WO2007016976A1

WO2007016976A1 - Optical reflectometry analysis based on first order and second order scatter signals

Info

Publication number: WO2007016976A1
Application number: PCT/EP2005/053857
Authority: WO
Inventors: Bernd Nebendahl; Josef Beller
Original assignee: Agilent Technologies, Inc.
Priority date: 2005-08-04
Filing date: 2005-08-04
Publication date: 2007-02-15
Also published as: DE212005000001U1

Abstract

The invention relates to determining a distributed property of a device under test - DUT- (3) by wavelength dependent selecting a first response signal (R1) and a second response signal (R2) returning from the DUT (3) in response to a probe signal (S) launched into a near end of the DUT (3), both response signals comprising a first order scatter signal originating from the forward traveling probe signal (S), and a second order scatter signal originating from the backwards traveling probe signal reflected at a far end of the DUT (3), and determining the distributed optical property of the DUT (3) on the base of the first order scatter signals and the second order scatter signals of the first response signal (P1) and the second response signal (P2).

Description

DESCRIPTION

OPTICAL REFLECTOMETRY ANALYSIS BASED ON FIRST ORDER AND SECOND

ORDER SCATTER SIGNALS

BACKGROUND ART

[0001] The present invention relates to determining an optical property of a device under test by optical reflectometry measurements.

[0002] For determining optical properties of an optical device under test (DUT), e.g. an optical fiber, it is known to apply a so-called optical time domain reflectometry (OTDR) or optical frequency domain reflectometry (OFDR). For that purpose an optical signal is coupled into the DUT, which travels along the DUT and which is partly scattered by the DUT, e.g. due to inhomogeneities in the silica structure (Rayleigh scattering) along the optical fiber or due to interaction of the optical signal with optical phonons (Raman scattering) or acoustical phonons (Brillouin scattering). Some of the scattered light travels back to the DUT input. The power of this returning light is measured and evaluated.

[0003] It is further known to separate different spectral components of the backscattered light from an optical fiber and to put these components into relation in order to obtain a physical property of the fiber. Well-known scatter signals returning at different wavelengths are so-called Raman scatterings. An arrangement for measuring a temperature distribution along an optical fiber by determining Raman scatterings is e.g. described in US 5,618,108.

DISCLOSURE

[0004] It is an object of the invention to provide an improved reflectometry measurement for determining an optical property of a device under test. The object is solved by the independent claims. Further embodiments are shown by the dependent claims.

[0005] According to embodiments of the present invention, an optical probe signal is launched into a near end of the DUT. The DUT is provided with an at least partially reflective mirror at its far end. A probe signal generated by a light source connected to the near end of the DUT is emitted into the DUT.

[0006] The DUT returns optical response signals back to the near end. These response signals are provided to an optical selector, that wavelength dependent selects a first response signal and a second response signal from the received response, wherein the first response signal therewith comprises a first fraction of the wavelength spectrum around a first center wavelength and the second response signal comprises a second fraction of the wavelength spectrum around a second center wavelength. Due to the mirror provided at the far end of the DUT, the response signals comprise each a first order scatter signal originating from the forward traveling probe signal and a second order scatter signal originating from the backwards traveling probe signal reflected at the mirror. An analyzer determines the distributed optical property of the DUT on the base of the first order scatter signals and the second order scatter signals of the first response signal and the second response signal.

[0007] To obtain time domain signals, different methods exist. A first method also known as optical time domain reflectometry (OTDR), directly measures the signals in the time domain. A further method, also known as optical frequency domain reflectometry (OFDR), uses a frequency-modulated signal, where the frequency is swept while the reflected intensity is recorded. This intensity versus frequency signal is transformed in the time domain using a Fourier transformation. After the transformation the obtained information is identical to the time domain method and any further corrections and calculations described here can be used.

[0008] By means of an optical detector, a first power function over the time or location is detected from the first response signal, and a second power function over the time or location is detected from the second response signal.

[0009] Thereby, the time scales of first power signal and the second power signal recorded over the time might be adapted to each other by individual time-to-location conversion factors on the base of the group velocity of the first response signal and the group velocity of the second response signal correspondingly.

[0010] In an embodiment, the analyzer determines a difference function between a logarithm of the first power functions and a logarithm of the second power function, and further determines an adjusted difference function by subtracting a linear function from the difference function. The linear function can be determined by fitting a linear function to the difference function.

[0011] In an alternative embodiment, the analyzer identifies each an inversion point at a transition between the first order scatter signal and the second order scatter signal of both power functions, and further determines an inverted first power function and an inverted second power function by mirroring the first power function and the second power function respectively at an axis through the inversion point perpendicular to the time or location axis. The analyzer then determines the property as a function of the first power signal, the inverted first power signal, the second power signal, and the inverted second power signal.

[0012] In an embodiment, the function is derived by taking a square root of a multiplication of the first power signal and the inverted first power signal and of the second power signal and the inverted second power signal and determining the function by setting these square roots into relation to each other.

[0013] In a further embodiment, a scaling factor of this function might be determined by measuring the property at one point, e.g. at the near end of the DUT and setting said function at this point into relation to said measured property.

[0014] In a further embodiment, the determination of the property is carried out in the logarithmic domain. Therefore, a logarithm of each the first power signal, the inverted first power signal, the second power signal, the inverted second power signal is determine. Further, a first average of the sum of the logarithm of the first power signal and the logarithm of the inverted first power signal, and a second average of the sum of the logarithm of the second power signal and the logarithm of the inverted second power signal are determined. The physical property is then determined by taking the difference between the first average and the second average.

[0015] With increasing time or location, the signal strength of the probing signal traveling through the fiber decreases and thus the corresponding response signals decrease with increasing time. Thus, disturbing noise signals superimposing to the response signals form an increasing part of the received signals. Especially signal to noise ratio (SNR) of second order scatter signals are much lower compared to the SNR of the corresponding first order scatter signals. As the physical property is determined on the base of both first and second order scatter signals, such noise might deteriorate the measurement. On the other hand, the first order scatter signal and the inverted second order scatter signal for physical reasons should only differ each in a constant value. Therefore, in an embodiment, the measurement can be substantially improved by applying an appropriate filter or estimation process.

[0016] Such estimation might be performed by estimating a linear change of the difference of the logarithm of the first (or second) power signal and the logarithm of the inverted first (or second) power signal and subtracting the estimated linear change divided by two from the logarithm of the first (or second) power signal.

[0017] The filter can be realized by calculating the difference of the logarithm of the first (or second) power signal and the logarithm of the inverted first (or second) power signal and applying an appropriate filtering (or smoothing) assuming that the process leading to the difference has only slow variations versus space. The filtered (or smoothed) signal divided by two will then be subtracted from the logarithm of the first (or second) power signal.

[0018] In an embodiment, the first response signal is selected to be one of: a Raman Antistokes signal with a center wavelength at a first response wavelength (λi) and a Raman Stokes signal with a center wavelength at a second response wavelength (λ₂), and the second response signal is the other of the Stokes signal or Antistokes signal, and wherein the property (τ) is a temperature profile along the DUT.

[0019] Regarding the low Raman signal level, a common reflectometry process might not deliver a sufficient signal-to-noise ratio (SNR) within an acceptable time frame. Therefore, in an embodiment, instead of using single pulses, a code correlation technique is performed. Therefore, a light source connected to the near end of the

DUT is controlled such that the stimulus signal comprises a plurality of optical pulses according to a digital sequence, and wherein the first power signal and the second power signal are derived by each performing an autocorrelation between the detected responses and said digital sequence.

[0020] Embodiments of the invention can be partly or entirely embodied or supported by one or more suitable software programs, which can be stored on or otherwise provided by any kind of data carrier, and which might be executed in or by any suitable data processing unit. Software programs or routines can be preferably applied in a control unit for controlling the light source, the optical selector, the optical detector and the analyzer, and in the analyzer itself.

BRIEF DESCRIPTION OF DRAWINGS

[0021] Other objects and many of the attendant advantages of embodiments of the present invention will be readily appreciated and become better understood by reference to the following more detailed description of embodiments in connection with the accompanied drawings. Features that are substantially or functionally equal or similar will be referred to by the same reference signs.

[0022] Fig.1 shows a basic setup of an OTDR system connected to an exemplary optical device under test,

[0023] Fig.2a shows a diagram with power functions of exemplary response signals returning from the DUT recorded by an OTDR system,

[0024] Fig.2b shows a diagram with attenuation normalized power functions according to Fig.2a,

[0025] Fig.2c shows a diagram with a temperature profile as result of an operation of the normalized functions of Fig.2b,

[0026] Fig.3 shows a diagram illustrating an alternative to Fig.2a - Fig.2c of deriving a temperature profile from the power functions of the response signals, and

[0027] Fig.4 shows an exemplary sketch diagram depicting the intensities of response signals returning from the DUT over the wavelength λ.

[0028] Fig.1 shows an OTDR system 1 , an optical source 2 and an optical device under test (DUT) 3. The OTDR system 1 comprises an optical coupler 10, an optical selector or splitter 11 an optical detector 12 and an analyzing unit 13 and a control unit 14. The optical coupler 10 optically connects the optical source 2, the DUT 3 and the optical splitter 11 to each other, such that a stimulus probe signal S emitted from the optical source 2 is provided to the DUT 3 and a response light returning from the DUT 3 is provided to the optical selector 11.

[0029] The control unit 14 controls the light source 2, the optical selector 11 , the detector 12 and the analyzer 13 by corresponding control signals C1- C4.

[0030] The probe signal S generated by the light source 2, preferably a laser diode with high output power, is preferably a narrow band signal at a probing or transmission wavelength λ₀.

[0031] The optical selector 11 wavelength dependently selects from the received response light a first spectral signal part or first response signal R1 around a first center wavelength λi and a second spectral part or second response signal R2 around second center wavelength λ₂, and provides the signals R1 and R2 to the optical detector 12.

[0032] The optical detector 12 performs an opto-electrical conversion of the response signals R1 and R2, generates a first power response over the time P1 and a second power response over the time P2, and provides these signals to the analyzer 13.

[0033] In an embodiment, the optical selector 11 provides both the first response signal R1 and the second response signal in parallel to a detector unit 12 comprising two separate detector elements for detecting the first power response P1 and the second power response P2 correspondingly.

[0034] Alternatively, the first response signal R1 and the second response signal R2 are provided sequentially to one detector element 12. Therefore, the control unit 14 might in a first step instruct the light source 2 to provide a first probe signal to the DUT 3, and the selector 11 to select the first response signal R1. In a second step, the light source 2 is instructed to provide a second probe signal and the selector 11 to select the second response signal R2. Switching the selection within the selector 11 might be performed by moving an optical shutter such that alternatively one of the selected response signals R1 or R2 is blocked form the detector 12. Further details of detecting different spectral components by means of an optical shutter are disclosed in the application EP 05105036.7 of the same applicant.

[0035] Scattering effects, e.g. Rayleigh scattering, Fresnel reflections, Raman or Brillouin scattering, cause a fraction of the forward traveling light to return from the DUT 3 eventually shifted in wavelength to the measurement system 1. Whereas the Raman and Brillouin backscattering signals are composed of Antistokes light and Stokes light returning at wavelengths different to the transmission wavelength λ₀, the Fresnel reflection or Rayleigh signals are returning the transmission wavelength λ₀.

[0036] In an embodiment, the physical property is a temperature profile along the DUT, e.g. an optical fiber installed along tunnel or a drill hole. As the Raman Antistokes light is subject to temperature changes of the DUT, the Raman scatterings are evaluated by selecting the so-called Antistokes part as one of the response signals, and the so-called Stokes part as the other response signal. In the following, by way of example and without limiting thereto, the first response signal R1 will be regarded as the Antistokes part and the second response signal R2 will be regarded as the Stokes part. The evaluation is thereby based on a ratio between the corresponding power responses P1 and P2.

[0037] The far end of the DUT 3 is provided with a broadband mirror that at least reflects a part of the probe signal S and of the signals R1 and R2 traveling from the near end to the far end of the DUT 3. The response signals R1 and R2 therefore each comprise a first order scatter signal part R11 and R21 originating from the forward traveling probe signal S, and a following second order scatter signal part R12 and R22 originating from the backwards traveling probe signal reflected at a far end of the DUT 3. Further, the response signals comprises each a power peak originating from the refection of the probe signal S at the mirror due to forward scattering. This power peaks mark the transition between the first order scatter signal part, and the second order scatter signal part of each response signal.

[0038] The group velocity of an optical material is depending on the wavelength. As the first response signal (Antistokes) R1 and the second response signal (Stokes) R2 are different in wavelength there is a difference in the group velocities of both the first and second response signal, so that events originated at one location occur at different times for both responses. Thus, the inversion time point t₀ of the first power signal P1 is different to the inversion time point t'_o of the first second response P2.

[0039] Therefore, in a further embodiment, the analyzer 13 performs an adjustment of the power responses P1 and P2.

[0040] P1 (x) = P1 (t*(ci C₀/ C1+C0))

[0041 ] P2(x) = P2(t*(c₂ C₀/ c₂+c₀))

[0042] with C₀ being the group velocity of the probe signal and Ci and C₂ being the group velocities of the first response signal R1 and the second response signal R2 respectively.

[0043] The inverted power respons signals can be written as follows:

[0044] P1'(x) = P1 (2x_o- x)

[0045] P2'(x) = P2(2x ₀ - x)

[0046] with x o = 1₀*(ci C₀/ Ci+c₀) = t' ₀* (c₂ C₀/ c₂+c₀), whereby x ₀ correponds to the length of the DUT 3 and is further also referred to as inversion point in the spatial domain. The curve parts with x < X₀ thereby represent the first order scatter signals scattering back from the probe signal traveling from the near fiber end at the location x=0 to the far fiber end at the location X₀, and the curve parts with X₀ < x ≤ 2X₀ represent the second order scatter signals scattering back (and reflected at the far fiber end) from the backward traveling probe signal S reflected at the far fiber end.

[0047] The analyzer 13 identifies, either in the time domain or in the spatial domain, the transition point time between the first order scatter signal parts from the received first or second power response. The identification of this time point can be carried out by identifying the power peak between the first order scatter signal and the second order scatter signal of the corresponding curve originating from the fiber end reflection. [0048] The first order scatter signal of the first response signal R1 can be written as follows:

[0049] P1 (x) = α(x,λo) K(x,λ_o,λi) α(x,λi), x < X₀

[0050] wherein α(x,λo) is the attenuation of the probing signal between the near end (at location 0) the location x, K(x,λ_o,λi) is the scatter factor from the probe signal S to the Raman Antistokes signal, R1 and α(x,λi) is the attenuation of the Raman Antistokes signal R1.

[0051] The second order scatter signal of the first response signal R1 can be denoted as follows:

[0052] P1 (x) = [α² (x,λo) /α(2xo-x,λo)] K(x,λo,λi) [α² (x,λi)/α(2xo-x,λi)], X > XQ

[0053] The first and second order scatter signal can be written accordingly:

[0054] P2(x) = α(x,λo) K(x,λ_o,λ2) α(x,λ2), x < X₀

[0055] wherein K(x,λ_o,λ₂) is the scatter factor from the probe signal S to the Raman Stokes signal R2 and α(x,λ2) is the attenuation of the Raman Stokes signal R2.

[0056] P2(x) = [α² (x,λo) /α(2xo-x,λo)] K(x,λ_o,λ₂) [α² (x,λ₂)/cc(2xo-x,λ₂)], _{x > Xo}

[0057] The invention is based on the insight that a suitable concatenation of a first order scatter signal with the inverted second order scatter signal form a function that is independent from the attenuation. Such suitable concatenation is the square root of a multiplication of the power response and the inverted power response mirrored at the axis X=X₀:

[0058] T1(x) = [P1 (x) * P1'(2xo- x)]^1/2=α(xo,λo) K(x,λo,λi) α(xo,λi)

[0059] T2(x) = [P2(x) * P2'(2xo- x)]^1/2=α(xo,λo) K(x,λo,λ₂) α(xo,λ₂)

[0060] R(x) = P1 (x)/P2(x)= K(x,λo,λi) α(xo,λi)/ K(x,λo,λ₂) α(xo,λ₂)

[0061] = α(xo,λi)/ α(xo,λ2) * F(τ(x)) , [0062] wherein α(xo,λi)/ oc(xo_>λ2) is a scaling factor that is independent from the location and the temperature and F(τ(x)) is a function that is dependent from the temperature profile τ(x) only.

[0063] This scaling factor can be determined by directly measuring the attenuations or from values taken from the fiber data sheet. Alternatively, in an embodiment, the scaling factor is determined by measuring the temperature at a single position, e.g. at the near fiber end within the reflectometer apparatus.

[0064] In the following the above method of determining the temperature profile of an optical fiber is described by means of diagrams shown in Fig.2a - Fig.2c.

[0065] Fig.2a shows a diagram with a first and a second power curves A1 and A2 depicted at logarithmic scale returning from the DUT measured at a detector of an OTDR system. The responses recorded overtime have been each converted into the spatial domain as described above, whereby x=0 marks the near fiber end and to X=X₀ marks the far fiber end. The curves parts A11 and A21 running from 0 to X₀ represent the first order scatter signals and the curve parts A12 and A22 running from X₀ to 2X₀ represent the second order signals.

[0066] The power curves A1 and A2 correspond to the logarithms of the first power response P1 and the second power response P2 described above:

[0067] A1 (x) = log P1(x)

[0068] A2(x) = log P2(x)

[0069] As the fiber attenuation for a homogeneous fiber is going with a negative e- function over the location (^e^"8*), both first and second power curves A1 and A2 show each a negative slope, wherein the slope angle represent the attenuation. In general the slope is different for the both functions A1 and A2.

[0070] Further, at the point x= X₀, a first reflection peak Pk1 and a second reflection peak Pk2 are shown for the first power response A1 and the second power response A2 respectively. These reflections result from a mirror at the far fiber end.

[0071] Fig.2b shows a diagram depicting, additionally to Fig.2a, mirrored or inverted power curves AV and A2'. The inverted power curves AV and A2':

[0072] AV(x) = A1 (2xo- x)

[0073] A2'(x) = A2(2xo- x)

[0074] As described above, a suitable concatenation in the non-logarithmic domain is the square root of a multiplication of the power response and the inverted power response. In the logarithmic domain, such operation corresponds to an average of the corresponding logarithmic power responses. Fig.2b shows each an average function B1 (x) coresponding to log T1 (x) and B2(x) coresponding to log T2(x) for 0 < x < X₀:

[0075] B1 (x) = 1/2 [A1 (x) + AV(2xo- x)] = log T1 (x)

[0076] B2(x) = 1 /2 [A2(x) + A2'(2xo - x)] = log T2(x)

[0077] Under an exemplary simple condition that the temperature over the length of the DUT 3 is constant and that the fiber is homogeneous over the length, the first order scatter signal part and the second order scatter signal part of each power response have the constant but in general different slope in the logarithmic domain, wherein the slope angle indicates the attenuation of the response signals. Therewith, both average functions B1 and B2 are horizontal straight lines.

[0078] Fig.2c shows a diagram with a temperature distribution as result of an operation of the power responses of Fig.2b:

[0079] τ (x) = constant/[B2(x) - B1 (x)] = constant/[log T2(x) - log T1 (x)]

[0080] The difference between the average functions B2 and B1 thereby is proportional to the reciprocal temperature τ (= constant/τ). In the case that the slopes would be the same for both functions A1 and A2 the constant would be equal to 1. Corresponding to the description above, the constant can be determined by additionally measuring the temperature of the fiber at a single point. As for the simple example described here the temperature is assumed to be constant over the whole fiber, the distance between both normalized responses of Fig.2c is constant over the length x, and thus the temperature curve shown here is a constant over the length of the fiber. [0081] With increasing time or location, the signal strength of the probing signal traveling through the fiber decreases and thus the received response signals R1 and R2 decrease. Thus, disturbing noise signals superimposing to the response signals form an increasing part of the received signals. This might especially apply for the second order scatter signals A1'(x) and A2'(x). Due to the fact that the first order scatter signal and the inverted second order scatter signal A1 (x) and A1'(2xo- x), or A2(x) and A2'(2xo - x) for physical reasons assuming a spatially homogeneous attenuation should only differ by a linearly increasing value, the influence of the noise to the measurement can be substantially decreased by an appropriate filter or average process or by fitting a straight line (or a low order polynomial) to the difference. Thus for determining the average functions B1 and B2 the following

[0082] B1 (x) = A1 (x) - 1 /2 AVERAGE [A1 (x) - A1 '(2X₀ - x)]

[0083] B2(x) = A2(x) - 1 /2 AVERAGE [A2(x) - A2'(2xo - x)]

[0084] Fig.3 shows a diagram illustrating an alternative to Fig.2a - Fig.2c of deriving a temperature profile from the power functions of the response signals. Fig.3 therefore shows the first and a second power curves A1 and A2 of Fig.2a. Further, a difference curve B is shown, that is derived by subtracting the first power curve A1 from the second power curve A2. Similar to the example described above, the temperature is assumed to be constant over the whole homogeneous fiber. Thus, the difference curve B is a linear function with a positive slope. This slope is due to the attenuation difference of the response signals. Without attenuation difference, this difference curve is proportional to the reciprocal temperature τ (= constant/τ). As the locations x=0 and x= 2X₀ both refer to the near end of the fiber, the temperature at this locations must be equal. Thus, an attenuation-adjusted difference function B' is determined by a adding or subtraction a linear function (or a low order polynomial) such that the values at the location x=0 and x= 2X₀ of the attenuation-adjusted difference function B' are equal:

[0085] B'(0) = B'(2xo)

[0086] For the same reason as described above, disturbing noise signals might superimposing to the response signals. Therefore, in a further embodiment, the linear function is determined by fitting a straight line (or a low order polynomial) to the difference function (B).

[0087] In the non-logarithmic domain, a function corresponding to the difference function is determined by taking the ratio between the second power function P2 and the first power function P1 , and fitting this ratio to en exponential function over the location (C*e^ax, with A and C being constants to be determined).

[0088] Regarding the low Raman signal level, a common OTDR process might not deliver a sufficient signal-to-noise ratio (SNR) within an acceptable time frame. Therefore, instead of using single pulses, a code correlation technique is used, which significantly improves signal strength and thus SNR.

[0089] Using such correlation technique, the light source 2 is modulated according to a specific predetermined digital sequence. The response signals S1 and S2 are composed of an overlay of a plurality of corresponding shifted impulse responses from the DUT. To determine the resulting impulse responses and therewith the resulting power responses P1 and P2, the analyzing circuit 13 might perform a correlation between each the probe signal and the partial responses received. As result of these correlations, the power responses P1 and P2 versus time are determined. Alternatively this correlation might be performed in the digital time domain by sampling both the partial responses and digitally convoluting the sampled sequences with the predetermined digital signal.

[0090] The autocorrelation function of a digital pseudo random code shows a maximum at zero shift and residual side lobes. Thus, neglecting the side lobes, the correlation of a probe signal S with a response signal from DUT 3 represents the impulse answer of DUT 3. Alternatively, complementary codes like so-called Golay codes can be used. Such codes have the advantage that the side lobes are cancelled out. Further information of applying impulse sequences in OTDR applications is described in the international application PCT/EP2004/052670 of the same applicant.

[0091] Fig.4 shows an exemplary sketch diagram depicting the intensity I of the first response signal R1 (Antistokes signal) and the second response signal R2 (Stokes signal) over the wavelength λ. Further, the transmission wavelength λ₀ the first center wavelength λ_u and second center wavelength λ₂ are depicted at the wavelength axis.

[0092] In a possible realization, the transmission wavelength λ₀ is about 1064 nm, the first center wavelength at

of the Antistokes signal is about 1014 nm and the second center wavelength λ₂ of the Stokes signal is about 1114 nm. Thus for selecting the first and second response signals R1 and R2, corresponding band pass filters might each have a range of about 100 nm around first and second center wavelengths λi and λ₂ respectively.

Claims

1. A method of determining a distributed property of a device under test - DUT- (3), comprising:

- wavelength dependent selecting a first response signal (R1 ) and a second response signal (R2) returning from the DUT (3) in response to a probe signal

(S) launched into a near end of the DUT (3), both response signals comprising a first order scatter signal originating from the forward traveling probe signal (S), and a second order scatter signal originating from the backwards traveling probe signal reflected at a far end of the DUT (3), and

- determining the distributed optical property of the DUT (3) on the base of the first order scatter signals and the second order scatter signals of the first response signal (P1 ) and the second response signal (P2).

2. The method of claim 1 , further comprising:

- detecting from the first response signal (R1 ) a first power function (P1 ) over the time or location and from the second response signal (R2) a second power function (P2) over the time or location,

- determining a difference function (B) between a logarithm of the first power functions (P1 ) and a logarithm of the second power function (P2), and

- determining an adjusted difference function (B') by subtracting a linear function from the difference function (B).

3. The method of claim 2, wherein the linear function is determined by fitting a linear function to the difference function (B).

4. The method of claim 1 , further comprising:

- detecting from the first response signal (R1 ) a first power function (P1 ) over the time or location and from the second response signal (R2) a second power function (P2) over the time or location, - identifying each an inversion point (X₀) at a transition between the first order scatter signal and the second order scatter signal of both power functions (P1 , P2),

- determining an inverted first power signal (P1 ') and an inverted second power signal (P2') by mirroring the first power function (P1 ) and the second power function (P2') respectively at an axis through the inversion point (X₀) perpendicular to the time or location axis, and

- determining the property as a function of the first power signal (P1 ), the inverted first power signal (PV), the second power signal (P2), and the inverted second power signal (P2').

5. The method of claim 4, wherein the property is determined on the base of a square root of a multiplication of the first power signal (P1 ) and the inverted first power signal (P2') and a square root of a multiplication of the second power signal (P2) and the inverted first power signal (P2').

6. The method of claim 4 or 5, wherein a scaling factor of the function of the first power function (P1 ), the inverted first power function (PV), the second power function (P2), and the inverted second power function (P2') is determined by measuring the property at one location, preferably at the near end of the DUT (3), and setting said function at this location into relation to said measured property.

7. The method of claim 4 or an one of the claims 5 or 6, wherein the property is determined by

- taking a first average (B1 ) of the sum of the logarithm of the first power function (A1 ) and the logarithm of the inverted first power function (AV),

- taking a second average (B2) of the sum of the logarithm of the second power function (A2) and the logarithm of the inverted second power function (A2'),

- determining a difference function between the first average and the second average, and

- deriving the property on the base of the difference function.

8. The method of claim 7, wherein the first average (B1) and/or the second average (B2) is determined by

- filtering the difference function of the logarithm of the first power function (A1 ) and the logarithm of the inverted first power function (AV) and the difference function of the logarithm of the second power function (A2) and the logarithm of the inverted second power function (A2') and

- subtracting this filtered function divided by two from the logarithm of the first power function (A1 ) and/or the logarithm of the second power function (A2).

9. The method of claim 7, wherein the first average (B1 ) and/or the second average (B2) is determined by

- fitting a defined function, preferably a straight line, to the difference function of the logarithm of the first power function (A1 ) and the logarithm of the inverted first power function (AV) and to the difference function of the logarithm of the second power function (A2) and the logarithm of the inverted second power function (A2') and

- subtracting this defined function divided by two from the logarithm of the first power function (A1 ) and/or the logarithm of the second power function (A2).

10. The system of claim 2 or any one of the above claims, wherein first power function (P1 ) and the second power function (P2) recorded over the time are individually converted to functions over the location on the base of each the group velocity of the first response signal (R1 ) and the group velocity of the second response signal (R2).

11. The method of claim 1 or any one of the above claims, wherein the first response signal (R1 ) is selected to be one of: a Raman Antistokes signal with a center wavelength at a first response wavelength (λi) and a Raman Stokes signal with a center wavelength at a second response wavelength (λ₂), and the second response signal (R2) is the other of the Stokes signal or Antistokes signal, and wherein the property (τ) is a temperature profile along the DUT (3).

12. The method of claim 1 or any one of the above claims, wherein a light source (2) connected to the near end of the DUT (3) is controlled such that the stimulus signal (S) comprises a plurality of optical pulses according to a digital sequence, and wherein the first power function (P1 ) and the second power function (P2) are derived by each performing an autocorrelation between the detected responses and said digital sequence.

13. An optical reflectometry system (1) for determining a distributed property of a device under test - DUT- (3), comprising:

-an optical selector (11 ) adapted for wavelength dependent selecting a first response signal (R1 ) and a second response signal (R2) returning from the

DUT (3) in response to a probe signal (S) launched into a near end of the DUT

(3), both response signals comprising a first order scatter signal originating from the forward traveling probe signal (S), and a second order scatter signal originating from the backwards traveling probe signal reflected at a far end of the DUT (3), and

-an analyzer (13) adapted for determining the distributed optical property of the DUT (3) on the base of the first order scatter signals and the second order scatter signals of the first response signal (P1 ) and the second response signal (P2).

14. A software program or product, preferably stored on a data carrier, for controlling or executing the method of claim 1 or any one of the above claims 2-13, when run on a data processing system such as a computer.