JPS6218883B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a method for detecting a core axis deviation of a fiber from an observation signal obtained when the fiber is observed at the abutting portion of fibers to be connected or at the connecting portion of already connected fibers.

In order to connect fibers with low loss, alignment is performed to minimize the misalignment between the fibers to be connected. Conventionally, as shown in Fig. 1, a light source 5 is attached to one of the fibers to be connected. A power monitor that receives light from the other fiber 2, monitors the light with a light receiver 6 coupled to the other fiber 2, and moves either or both of the fiber fixing bases 3 and 4 so that the amount of light is maximized. law is used. Such a power monitoring method is essentially limited to the case where the monitor light can be input into the fiber to be connected and the monitor light can be taken out from within the fiber for reception. However, if this is not possible, such as when connecting via a transmission line that includes a repeater, it is necessary to align the axes using another method.

Therefore, if the cores of the fibers 1 and 2 near the connection point can be visually observed using a microscope 11 or a video monitor 12, as shown in FIG. You can align yourself. However, if the core of the fiber is aligned while visually observing it, the operator's familiarity and habits, as well as the quality of the working environment, will be reflected in the alignment characteristics and working time, and it is not always possible to maintain excellent characteristics. do not have. Therefore, it is desirable to detect the axis misalignment of the core from an image of the fiber including the core, and feed that information back to the axis alignment mechanism to automatically eliminate the axis misalignment. Furthermore, in order to evaluate splicing loss from the shape of the core after splicing, a technique for quantitatively determining axial and angular misalignments in the vicinity of the splicing point is required.

In view of the above circumstances, the present invention provides a method for detecting fiber core axis misalignment that can be applied to the above-mentioned automatic core axis alignment and core axis shape inspection after connection. axis position,
Alternatively, the core axis misalignment can be detected.

The appearance of the core when observing a fiber differs slightly depending on the observation method and optical characteristics of the observation system, but in any case, as an image signal scanned in a direction perpendicular to the axis of a certain fiber, , Figure 3a
A light/dark (brightness) signal that characterizes the fiber outer diameter and core boundary as shown in Figure 1, or a color signal when identifying by color information is obtained. In this figure, the core boundaries are identified as dips in the luminance signals at 21 and 22. Therefore, by searching for such dark areas on the luminance signal, the boundary points of the core can be determined, and the position of the central axis of the core can be determined from the intermediate points. However, in the case of normally obtained images, as is clear from Figure a, the change in brightness at the core is extremely small compared to that at the outer periphery of the fiber.
Furthermore, a white noise component is superimposed on the image signal from the TV. Therefore, as shown in FIG. 3b, the actually obtained signal waveform near the core includes, in addition to the lowered brightness parts 21 and 22, which are the true core boundaries.
Some parts 23, 24 and the like have reduced brightness, and it is extremely difficult to specify them. Also,
If you try to find this core boundary using a uniformly programmed calculation routine,
There is a risk of finding incorrect points such as 23 and 24 instead of the true values 21 and 22, and the central axis 26 of the core finally found will also be different from the true central axis 25. The correct core position is often not found.

The present invention is capable of accurately extracting core position information even from signals with small contrast and superimposed noise, and detecting axis misalignment between connected fibers and connected fibers. It is a feature.

Hereinafter, the present invention will be explained in detail based on examples. FIGS. 4a to 4d are explanatory diagrams showing embodiments of the present invention. For example, as shown in FIG. 4a, a-a', b-
Image signals V _a , V b , . . . V _o are obtained by scanning in a direction perpendicular to the fiber axis at each point b' . _{. .} , n−n′.
Such a signal waveform is generated when the light source 4 is placed on the axis perpendicular to the axes of the fibers A and B, as shown in FIG.
1, and a microscope 11 consisting of an objective lens 42 and an eyepiece 43 is placed on the opposite side of the fiber.The fiber image obtained by placing the microscope 11 on the opposite side of the fiber is converted into an image signal by a video camera 45. can be easily obtained by scanning. Note that this signal is a time signal from a reference point such as an image edge.

In a normal observation signal, the signal components that give core position information in each waveform include the outer diameter of the fiber, etc.
It is small compared to signal components that provide other information, and also includes noise from the observation system. Therefore, in order to focus only on information near the core, the fiber center position
For example, V _a is determined, and only the signal at an appropriate interval Δx at both ends thereof is extracted. Δx is normally 1/5 to 1/7 of the outer diameter of the fiber.

Then, using this waveform as a reference signal, the cross-correlation function R(τ) near the fiber center of the image signals V _{b .} . . V _o observed at other positions is determined from the following equation.

R(τ)=1/2Δx∫ ^xc+ 〓 ^x _xc− 〓 _x V _a (t)·V _i (t−τ) dt...(1) where i=b...n.

The cross-correlation function thus obtained gives a maximum value at the position where the two waveforms are most similar, as shown in FIG. 4C. τ ₀ when giving this maximum value gives the amount of deviation of the core position of the waveform V _i from the core position of the waveform V _a .

In this way, when calculating from the cross-correlation function,
The position of the core is given as a deviation from a particular waveform, in this example _Va . The advantage of this method is that it is almost unaffected by noise components as already explained in FIG. 3b.

In other words, the noise components of waveforms V _a and V _i are respectively
When Na and Ni are taken, if we take the cross-correlation function using only Na and Ni, we get 1/2Δx∫ ^xc+ 〓 ^x _xc- 〓 _x Na(t)・Ni(t-
τ)dt, but if Na(t) and Ni(t) have completely independent waveforms in time, the above equation will always be 0. In general, the noise superimposed on the image signal is white noise, so it satisfies this property.

Therefore, in theory, the cross-correlation function between signals with superimposed noise and the cross-correlation function taken between ideal signals with no noise are exactly the same, and the core position where the presence or absence of noise is determined This makes it possible to determine the core position very accurately without causing any errors in the information.

The above is based on a certain sampling line (i=
This is an explanation of the principle of determining the core position for any of b...n, and the result of doing this for all points other than a is shown in Figure 4 d, and when a is taken as the standard, Displacement amount τi (i=b...
n) is found for all points.

The amount of axial deviation near the connection point can be determined by the amount of deviation between the extension lines at the connection center point, if the positional deviation of at least two points for each fiber A and fiber B, a total of four points, is clear. In addition, if the detected value of positional deviation at each point contains an error, increase the number of measurement points and apply the least squares method to each fiber A and B for the amount of positional deviation τi. By determining the core axis by

Furthermore, since it is possible to determine not only axis misalignment but also axis inclination and bending from this information, it can be applied not only to axis alignment before connection, but also to estimation of connection loss after connection.

Furthermore, if the fiber surface is scratched or has air bubbles, a waveform that is considerably different from the normal waveform will be obtained, and in such cases the correlation coefficient will be significantly lower than when it is normal. Therefore, by setting an appropriate level for the correlation coefficient, it is possible to determine whether the fiber is abnormal at the sampling position depending on whether the correlation coefficient is satisfied or not.

Normally, in the observed waveforms of fibers, the appearance of the core often differs due to differences in fibers, focal shifts, and optical systems. However, since this method uses one of the actually observed waveforms as a reference waveform, it also has the advantage of being able to determine axis deviation with extremely high accuracy.

FIGS. 6a to 6c are explanatory diagrams showing other embodiments of the present invention. As shown in FIG. 6a, the image signals V _a , V _{b .} . . V _o at each point of the fiber are determined in the same way as in FIG. The difference is that ₀ is used. The reference waveform at this time is preferably one that is as close as possible to the observed waveform near the core, as shown in Figure 6b.
The position corresponding to the center of the core is set to 0, and the range ±Δ is such that information about the outer diameter of the fiber is not included.
It shall be defined by x. The image signal V obtained from each point of the fiber using such a waveform as a reference waveform
Similar to equation (1), cross-correlation is calculated for _a , V _b . . . V _o using the following equation.

R(τ)=1/2Δx∫〓 ^x ₋ 〓 _x V ₀ (t)V _i (t−
τ) dt... (2) However, i=a, b...n.

The obtained cross-correlation function R(τ) is shown in Figure 4C.
Similarly, it has a functional form with the maximum value. τ ₀ giving this maximum value is the point giving the center of the core in each image signal waveform.

Therefore, similarly to Fig. 4, by finding the center positions of these cores at each point of fibers A and B as shown in Fig. 6C, it is possible to detect core axis deviations, angular deviations, etc. It is possible. Also,
Similarly to the case of FIG. 4, it is also possible to determine whether there is an abnormality in the fiber from the value of the correlation coefficient.

The feature of this method is that the amount determined is an absolute value corresponding to the center position of the core, and is not a relative value of the amount of deviation from a certain reference waveform as in the method shown in FIG. Therefore, there is an advantage that the deviation between the core center and the fiber center, which is separately determined from the outer diameter measurement, or the eccentricity of the core, can also be determined.

Furthermore, the way the core is seen differs depending on the type of observation system. In other words, instead of observing the transmitted light with a normal microscope as shown in FIG. 5, which has been described in the Examples, for example, in a method of observing with an interference microscope, the observed waveform is at the core boundary as shown in FIG. When the fiber is irradiated with ultraviolet light to cause the core to emit fluorescence and observed under a microscope, the waveform is as shown in Figure 8, where the brightness of the core increases. A waveform with a unique shape can be obtained. However, when using the correlation method of the present invention, a common waveform is used as a reference waveform, or an optimal reference waveform is selected for each observation system and stored in a ROM for calculation as unique to that equipment. In addition, we categorize the appearance of fibers into several types.
By preparing several reference waveforms for these and storing them in ROM, and having the processing system determine the waveform with the highest correlation coefficient, the accuracy can be maintained. It becomes possible to determine the core center position.

As explained above, according to the present invention, for signal components near the center of an image signal obtained by observing a fiber, the waveforms at each point are correlated using one of the observed waveforms or a waveform prepared in advance as a reference waveform. Since the core position deviation or the core center position is determined by taking , it is almost unaffected by noise, and the observed waveform is slightly different because the focal point of the observation system is slightly shifted. Even in cases where these cores are misaligned, it is possible to accurately determine the center position of the core, and from this information, it is possible to determine the misalignment between the fibers to be connected, the misalignment of the axes near the connection point after connection, angular misalignment, bending, etc. can be evaluated. All of these processes can be performed by a computing device such as a computer, so if this technology is applied to fiber splicing equipment, it will be possible to achieve high-accuracy axis alignment in the pre-connection stage, and high-accuracy splice loss control in the post-connection stage. This has the advantage that all estimations can be made automatically.

[Brief explanation of the drawing]

Figure 1 is an explanatory diagram of the conventional fiber shaft alignment method;
Figure 2 is an explanatory diagram of a method for axial alignment by observing the fiber, Figures 3a and b are waveform diagrams showing examples of observed waveforms when optically observing the fiber, Figure 4a
- d are explanatory diagrams showing one embodiment of the present invention, Fig. 5 is a layout diagram of an observation device for observing a transmitted image of a fiber in the present invention, and Figs. 6 a - c are other embodiments of the present invention. FIGS. 7 and 8 are diagrams showing observed waveforms when the fiber is observed optically, and FIG. 7 is a waveform of the observed waveform obtained by observation using an interference microscope. 8 are waveform diagrams of observed waveforms obtained when the core emits fluorescence. 1, 2, A, B...Fiber, 11...Microscope, 12...Video monitor, 41...Light source, 42
...Objective lens, 43...Eyepiece lens, 45...
Video camera.

Claims (1)

[Claims] 1. Observing the fibers from a direction perpendicular to the axis of the fibers near the abutting portion of two fibers to be connected that are butted against each other in a straight line, or near the connection point of already connected fibers. In the method of detecting the core position, one of the signal components near the fiber center of the observed brightness signals of the fiber taken in the direction orthogonal to the axis at multiple points on the fiber is used as a reference signal, and the other signals are Calculate the cross-correlation function between the fibers and detect the core deviation at each point from the deviation between the signals that maximizes the function value, and detect the axis deviation between the connected fibers or between the already connected fibers. Characteristic method for detecting fiber core axis misalignment. 2. A method of detecting the core position by observing the fiber from a direction perpendicular to the axis of the fiber near the abutting part of two fibers to be connected that are butted against each other in a straight line, or near the connection point of already connected fibers. In this process, the cross-correlation function between the signal component near the fiber center of the observed brightness signal of the fiber taken in the direction perpendicular to the axis at multiple points on the fiber and a predetermined specific reference signal is calculated, and the function value is calculated. 1. A method for detecting core axis deviation of fibers, characterized in that the maximum core center position is determined at each point and the axis deviation between the connected fibers or between the connected fibers is detected.