Method of calibrating interferometer and interferometer using the same

This invention is related to a method of calibrating an interferometer and an interferometer using the same.

Optical coherence tomography (SS-OCT) for obtaining a tomographic image of a biographic body with a wavelength-swept light source as a light source of an interferometer has been conventionally known (refer to, for example, Patent Literatures 1 and 2). In such tomography, the light from the wavelength-swept light source is divided into reference light that is guided to a reference optical path and measurement light that is guided to a measurement optical path, the reference light reflected on a reference mirror disposed in the reference optical path and the measurement light scattered and reflected on the biographic body as a measurement target disposed in the measurement optical path are again synthesized to be superimposed, so that a time-varying spectrum interference signal (beat signal) is generated. This spectrum interference signal (beat signal) is detected by a detector, the beat signal is sampled at equal time intervals on a time axis, and this spectrum interference signal (beat signal) is Fourier-transformed, so that the peak of each frequency on a frequency axis is detected.

However, an inverse proportional relationship is established between the wavelength and the position corresponding to the depth of the biographic body. The wave number does not therefore linearly vary on the time axis. When a point-spread-function of the interference signal is obtained on the frequency axis by the Fourier transform, the waveform of the point-spread-function is distorted in accordance with an increase in a depth of the biographic body. As a result, the tomographic image information is deteriorated.

To solve such deterioration, Patent Literature 1 discloses a wavelength-swept light source that linearly varies the wave number on the time axis. The wavelength-swept light source includes an oscillator having an electric optical deflector, an interferometer that outputs, relative to interference light by the difference between two optical paths by traveling oscillation output light from the oscillator in two optical paths each having a different optical path length, an electric signal representing interference intensity containing an AC component of a frequency proportional to the variation in the wave number of the oscillation output light of the optical path length, an error signal generation circuit that generates a feedback signal for maintaining a constant frequency of the AC component of the electric signal, and a feedback controller having a control voltage generation circuit that generates corrected control voltage to be supplied to the electric optical deflector by applying the feedback signal to the gain relative to the lamp voltage signal.

Patent Literature 2 discloses that a spectrum interference signal is Hilbert transformed, phase information is obtained by obtaining an arc tangent of the ratio between the spectrum interference signal and the Hilbert transformed spectrum interference signal, a function representing time dependency of a sweep wavelength is obtained after the phase information is unwrapped, and calibration relative to the time dependency of a sweep light source is executed by using the time dependency of the sweep wavelength.

Patent Literature

Patent Literature 1: JP2012-150081A

Patent Literature 2: JP2007-101365A

Technical Problem

However, in the technique disclosed in Patent Literature 1, the wave number is linearly varied on the time axis. More specifically, the feedback control for maintaining a constant frequency of the AC component is executed by detecting the frequency of the AC component of the electric signal obtained from the light detector of the interferometer such that a variation in the oscillation wavelength becomes constant relative to the time for maintaining the constant variation in the wave number relative to the time, so as to maintain the constant frequency of the alternating current. For this reason, it is necessary to constantly execute the feedback control during the measurement in this technique.

On the other hand, in the technique disclosed in Patent Literature 2, a function representing time dependency of a sweep wavelength is obtained, and calibration is performed to a wavelength-swept light source by using the time dependency of the sweep wavelength. Accordingly, when data is obtained by sampling a spectrum interference signal at unequal time intervals on the time axis, it is not necessary to constantly execute the feedback control during the measurement.

However, it is difficult to execute the sampling at unequal time intervals on the time axis.

The present invention has been made in view of the above circumferences. An object of the present invention is to provide a method of calibrating an interferometer capable of linearly varying a wave number on a time axis without constantly executing feedback control to a wavelength-swept light source during measurement, and an interferometer using the method.

Solution to Problem

A method of calibrating an interferometer according to the present invention includes a first step of generating a spectrum interference signal through sweep of a wavelength of light of a wavelength-swept light source by varying, relative to a time, voltage to be applied to the wavelength-swept light source within one period, a second step of obtaining a point-spread-function by sampling the spectrum interference signal at equal time intervals on a time axis, and Fourier transforming the sampled spectrum interference signal by multiplying by a first window function, a third step of obtaining a complex analysis signal including frequency information of the light of the wavelength-swept light source through Hilbert transform with inverse Fourier transform by multiplying the point-spread-function by a second window function, a fourth step of obtaining a variation in a frequency of the light relative to a time within one period at equal time intervals by unwrapping phase information of the complex analysis signal, a fifth step of obtaining a correspondence relationship between the variation in the frequency of the light within one period, which is sampled at the equal time intervals, and a variation in the voltage within one period, which is sampled at the equal time intervals, a sixth step of obtaining correspondence voltage that is obtained when the variation in the frequency of the light within one period is subdivided by using the correspondence relationship between the variation in the voltage and the variation in the frequency of the light obtained in the fifth step, and a seventh step of substituting the voltage to be applied to the wavelength-swept light source in the first step into the correspondence voltage obtained by the sixth step.

Advantageous Effects

According to the present invention, the frequency of the light can be varied at equal time intervals within one period by applying the correspondence voltage after the calibration, a wave number can be linearly varied on a time axis without constantly executing the feedback control to the wavelength-swept light source during the measurement.

FIG. 1 is a schematic diagram illustrating a configuration of a wavelength-swept light source for use in a method of calibrating an interferometer according to an embodiment of the present invention.

FIG. 2 is a schematic diagram illustrating a configuration of a linear chirped fiber Bragg grating illustrated in FIG. 1 .

FIG. 3 is a block diagram illustrating a configuration of an interference optical system illustrated in FIG. 1 .

FIG. 4 is a waveform chart illustrating one example of a waveform of a window function and one example of a waveform of a spectrum interference signal output from the interferometer illustrated in FIG. 1 .

FIG. 5 is a graph showing a relationship between a depth z and a variation in a light frequency.

FIG. 6 is a graph describing distortion of a waveform of a point-spread-function before calibration according to the embodiment of the present invention.

FIG. 7 is a graph showing a waveform of a second window function for executing a Hilbert transform process to the point-spread-function according to the present invention.

FIG. 8 is a graph showing a curve representing a relationship between a time and a variation in a light frequency obtained by executing an unwrapping process in view of a phase of a complex analysis signal obtained by the Hilbert transform process.

FIG. 9 is a graph showing a curve representing a variation in voltage before the calibration according to the present invention.

FIG. 10 is a graph showing a curve representing a correspondence relationship between voltage and a light frequency obtained by using the curve in FIG. 8 and the curve in FIG. 9 .

FIG. 11 is a graph showing a waveform of a point-spread-function obtained by sweeping a wavelength with correspondence voltage obtained by using the curve in FIG. 10 .

FIG. 12 is a waveform chart in which the normalized dot-spread-functions in FIG. 11 are superimposed.

FIG. 13 is a waveform chart in which the normalized dot-spread-functions are superimposed, and the dot-spread functions are obtained by performing a sampling process to the spectrum interference signal obtained by the calibration of the present invention.

Embodiment

FIG. 1 is a view illustrating a wavelength-swept light source (light source) for use in one embodiment of an interferometer using a method of calibrating an interferometer.

Referring to FIG. 1 , reference number 1 denotes a semiconductor optical amplifier (SOA), 2 denotes an optical isolator, 3 denotes a sweep modulator (EOM), 4 denotes a circulator, and 5 denotes a dispersion compensator.

DC voltage is applied to the sweep modulator 3 from a DC circuit 3 A, and an oscillation signal RF from an oscillator (RF SG) 3 B is input to the sweep modulator 3 through an amplifier (Amp) 3 C. Saw-tooth voltage V that periodically varies by a sweep frequency is applied to the oscillator 3 B from a frequency generator (FG) 3 D.

The oscillation signal RF is modulated by the voltage V. A modulation signal MO modulated by the voltage V is input to the sweep modulator 3 . The sweep modulator 3 is driven by the modulation signal MO, and operates, along with a variation in the frequency of the modulation signal MO, as a voltage controller that generates the after-described spectrum interference signal by sweeping the wavelength of the light of the wavelength-swept light source.

More specifically, at least the frequency generator 3 D operates as the voltage controller that generates the after-described spectrum interference signal by sweeping the wavelength of the light of the wavelength-swept light source through a variation in the voltage applied to the wavelength-swept light source relative to the time within one period.

The semiconductor optical amplifier 1 , the optical isolator 2 , the sweep modulator 3 , the circulator 4 , and the dispersion compensator 5 constitute a ring resonator 6 .

The semiconductor optical amplifier 1 includes a waveguide structure 1 a . The waveguide structure 1 a has on the one end face thereof an incident end face 1 b and the other end face thereof an emission end face 1 c.

Injection current I is injected into the waveguide structure 1 a , and carrier is generated in the waveguide structure 1 a . The carrier is consumed by a stimulated emission phenomenon of the light pulse incident on the incident end face 1 b of the waveguide structure 1 a . As a result, the pulse of the laser light P is amplified, and the pulse of the laser light P is emitted from the emission end face 1 c.

The pulse of the laser light P emitted from the emission end face 1 c allows the light to pass through only in one direction, and is guided to the sweep modulator 3 through the optical isolator 2 as an optical element that blocks return light.

An electro-optical modulator (EOM) is used for the sweep modulator 3 . In this case, the circulator 4 includes three ports. An emission light guide fiber 7 that guides the pulse of the laser light P output from the sweep modulator 3 is connected to a first port 4 a of the circulator 4 .

The dispersion compensator 5 is connected to a second port 4 b of the circulator 4 . A liner chirped fiber Bragg grating (LCFBG) schematically illustrated in FIG. 2 is used for the dispersion compensator 5 .

In the liner chirped fiber Bragg grating, the period of grating varies such that a reflection position of a low frequency component of the pulse linearly differs from that of a high frequency component. The liner chirped fiber Bragg grating includes a diffracting grating in the fiber.

The liner chirped fiber Bragg grating has characteristics of both normal dispersion and abnormal dispersion according to its direction, and can be used in a normal dispersion area and an abnormal dispersion area according to a connection type of the second port 4 b of the circulator 4 of the liner chirped fiber Bragg grating. In addition, in FIG. 1 , reference number 5 d denotes an incident end face and 5 e denotes a transmission end face.

In the normal dispersion area, a long-wavelength pulse component is first reflected, and a short-wavelength pulse component is reflected afterward. In the abnormal dispersion area, a short-wavelength pulse component is first reflected, and a long-wavelength pulse component is reflected afterward. In this embodiment, the liner chirped fiber Bragg grating is used in both of the areas.

A third port 4 c of the circulator 4 is connected to a feedback light guiding fiber 8 that feeds the pulse of the laser light reflected on the liner chirped fiber Bragg grating back to the semiconductor optical amplifier 1 .

The pulse of the laser light P output from the transmission end face 5 e of the liner chirped fiber Bragg grating is guided to an interference optical system (interferometer) 10 of an optical coherence tomography apparatus through an optical isolator 9 .

According to the wavelength-swept light source, an intensity modulation frequency is varied by varying the voltage V within one period, and the oscillation wavelength of the laser light P is varied.

A known interference optical system can be used. The interference optical system 10 divides the optical path of the pulse of the laser light P into an optical path for a reference light P 1 and an optical path for a measurement light P 2 by a coupler 10 a as an optical path divider as illustrated in FIG. 3 .

The reference light P 1 is guided to a collimator lens 10 c through a circulator 10 b , is guided to a fixed reference mirror 10 d disposed in the reference optical path as parallel light flux, and returns to the collimator lens 10 c again after being reflected on the fixed reference mirror 10 d.

The measurement light P 2 is guided to a collimator lens 10 c ′ through a circulator 10 b ′ to be converted into parallel light flux, and then is guided to a measurement target 11 disposed in the measurement optical path.

In this case, a total reflection mirror is disposed as the measurement target 11 for the descriptive purpose. However, a body tissue such as a fundus may be used as the measurement target 11 .

The measurement light P 2 is reflected on the measurement target 11 , and returns to the collimator lens 10 c ′ again.

The reference light P 1 is collected by a collimator lens 10 c , is guided to a coupler 10 a ′ as an optical path combining part through the circulator 10 b , and the measurement light P 2 is guided to the coupler 10 a ′ through a collimator lens 10 b′.

The after-described spectrum interference signal (interference signal) is generated by the interference of the reference light P 1 and the measurement light P 2 .

While the measurement target 11 is placed in the reference position 0 (for example, optical path difference with the fixed reference mirror 10 d is 1 mm), when the voltage V of the wavelength-swept light source is temporally varied within one period, and the spectrum interference signal S(t) is sampled at equal time intervals, the spectrum interference signal S(t) shown in FIG. 4 is obtained. This spectrum interference signal S(t) is input to a processor 12 .

This processor 12 includes a Fourier transform processor 12 a , a Hilbert transform processor 12 b , an inverse Fourier transform processor 12 c , an unwrapping processor 12 d , a substitution processor 12 e , and a tomographic image obtainer (tomographic image formation section) 12 f.

A known Fourier transform processor 12 a can be used. The Fourier transform processor 12 a obtains a point-spread-function (spread function) SPF(f) by multiplying the spectrum interference signal S(t) by a first window function W(t) illustrated in FIG. 4 . In addition, as ω=2πf, when the point-spread-function is obtained by the frequency f, the point-spread-function PSF(f) is obtained by the following equation where C is a coefficient.

PSF( f )= C∫S ( t ) W ( t )exp( i 2π ft ) dt

The window function W(t) is used for obtaining a smooth point-spread-function PSF(f).

The following relational expression is established among the frequency f, an optical path difference cΔt, and the depth z. In addition, c represents a light speed, and Δt represents a time difference between the light returned from the fixed reference mirror 10 d and the light returned from the measurement target 11 .

As illustrated in FIG. 5 , t represents a sweep time of one period T. A delay time Δt corresponding to the optical path difference is generated between the time required for the reference light P 1 returning from the fixed reference mirror 10 d and the time required for the measurement light P 2 returning from the measurement target 11 . The frequency of the measurement light is varied by f L +Δf relative to the frequency f L of the reference light. In this case, Δf is a beat frequency of the spectrum interference signal S(t).

The beat frequency Δf corresponds to the variable number f of the point-spread-function PSF(f), and the relational expression of f=ΔF·Δt/T is obtained. In this case, ΔF represents a sweep frequency width. When the terms of both sides of this expression are multiplied with a light speed c, cf=cΔF·Δt/T is obtained. Therefore, cf·T/ΔF=cΔt is obtained.

As the depth z is an optical path difference cΔt/2, the depth z can be obtained by obtaining the sweep frequency width ΔF and the variable number f.

When the point-spread frequency PSF(f) is obtained by Fourier transforming the spectrum interference signal S(t) in the Fourier transform processor 12 a , the waveform illustrated by reference number I 0 in FIG. 6 is obtained.

As illustrated in FIG. 3 , when the measurement target 11 is moved from the reference position 0 in the optical axis direction (depth direction Z), the waveforms I 0 , I 1 , I 2 , . . . , I m of the obtained point-spread-functions PSF(f) are distorted as shown in FIG. 6 in accordance with an increase in the optical path difference cΔt between the measurement target 11 and the fixed reference mirror 10 d.

In this embodiment, after a first step of generating the spectrum interference signal S(t) through the sweep of the wavelength of the light of the wavelength-swept light source by varying, relative to a time, the voltage V to be applied to the wavelength-swept light source within one period is executed, a second step of obtaining the point-spread-function PSF(f) by sampling the spectrum interference signal S(t) at the equal time intervals on the time axis, Fourier transforming the sampled spectrum interference signal S(t), and calculating the value obtained by the Fourier transform is executed in the Fourier transform processor 12 a.

Next, in the Hilbert transform processor 12 b , a third step of obtaining an analysis signal S H (t) (SH(t)) including the frequency information of the light of the wavelength-swept light source by determining a second window function R(f) illustrated in FIG. 7 from the point-spread-function PSF(f), and executing the Hilbert transform by inverse Fourier transforming the product of the window function R(f) and the point-spread-function PSF(f) in the inverse Fourier transform processor 12 c is executed.

In addition, in the Hilbert transform process, a process of doubling the positive part with the negative part as 0, and the window function R(f) is used such that the circumference of the peak of the point-spread-function PSF(f) for reducing the noise of the complex analysis signal S H (t) is only used for the calculation.

The complex analysis signal S H is obtained with the following relational expression with the window function R(f) and the PSF(f).

S H ( t )= C ∫PSF( f ) R ( f )exp( i 2π ft ) df

Next, a fourth step of obtaining a variation in the frequency of the light relative to the time within one period T at equal time intervals through a known unwrapping process to the phase information in the unwrapping processor 12 d based on the phase information of the complex analysis signal S H is executed.

The frequency curve of the frequency f of the light relative to the time t, Q 1 =t(f) is thereby obtained as shown in FIG. 8 .

In FIG. 8 , the horizontal axis represents the time t, and the vertical axis represents the variation in the frequency f of the light. The variation in the time t relative to the variation in the frequency f is expressed by the third order polynomial for convenience.

t ( f )= a 0 +a 1 f+a 2 f 2 +a 3 f 3

The frequency curve Q 1 of the light relative to the time axis illustrated in FIG. 8 is a curve that is obtained when a predetermined sweep frequency curve Q 2 in which the voltage V illustrated in FIG. 9 varies in a time within one period T as illustrated in FIG. 9 .

This voltage V(t) is expressed by the third order polynomial as the function of the sweep frequency f for convenience.

V ( f )= C 0 +C 1 f+C 2 f 2 +C 3 f 3

In addition, reference numerals a and C represent a coefficient.

Next, in the substitution processor 12 e, after a fifth step of obtaining a curve Q 3 having a correspondence relationship between the variation in the frequency f of the light within one period T, which is sampled at equal time intervals, and the variation in the voltage V within one period T, which is sampled at equal time intervals, as illustrated in FIG. 10 , based on the curves Q 1 and Q 2 is executed, a sixth step of converting the variation in the frequency f within one period T into correspondence voltage CV when the variation in the frequency f within one period T is subdivided into equal intervals is executed.

Next, the processor 12 executes a seventh step of substituting the voltage V to be applied to the waveform-swept light source applied in the first step into the correspondence voltage CV obtained by the sixth step, and thereby executes the calibration of the control voltage V. Then, the spectrum interference signal S(t) is again obtained by sweeping the waveform-swept light source with the correspondence voltage CV.

When the point-spread-function PSF(f) is obtained by the Fourier transform processor 2 a by Fourier transforming the spectrum interference signal S(t) obtained as described above, the point-spread-functions PSF(f) in which the distortion of the waveforms I 0 , I 1 , . . . , I m corresponding to the depth z is improved are obtained as shown in FIG. 11 .

FIG. 12 is a graph in which the horizontal axis corresponds to the depth z, and the point-spread-functions SPF(f) each having a normalized amplitude intensity I as illustrated in FIG. 11 are superimposed. It can be recognized from FIG. 12 , the superimposing level of the intensity of the peak is improved.

In addition, when the resampling technique disclosed in JP2010-3246A is used in conjugation with the technique according to this embodiment, the point-spread-function PSF(f) in which the distortion of the waveform relative to the depth z is improved can be obtained as illustrated in FIG. 13 .

When the tomographic image (not shown) is formed by analyzing the point-spread-function PSF(f) in which the waveform distortion is improved by the tomographic image obtainer (tomographic image forming section) 12 f , a high quality tomographic image with higher resolution can be obtained.

The present application is based on and claims priority from Japanese Patent Application No. 2014-34812, filed on Feb. 26, 2014, the disclosure of which is hereby incorporated by reference in its entirety.