Furthermore, the imaging units 105A and 105B of this embodiment are composed of fixed-focus cameras whose focus positions are fixed at a predetermined distance from the imaging units 105A and 105B and are configured to be movable along the imaging direction of the imaging units 105A and 105B. Therefore, the imaging units 105A and 105B can image the optical fibers 10A and 10B at multiple focus positions by moving relative to the optical fibers 10A and 10B. These focus positions are focus positions in the radial direction of the optical fibers 10A and 10B along the imaging direction of the imaging units 105A and 105B. If the imaging units 105A and 105B have a focus adjustment function that can adjust the focus position, the imaging units 105A and 105B may use this function to image the optical fibers 10A and 10B at multiple focus positions. The focus position is preferably adjusted by the control unit 115, which will be described later. That is, if the image capturing units 105A and 105B are composed of fixed-focus cameras, the desired focus position is achieved by moving the image capturing units 105A and 105B in the radial direction of the optical fibers 10A and 10B using a moving means (not shown) in response to a control signal from the control unit 115. Also, if the image capturing units 105A and 105B have a focus adjustment function, the desired focus position is achieved by having the image capturing units 105A and 105B adjust their focus in response to a control signal from the control unit 115. The image capturing units 105A and 105B may be integrated to simultaneously capture images of one end of each of the pair of optical fibers 10A and 10B, and may be configured so that the focus position of the image capturing unit 105A with respect to the optical fiber 10A and the focus position of the image capturing unit 105B with respect to the optical fiber 10B are the same.

[0044] The processing unit 110 may be, for example, an integrated circuit such as a microcontroller, an integrated circuit (IC), a large-scale integrated circuit (LSI), or an application-specific integrated circuit (ASIC), or an NC (numerical control) device. Furthermore, when an NC device is used, the processing unit 110 may or may not use a machine learning device. The control unit 115 of the processing unit 110 controls the operations of the fusion splicer 101, the rotation units 102A and 102B, the imaging units 105A and 105B, the image processing unit 111, the feature calculation unit 112, the cross-correlation calculation unit 113A, the difference calculation unit 113B, the focus position selection unit 114, and the like.

[0045] The processing unit 110 is electrically connected to a memory 120. The memory 120 is, for example, a non-transitory recording medium, and is preferably a semiconductor recording medium such as a random access memory (RAM) or a read only memory (ROM), but may include any known type of recording medium, such as an optical recording medium or a magnetic recording medium. Note that the term "non-transitory" recording medium includes all computer-readable recording media except for transient, propagating signals, and does not exclude volatile recording media.

[0046] Image processing unit 111 processes the image signals input from imaging units 105A and 105B. At this time, for example, noise may be removed from the image, or signals indicating each pixel of the image may be binarized. The signals processed by image processing unit 111 are output from image processing unit 111 and input to feature amount calculation unit 112. Note that if image processing is not required, image processing unit 111 is not required, and in this case, the image signals output from imaging units 105A and 105B may be input directly to feature amount calculation unit 112.

[0047] The feature amount calculation unit 112 calculates, for each of the optical fibers 10A and 10B, feature amounts that digitize the features of the side images captured by the imaging units 105A and 105B. Therefore, when the imaging units 105A and 105B capture side images of the optical fibers 10A and 10B for one revolution, the feature amount calculation unit 112 calculates feature amounts for one revolution of the optical fibers 10A and 10B. For example, when the imaging units 105A and 105B capture images of the optical fibers 10A and 10B every time the rotating units 102A and 102B rotate the optical fibers 10A and 10B by a rotation angle of 0.1 degrees, the feature amount calculation unit 112 calculates 3600 feature amounts for each of the optical fibers 10A and 10B. Therefore, the feature amount for one revolution is composed of, for each of the optical fibers 10A and 10B, data that combines, for example, the rotation angle of the optical fiber and the feature amount at that rotation angle. The method for calculating the feature amounts of the side surface images is not particularly limited as long as it can digitize the features of the side surface images. For example, a method can be used to perform processes such as edge detection and region extraction using the brightness distribution of the side surface images, calculate local feature amounts and global feature amounts from geometric features such as the width, area, and metric tensor of each region, and analytical features such as the brightness gradient, Laplacian, and Fourier coefficient, and then appropriately combine these feature amounts to obtain the feature amounts. Machine learning may be used to calculate the feature amounts. In this embodiment, as described below, under the control of the control unit 115, the imaging units 105A and 105B capture side surface images of the optical fibers 10A and 10B for one revolution in the circumferential direction at multiple focus positions, and the feature amount calculation unit 112 calculates the feature amounts for one revolution of the optical fibers 10A and 10B for each focus position.

[0048] FIG. 4 is a diagram showing feature profiles for one revolution of the optical fibers 10A and 10B when the focus positions of the imaging units 105A and 105B are 0.71. The focus position of 0.71 means that the relative value obtained by dividing the coordinates of the focus position by the coordinates of a standard focus position is 0.71. Hereinafter, this profile may be referred to as a feature profile. In this embodiment, as described above, the optical fibers 10A and 10B have four cores 11 arranged at approximately equal intervals on a circumference centered on the central axis C of the cladding 12. Therefore, the optical fiber 10A has a refractive index distribution that is similar to each other in a four-fold rotational symmetric manner along the circumferential direction about the central axis C of the cladding 12, and the optical fiber 10B has a refractive index distribution that is similar to each other in a four-fold rotational symmetric manner along the circumferential direction about the central axis C of the cladding 12. For this reason, as shown in Figure 4, the feature profile of optical fiber 10A, indicated by a solid line, consists of four mutually similar repeating patterns, and the feature profile of optical fiber 10B, indicated by a dashed line, also consists of four mutually similar repeating patterns. Note that it is important that the feature profiles consist of mutually similar repeating patterns, and it is not necessary to define boundaries between these patterns. In Figure 4, one of the repeating patterns is indicated by Pt71.

[0049] FIG. 5 is a diagram showing feature amount profiles for one revolution of the optical fibers 10A and 10B at a focus position of 0.56 of the imaging units 105A and 105B. As shown in FIG. 5, the feature amount profile of the optical fiber 10A and the feature amount profile of the optical fiber 10B each consist of four repetition patterns that are similar to each other. In FIG. 5, one of the repetition patterns is indicated by Pt56. As is clear from FIGS. 4 and 5, when the focus positions of the imaging units 105A and 105B differ, the feature amounts change, and the feature amount profiles also change. Note that even if the feature amounts for one revolution are not visualized as in FIGS. 4 and 5, the feature amount calculation unit 112 can grasp this repetition pattern. A technique such as pattern recognition is used to grasp this, and machine learning may be used as this technique.

[0050] In this embodiment, an example in which the repeat pattern is four times is shown, but if the optical fibers 10A and 10B have refractive index distributions that are similar to each other in an n-fold rotational symmetry, that is, two or more times, along the circumferential direction centered on the central axis C of the cladding 12, the feature profile for one revolution will consist of n-fold repeat patterns. Signals indicative of the feature amounts for one revolution of each of the optical fibers 10A and 10B calculated in this manner by the feature calculation unit 112 are stored in the memory 120.

[0051] The cross-correlation calculation unit 113A changes the relative angle of the optical fibers 10A and 10B in the circumferential direction and calculates the cross-correlation between the feature quantities for one revolution of each optical fiber at each relative angle. The cross-correlation calculation unit 113A changes the relative angle of the feature quantities for one revolution of each optical fiber 10A and 10B on the data. Specifically, the cross-correlation calculation unit 113A calculates the cross-correlation between the feature quantities for one revolution of each optical fiber 10A and 10B at each relative angle, while shifting the relative angle of the feature quantities for one revolution of each optical fiber 10A and 10B by 0.1 degrees, for example. The cross-correlation is determined, for example, by a cross-correlation function. The closer the cross-correlation is to 1, the higher the cross-correlation between the feature quantities for one revolution of each optical fiber 10A and 10B. The closer the cross-correlation is to 0, the lower the cross-correlation between the feature quantities for one revolution of each optical fiber 10A and 10B. FIG. 6 is a diagram showing a profile of the relationship between the relative angle of the optical fibers 10A and 10B when the focus positions of the imaging units 105A and 105B are 0.71 and the cross-correlation between the feature quantities for one revolution of the optical fibers 10A and 10B. This relationship is made up of data combining the relative angle in the circumferential direction of the optical fibers 10A and 10B and the feature quantities at that relative angle. Note that in FIG. 6, the cross-correlation is normalized. Furthermore, the relative angle of the optical fibers 10A and 10B is also the relative angle in data between the feature quantities for one revolution of the optical fiber 10A and the feature quantities for one revolution of the optical fiber 10B. Hereinafter, this profile may be referred to as a cross-correlation profile.

[0052] As shown in FIG. 6, four large peaks Pk1 to Pk4 appear in this cross-correlation profile. This is for the following reason. FIGS. 4 and 5 show a state in which the cross-correlation between the characteristic quantities of one revolution of the optical fiber 10A and the characteristic quantities of one revolution of the optical fiber 10B is high. Therefore, when the relative angle between the optical fibers 10A and 10B is shifted, the cross-correlation decreases. However, as described above, since each of the characteristic quantities of one revolution of the optical fibers 10A and 10B consists of four similar repeating patterns, when the relative angle between the characteristic quantities of one revolution of the optical fibers 10A and 10B changes by one revolution, the high cross-correlation state appears four times, as described above. Therefore, when the optical fibers 10A and 10B have similar refractive index profiles with n-fold rotational symmetry (two or more times) along the circumferential direction around the central axis C of the cladding 12, when the relative angle between the optical fibers 10A and 10B changes by one revolution, the high cross-correlation state appears n times. The cross-correlation calculation unit 113A calculates the cross-correlation between the feature amounts for one revolution of the optical fibers 10A and 10B for each focus position of the imaging units 105A and 105B.

[0053] FIG. 7 is a diagram showing the cross-correlation profile when the focus position of the imaging units 105A and 105B is 0.56. Note that the cross-correlation is also normalized in FIG. 7. When the focus position changes, the feature quantities of the optical fibers 10A and 10B change as shown in FIGS. 4 and 5, and therefore the cross-correlation also changes as shown in FIGS. 6 and 7. As shown in FIGS. 6 and 7, it can be seen that the change in cross-correlation is smaller for the optical fibers 10A and 10B imaged at a focus position of 0.56 than for the optical fibers 10A and 10B imaged at a focus position of 0.71. However, in this embodiment, the magnitude of the change in cross-correlation is not used for aligning the optical fibers 10A and 10B in the rotational direction. A signal indicating the relationship between the calculated relative angle and the cross-correlation of the profile is stored in the memory 120.

[0054] The difference calculation unit 113B calculates, for each focus position, differences between multiple peak values ​​among the n-th largest peaks in the cross-correlation, and calculates the asymmetry between the characteristic quantity for one revolution of the optical fiber 10A and the characteristic quantity for one revolution of the optical fiber 10B based on the differences. The asymmetry is a quantity indicating the degree of difference between two quantities. Here, it is a quantity indicating the degree of difference between the characteristic quantity for one revolution of the optical fiber 10A and the characteristic quantity for one revolution of the optical fiber 10B when the optical fiber 10A and the optical fiber 10B are at a predetermined relative angle. In this embodiment, the second largest peaks Pk1 and Pk2 are used, and the difference between the two peaks Pk1 and Pk2 is calculated using the ratio (value of the first largest peak Pk1) / (value of the second largest peak Pk2), and the calculated difference is used as the asymmetry. In this case, the larger the calculated ratio, the more clearly the asymmetry of the optical fibers 10A and 10B is indicated. This is for the following reason. When the cross-correlation is high, that is, when peaks Pk1 to Pk4 appear, the feature quantity profile of the optical fiber 10A and the feature quantity profile of the optical fiber 10B generally match. Therefore, the above ratio is a value comparing a state in which the optical fibers 10A and 10B are opposed to each other at the most appropriate alignment angle with a state in which the optical fibers 10A and 10B are opposed to each other at the second most appropriate alignment angle. Therefore, the larger this value, the more clearly it indicates the subtle differences between the structures of the optical fibers 10A and 10B in each state.

[0055] Since the cross-correlation between the feature quantities for one revolution of the optical fibers 10A and 10B while changing the relative angle between the optical fibers 10A and 10B is calculated for each focus position, the asymmetry expressed by the ratio is also calculated for each focus position. FIG. 8 is a diagram showing the relationship between the focus position and the asymmetry. As shown in FIG. 8, it can be seen that the calculated asymmetry changes depending on the focus position. In other words, it can be seen that the degree of display of the subtle difference between the structure of the optical fiber 10A and the structure of the optical fiber 10B changes depending on the focus position. The calculated asymmetry is stored in memory 120.

[0056] The focus position selector 114 selects a specific focus position from among focus positions with a predetermined asymmetry degree or greater. For example, when the standard deviation of all asymmetries is σ, the focus position selector 114 may select one of the focus positions with an asymmetry degree of 1 + 1.96σ or greater. In this case, the alignment of the optical fibers 10A and 10B, which will be described later, can be achieved with a probability of approximately 95% or greater. In the example of FIG. 8, the focus position selector 114 most clearly shows the subtle difference between the structure of the optical fiber 10A and the structure of the optical fiber 10B when the focus position is 0.56. Therefore, in the example of FIG. 8, to achieve the highest probability of achieving the alignment of the optical fibers 10A and 10B, which will be described later, the focus position selector 114 selects the focus position of 0.56.

[0057] The input unit 130 includes an input device such as a touch panel, and is electrically connected to the processing unit 110. For example, when it is known that the optical fibers 10A and 10B have similar refractive index distributions with n-fold rotational symmetry along the circumferential direction centered on the central axis C of the cladding 12, n is input to the input unit 130. Note that the feature calculation unit 112 may obtain n from the feature amounts for one revolution of the optical fibers 10A and 10B.

[0058] Next, a method for manufacturing the optical fiber connection body 1 will be described.

[0059] Fig. 9 is a flowchart showing steps of a method for manufacturing the optical fiber connection body 1. As shown in Fig. 9, the method for manufacturing the optical fiber connection body 1 mainly includes a focus position adjusting step P1, an imaging step P2, a feature amount calculating step P3, an asymmetry calculating step P4, a determining step P5, a focus position selecting step P6, a rotational alignment step P7, and a fusion splicing step P8. The asymmetry calculating step P4 includes a cross-correlation calculating step P4A and a difference calculating step P4B.

[0060] In this embodiment, the explanation will be given assuming that in the starting state, optical fiber 10A is placed on rotating part 102A, optical fiber 10B is placed on rotating part 102B, and the end faces of optical fibers 10A and 10B are opposed to each other so that the central axes C of the optical fibers 10A and 10B coincide.

[0061] (Focus position adjustment process P1) This step is a step in which the imaging units 105A and 105B adjust the focus positions. In this step, the control unit 115 first sends a control signal to the imaging units 105A and 105B to adjust the focus positions. If the imaging units 105A and 105B are fixed-focus cameras as described above, the control signal drives a moving means (not shown) to move the imaging units 105A and 105B along the imaging direction of the imaging units 105A and 105B and stop them at desired positions. In this way, the distance between the imaging unit 105A and the optical fiber 10A and the distance between the imaging unit 105B and the optical fiber 10B are each adjusted, and the focus positions of the optical fibers 10A and 10B in the radial direction along the imaging direction of the imaging units 105A and 105B are adjusted. Furthermore, if the imaging units 105A and 105B have a focus adjustment function, the imaging units 105A and 105B adjust the focus position in the direction perpendicular to the longitudinal direction of the optical fibers 10A and 10B in accordance with the control signal and stop at the desired position, thereby adjusting the focus position in the radial direction of the optical fibers 10A and 10B.

[0062] (Imaging process P2) This step is a step of capturing side images of the pair of optical fibers 10A and 10B for one revolution in the circumferential direction. In this step, the control unit 115 sends a control signal to the rotating units 102A and 102B to rotate the optical fibers 10A and 10B around the central axis C by a predetermined rotation angle. As described above, the predetermined rotation angle is, for example, 0.1 degrees. Furthermore, the control unit 115 sends an imaging control signal to the imaging units 105A and 105B every time the optical fibers 10A and 10B are rotated by the predetermined rotation angle, and the imaging units 105A and 105B capture side images of the optical fibers 10A and 10B. In this way, the control unit 115 causes the imaging units 105A and 105B to capture images of the optical fibers 10A and 10B for one revolution in the circumferential direction of the optical fibers 10A and 10B. Therefore, if the predetermined rotation angle is 0.1 degrees as described above, imaging units 105A and 105B each capture 3,600 side images. The captured images are input to image processing unit 111, and control unit 115 controls image processing unit 111 to cause image processing unit 111 to perform predetermined image processing. Image processing unit 111 outputs image data that has been subjected to image processing, and control unit 115 stores the image data in memory 120.

[0063] (Feature calculation process P3) This step is a step of calculating feature quantities that quantify the features of each captured side image. In this step, the feature quantity calculation unit 112 first reads out from the memory 120 the data of each side image of the optical fibers 10A and 10B stored in the imaging step P2. From this side image data, feature quantities of each side image are calculated. In the imaging step P2, side images for one revolution of the optical fibers 10A and 10B are captured. Therefore, by calculating feature quantities for each side image data, feature quantities for one revolution of the optical fibers 10A and 10B are calculated in this step. When these calculated results are plotted for each rotation angle, the feature quantity profiles shown in FIGS. 4 and 5 are obtained. The feature quantity calculation unit 112 outputs data indicating the feature quantities for one revolution of the optical fibers 10A and 10B, and the control unit 115 stores the data in the memory 120.

[0064] (Cross-correlation calculation process P4A) This step is a step of changing the relative angle in the circumferential direction of the optical fibers 10A and 10B and calculating the cross-correlation between the feature quantities for one revolution of the optical fibers 10A and 10B at each relative angle. In this step, the cross-correlation calculation unit 113A first reads data indicating the feature quantities for one revolution of the optical fibers 10A and 10B stored in the memory 120. Next, the cross-correlation calculation unit 113A calculates the cross-correlation between the feature quantities for one revolution of the optical fibers 10A and 10B when the optical fibers 10A and 10B are at a specific relative angle. Next, the cross-correlation calculation unit 113A changes the relative angle of the feature quantities for one revolution of the optical fibers 10A and 10B by a predetermined angle and calculates the cross-correlation between the feature quantities for one revolution of the optical fibers 10A and 10B at the changed relative angle. The relative angle changed at this time is preferably the same as the rotation angle of the optical fibers 10A, 10B that the imaging units 105A, 105B rotate each time they capture one side image in the imaging step P2, from the viewpoint of being able to use all of the captured image data. In this case, if the rotation angle is 0.1 degrees, a cross-correlation of 3600 is calculated. When this cross-correlation is plotted for each relative angle, the cross-correlation profiles shown in Figures 6 and 7 are obtained. The cross-correlation calculation unit 113A outputs data indicating the cross-correlation between the feature amounts for one revolution of the optical fibers 10A, 10B at each relative angle of the optical fibers 10A, 10B thus calculated, and the control unit 115 stores the data in the memory 120.

[0065] (Difference calculation process P4B) This step calculates the difference between the peak values ​​of the nth largest peak in the cross-correlation and determines the asymmetry based on the difference. As described above, n is the number of times that the optical fibers 10A and 10B repeat similar refractive index distributions in a rotationally symmetric manner along the circumferential direction around the central axis C of the cladding 12. In this embodiment, as shown in FIG. 2, n is 4, and as described above, four large peaks Pk1 to Pk4 appear in the feature profile. In this step, the difference calculation unit 113B reads data stored in the memory 120 that indicates the cross-correlation of the feature values ​​of one rotation of the optical fibers 10A and 10B at each relative angle between the optical fibers 10A and 10B. Next, the difference calculation unit 113B calculates the difference between the peak values ​​of the nth largest peak in the cross-correlation and calculates the asymmetry based on the difference. As described above, in this embodiment, the difference calculation unit 113B calculates the asymmetry degree by dividing the value of the first largest peak Pk1 by the value of the second largest peak Pk2, which is the ratio of the values ​​of the second largest peaks Pk1 and Pk2, and outputs data indicating the calculated asymmetry degree. The control unit 115 stores the data in the memory 120.

[0066] (Judgment process P5) This step is a step of determining whether the asymmetry calculation step P4 has been performed at a predetermined plurality of focus positions. In this step, if the asymmetry calculation step P4 has been completed at a predetermined plurality of focus positions, the control unit 115 proceeds to the focus position selection step P6. If the asymmetry calculation step P4 has not been completed at a predetermined plurality of focus positions, the control unit 115 returns to the focus position adjustment step P1. In the second or subsequent focus position adjustment step P1, the control unit 115 controls the image capture units 105A and 105B so that the focus positions of the image capture units 105A and 105B are different from the focus positions used in the previous image capture step P2. When proceeding to the focus position selection step P6, the image capture step P2 through the asymmetry calculation step P4 have been completed at a predetermined plurality of focus positions. The predetermined plurality of focus positions may be stored in the memory 120 in advance or may be input from the input unit 130 and stored in the memory 120.

[0067] In order to proceed to the focus position selection step P6, it is sufficient that the asymmetry calculation step P4 has been completed at a predetermined number of focus positions, and it is not necessary to perform the imaging step P2 through the difference calculation step P4B sequentially for each focus position. For example, part of the feature calculation step P3 may be performed while the imaging step P2 is being performed, or the feature calculation step P3 through the difference calculation step P4B may be performed at each focus position after the imaging step P2 has been completed at all focus positions.

[0068] (Focus position selection process P6) This step selects a specific focus position from among focus positions with a predetermined asymmetry degree or greater. In this step, the control unit 115 first reads the asymmetry degrees at multiple focus positions stored in the memory 120. In this embodiment, the read asymmetry degrees are arranged by focus position, as shown in FIG. 8. The predetermined asymmetry degree is greater than the smallest asymmetry degree. In this case, it is preferable for the focus position selection unit 114 to use the standard deviation σ of all asymmetries to select a focus position with an asymmetry degree of 1 + 1.96σ or greater, from the perspective of statistically achieving proper alignment with a probability of 95% or greater. In the case of FIG. 8, σ is 0.0027, so 1 + 1.96σ is 1.0053. In this case, the asymmetry degree is greater than 1 + 1.96σ only when the focus position is 0.56. Therefore, the focus position selection unit 114 selects the focus position of 0.56 and outputs data indicating the selected focus position. Although the selection of the focus position is not limited to this example, it is preferable to select the focus position with the greatest degree of asymmetry in order to maximize the probability of proper alignment. The control unit 115 stores the data in the memory 120. The focus position selection unit 114 may set different focus positions between the image capturing unit 105A and the image capturing unit 105B as specific focus positions among focus positions with a predetermined degree of asymmetry or greater.

[0069] (Rotation alignment process P7) This step is a step of circumferentially aligning the optical fibers 10A and 10B by rotating the optical fibers 10A and 10B relatively around the central axis C. In this step, the control unit 115 selects the relative angle between the optical fibers 10A and 10B at the peak Pk1 where the cross-correlation is greatest at the focus position selected in the focus position selection step P6. Next, the control unit 115 controls the rotation units 102A and 102B so that the optical fibers 10A and 10B are at the selected relative angle. In this way, the optical fibers 10A and 10B are aligned. This step is performed by the control unit 115 and the rotation units 102A and 102B. That is, in this embodiment, the control unit 115 and the rotation units 102A and 102B can be understood as a rotation alignment unit that circumferentially aligns the optical fibers 10A and 10B. Unlike the above description, in this step, the imaging units 105A and 105B may again capture side images of the optical fibers 10A and 10B at the focus position selected in the focus position selection step P6, and the optical fibers 10A and 10B may be rotationally aligned based on these side images.

[0070] (Fusion splicing process P8) This step is a step of fusion splicing the pair of optical fibers 10A, 10B after the pair of optical fibers 10A, 10B have been aligned by the above step. In this step, the control unit 115 sends a control signal to the fusion splicing unit 101 to cause the fusion splicing unit 101 to fusion splice one end of the optical fiber 10A and one end of the optical fiber 10B. As described above, if the fusion splicing unit 101 includes a pair of electrodes, the control unit 115 controls a power supply circuit (not shown) to discharge electricity from the pair of electrodes, and the fusion splice is performed by the heat generated by this discharge.

[0071] In this way, the optical fiber connection body 1 shown in FIG. 1 is manufactured.

[0072] Next, a modified example of the difference calculation step P4B will be described.

[0073] In the above embodiment, the difference in peak values ​​was calculated using the ratio between the value of the first largest peak Pk1 and the value of the second largest peak Pk2, and the asymmetry degree was calculated based on this difference. However, the peaks used to calculate the difference ratio are not limited to the first largest peak Pk1 and the second largest peak Pk2. FIG. 10 is a diagram showing the relationship between the focus position and the ratio of peak values ​​resulting from a combination of two peaks. As shown in FIG. 10, the ratio between the value of the first largest peak Pk1 and the value of the second largest peak Pk2 and the ratio of peak values ​​resulting from other combinations of peaks tend to change with the focus position in roughly the same way. In other words, when calculating the difference between two peak values ​​using the ratio between two of the n-th largest peak values ​​in the cross-correlation, the trend is roughly the same regardless of the combination of peak values. Therefore, the asymmetry degree can be appropriately calculated using the ratio of any combination of two peak values.

[0074] Fig. 11 is a diagram showing the relationship between the focus position and the difference in peak values ​​resulting from a combination of two peaks. As shown in Fig. 10 and Fig. 11, the ratio of peak values ​​resulting from a combination of two peaks shown in Fig. 10 and the difference in peak values ​​resulting from a combination of two peaks shown in Fig. 11 show roughly the same tendency in terms of change with respect to the focus position. Therefore, the asymmetry can be appropriately determined even if the difference between two peak values ​​is calculated using the difference between two of the n-th largest peak values ​​in the cross-correlation.

[0075] FIG. 12 is a diagram showing the relationship between the focus position and the standard deviation of all peak values. As shown in FIG. 12, even when the standard deviation of all peak values ​​is used, the change with respect to the focus position tends to be roughly the same as in FIGS. 10 and 11. It is believed that this tendency also applies when the standard deviation of multiple peak values ​​among the nth largest peak values ​​in the cross-correlation is used. It is also believed that the same tendency also applies when variance is used instead of standard deviation. Therefore, the asymmetry can be appropriately determined even when the difference is calculated using the standard deviation or variance of multiple peak values ​​among the nth largest peak values ​​in the cross-correlation.

[0076] The difference calculation step P4B may be performed by any other method as long as it calculates, for each focus position, the difference between the peak values ​​of the nth largest peak in the cross-correlation. By using any other method, the asymmetry degree can be calculated appropriately by calculating, for each focus position, the difference between the peak values ​​of the nth largest peak in the cross-correlation and calculating the asymmetry degree based on the difference.

[0077] As described above, the method for aligning the optical fibers 10A and 10B of this embodiment includes an imaging process P2 for capturing side images of the pair of optical fibers 10A and 10B in the circumferential direction for one revolution at multiple focus positions; a feature calculation process P3 for calculating, for each focus position, feature values ​​that quantify the features of the side images for one revolution of the optical fibers 10A and 10B; an asymmetry calculation process P4 for calculating, for each focus position, the asymmetry between the feature values ​​for one revolution of the optical fibers 10A and 10B; a focus position selection process P6 for selecting a specific focus position from focus positions with a predetermined asymmetry or greater; and a rotational alignment process P7 for circumferentially aligning the optical fibers 10A and 10B based on the side images for one revolution of the optical fibers 10A and 10B at the selected focus positions.

[0078] In addition, the alignment device 200 for the optical fibers 10A and 10B of this embodiment includes imaging units 105A and 105B that image side images of the pair of optical fibers 10A and 10B in the circumferential direction for one revolution at multiple focus positions, a feature calculation unit 112 that calculates feature amounts for one revolution of the optical fibers 10A and 10B by quantifying the features of the side images for each focus position, an asymmetry calculation unit 113 that calculates the asymmetry between the feature amounts for one revolution of the optical fibers 10A and 10B for each focus position, a focus position selection unit 114 that selects a specific focus position from focus positions with a predetermined asymmetry or greater, and a rotary alignment unit that circumferentially aligns the pair of optical fibers 10A and 10B based on the side images for one revolution of each of the optical fibers 10A and 10B at the selected focus position.

[0079] In such an alignment method and alignment device 200, the selected focus position is a position where the asymmetry of the feature quantities for one revolution based on the side image of one revolution of the pair of optical fibers 10A, 10B is equal to or greater than a predetermined asymmetry. Therefore, the side image captured at the selected focus position clearly shows the structural difference between the optical fibers 10A, 10B than a side image captured at a focus position with a smaller asymmetry. In this way, a focus position where the structural difference between the optical fibers 10A, 10B is clear is selected, and alignment is performed using a side image where the structural difference is clear. Therefore, the alignment method and alignment device 200 for the optical fibers 10A, 10B of this embodiment can appropriately perform circumferential alignment.

[0080] Furthermore, in the alignment method of this embodiment, the asymmetry calculation step P4 includes a cross-correlation calculation step P4A and a difference calculation step P4B. When the feature quantities for one revolution consist of n-times repeated patterns that are similar to each other, the cross-correlation calculation step P4A changes the relative angle in the circumferential direction of the optical fibers 10A and 10B for each focus position, and calculates the cross-correlation between the feature quantities for one revolution of the optical fibers 10A and 10B at each relative angle. The difference calculation step P4B calculates the difference between multiple peak values ​​among the n-th largest peaks Pk1 to Pkn of the cross-correlation for each focus position, and determines the asymmetry based on the difference. Furthermore, in the alignment device 200 of this embodiment, the asymmetry calculation unit 113 includes a cross-correlation calculation unit 113A and a difference calculation unit 113B, and when the feature quantities for one revolution consist of n-times repetition patterns that are similar to each other two or more times, the cross-correlation calculation unit 113A changes the relative angle in the circumferential direction of the optical fibers 10A and 10B for each focus position and calculates the cross-correlation between the feature quantities for one revolution of the optical fibers 10A and 10B at each relative angle, and the difference calculation unit 113B calculates the difference between multiple peak values ​​from the n-th largest peaks Pk1 to Pkn of the cross-correlation for each focus position and finds the asymmetry based on the difference.

[0081] When the feature quantities for one revolution consist of n-times or more repeating patterns, the optical fibers 10A and 10B each have a refractive index distribution similar to each other in an n-times rotationally symmetric manner along the circumferential direction around the central axis C of the cladding 12. When aligning such optical fibers 10A and 10B, the relative angle between the optical fibers 10A and 10B is changed, and the cross-correlation between the feature quantities for one revolution of the side images of the optical fibers 10A and 10B is calculated, the same number of large peaks as the number of repeating patterns are calculated. These large peaks are due to the influence of the refractive index distributions forming each pattern. Therefore, the deviation between the first n-th large peaks in the cross-correlation, which is the same number as the number of repeating patterns, indicates deviations in the refractive index distributions forming each repeating pattern. Therefore, the asymmetry can be easily calculated by calculating the difference between multiple peak values ​​among the n peaks due to the influence of the refractive index distribution and calculating the asymmetry based on this difference. In this embodiment, the calculated difference is used as the asymmetry. Note that the calculated difference may also be converted using a predetermined formula to calculate the asymmetry.

[0082] The manufacturing method for the optical fiber connection body 1 of this embodiment also includes a fusion splicing step P8 of fusion-splicing the optical fibers 10A, 10B after aligning the pair of optical fibers 10A, 10B by the above-described method for aligning the optical fibers 10A, 10B. The optical fiber fusion splicer 100 of this embodiment also includes the above-described aligning device 200 for the optical fibers 10A, 10B, and a fusion splicing section 101 that fuses the optical fibers 10A, 10B aligned by the aligning device 200. According to this manufacturing method for the optical fiber connection body 1 and fusion splicer, it is possible to obtain an optical fiber connection body 1 that is properly aligned in the circumferential direction.

[0083] Although the present invention has been described above using the above embodiment as an example, the present invention is not limited to the above embodiment.

[0084] For example, in the above embodiment, the optical fibers 10A and 10B have been described as having four cores 11. However, in the case of a multi-core fiber, the number of cores 11 is not limited to four. Furthermore, the optical fibers 10A and 10B may each include a trench layer having a refractive index lower than that of the cladding 12 so as to surround each core 11.

[0085] Furthermore, in the above-described embodiment, the optical fibers 10A and 10B are described as multi-core fibers. However, the optical fibers 10A and 10B in the above-described embodiment may have similar refractive index profiles with n-fold rotational symmetry along the circumferential direction around the central axis C of the cladding 12. Therefore, for example, the optical fibers 10A and 10B may be stress-applying optical fibers that have one core 11 arranged along the central axis C of the cladding 12 and further have a pair of stress-applying portions sandwiching the core 11. In this case, the feature quantities of one revolution of the optical fibers 10A and 10B are composed of two repeating patterns that are similar to each other.

[0086] In the above embodiment, the asymmetry calculation unit 113 includes the cross-correlation calculation unit 113A and the difference calculation unit 113B, and the asymmetry calculation step P4 includes the cross-correlation calculation step P4A and the difference calculation step P4B. However, as long as the asymmetry between the feature amounts for one revolution of the optical fibers 10A and 10B is calculated for each focus position, the asymmetry calculation unit 113 does not have to include the cross-correlation calculation unit 113A and the difference calculation unit 113B, and the asymmetry calculation step P4 does not have to include the cross-correlation calculation step P4A and the difference calculation step P4B. For example, if the optical fibers 10A and 10B are optical fibers that include a cladding 12 and one core 11 arranged along the central axis C of the cladding 12 and the core 11 is unevenly distributed, the optical fibers 10A and 10B do not have refractive index distributions that are similar to each other in two-fold or more rotational symmetry along the circumferential direction around the central axis C of the cladding 12. In this case, in the asymmetry calculation step P4, the asymmetry calculation unit 113 may calculate the asymmetry between the feature amounts for one revolution of the optical fibers 10A and 10B for each focus position, for example.

[0087] According to the present invention, there can be provided an optical fiber alignment method capable of properly performing circumferential alignment, a method for manufacturing an optical fiber connection body using the alignment method, an optical fiber alignment device capable of properly performing circumferential alignment, and an optical fiber fusion splicer using the alignment device, which can be used in fields such as optical communications, for example.

Claims

1. an imaging step of capturing side images of the pair of optical fibers in a circumferential direction for one revolution at a plurality of focus positions; a feature amount calculation step of calculating, for each focus position, a feature amount obtained by quantifying the feature of the side image for one revolution of each of the optical fibers; an asymmetry calculation step of calculating an asymmetry which is an amount indicating a degree of difference between the feature amounts for one revolution of each of the optical fibers for each of the focus positions; a focus position selection step of selecting a specific focus position from among the focus positions having a predetermined asymmetry degree greater than the smallest asymmetry degree; a rotational alignment step of aligning the pair of optical fibers in a circumferential direction based on the side image of one revolution of each of the optical fibers at the selected focus position; Equipped with the pair of optical fibers have refractive index profiles similar to each other in an n-fold rotational symmetry of at least two times along a circumferential direction around a central axis of the cladding, the asymmetry calculation step includes a cross-correlation calculation step and a difference calculation step, The feature quantity for one revolution is composed of n-times repeated patterns that are similar to each other and are repeated two or more times, In the cross-correlation calculation step, a relative angle of each of the optical fibers in the circumferential direction is changed for each of the focus positions, and a cross-correlation between the feature amounts for one revolution of each of the optical fibers is calculated at each relative angle; In the difference calculation step, a difference between a plurality of peak values ​​among the nth largest peaks in the cross-correlation is calculated for each focus position, and the asymmetry degree is calculated based on the difference.

1. A method for aligning an optical fiber, comprising:

2. In the difference calculation step, the difference is calculated based on the standard deviation or variance of the plurality of peak values.

2. The method for aligning an optical fiber according to claim 1.

3. In the difference calculation step, the difference is calculated as a ratio or difference between two peak values ​​among the nth largest peak values.

2. The method for aligning an optical fiber according to claim 1.

4. In the focus position selection step, when the standard deviation of all the asymmetry degrees is σ, the focus position having the asymmetry degree of 1+1.96σ or more is selected.

2. The method for aligning an optical fiber according to claim 1.

5. In the focus position selection step, the focus position with the maximum degree of asymmetry is selected.

2. The method for aligning an optical fiber according to claim 1.

6. 6. The optical fiber aligning method according to claim 1, further comprising a fusion splicing step of fusion splicing the pair of optical fibers after aligning the pair of optical fibers by the method according to claim 1.

1. A method for manufacturing an optical fiber connector, comprising:

7. an imaging unit that captures side images of the pair of optical fibers in a circumferential direction for one revolution at a plurality of focus positions; a feature amount calculation unit that calculates, for each focus position, a feature amount obtained by quantifying the feature of the side image for one revolution of each of the optical fibers; an asymmetry calculation unit that calculates an asymmetry that is an amount indicating a degree of difference between the feature amounts for one revolution of each of the optical fibers for each of the focus positions; a focus position selector that selects a specific focus position from among the focus positions having a predetermined asymmetry degree greater than the smallest asymmetry degree; a rotary alignment unit that aligns the pair of optical fibers in the circumferential direction based on the side image of one revolution of each of the optical fibers at the selected focus position; Equipped with the pair of optical fibers have refractive index profiles similar to each other in an n-fold rotational symmetry of at least two times along a circumferential direction around a central axis of the cladding, the asymmetry calculation unit includes a cross-correlation calculation unit and a difference calculation unit, The feature quantity for one revolution is composed of n-times repeated patterns that are similar to each other and are repeated two or more times, the cross-correlation calculation unit changes a relative angle of each of the optical fibers in the circumferential direction for each of the focus positions, and calculates a cross-correlation between the feature amounts for one revolution of each of the optical fibers at each relative angle; The difference calculation unit calculates, for each focus position, differences between a plurality of peak values ​​among the nth largest peaks in the cross-correlation, and obtains the asymmetry degree based on the differences. An optical fiber alignment device characterized by the above.

8. The difference calculation unit calculates the difference based on a standard deviation or a variance of the plurality of peak values.

8. The optical fiber alignment device according to claim 7.

9. The difference calculation unit calculates the difference as a ratio or difference between two peak values ​​among the nth largest peak values.

8. The optical fiber alignment device according to claim 7.

10. The focus position selection unit selects the focus position having the asymmetry of 1+1.96σ or more, where σ is the standard deviation of all the asymmetry degrees.

8. The optical fiber alignment device according to claim 7.

11. The focus position selection unit selects the focus position at which the degree of asymmetry is greatest.

8. The optical fiber alignment device according to claim 7.

12. An optical fiber aligning device according to any one of claims 7 to 11; a fusion splicer that fusion-splices the pair of optical fibers aligned by the alignment device; Equipped with 1. An optical fiber fusion splicer comprising:

Images