JPH05273043A - Radiation temperature measuring apparatus, emissivity measuring apparatus and decision method of emissivity cumulative ratio-emissivity correlation - Google Patents

Abstract

PURPOSE:To measure emissivity and temperature near the surface of a process. CONSTITUTION:For example, a monochromatic type radiation ther-mometer 120, a dichromatic type radiation thermometer 130 and an arithmetic unit are arranged and a rolled steel plate or aluminum during a rolling process is used as object 110 to be measured. A stationary state in which emissivity takes a fixed value hourly (where no oxidation film is formed on the surface thereof or the thickness (d) of the oxide film grows sufficiently) is detected. The temperature containing an estimation (error range) of measuring errors under such a condition is measured using the emissivity corresponding to the condition.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an emissivity temperature measuring device and an emissivity measuring device for measuring emissivity and temperature from light of two or more wavelengths.

[0002]

2. Description of the Related Art In the manufacturing process of semiconductor materials, metal materials, etc., the surface of the material to be processed is oxidized, alloyed,
Various reactions such as vapor deposition occur artificially or spontaneously, and the physical properties of the material change significantly with this reaction. However, sensing inside such a material is extremely difficult, and it is not practically desirable to attach many sensors and capture much data. Therefore, the surface temperature is selected as a parameter having the greatest influence among the above-mentioned physical property values and state values, and this is used for online sensing, or a theoretically calculated change state predicted value is used for control.

For online material surface temperature sensing, there are contact type and non-contact type temperature sensing techniques.
The contact-type temperature sensing technology using a thermocouple, thermistor, etc. is practically used because there are too many problems such as temperature limitation, only the temperature at the sensing point is measured, and contamination (contamination) due to contact. The processes involved are extremely limited. On the other hand, there is “radiation temperature measurement” as a non-contact temperature measurement. This is a practical technique in the process line of metal materials such as steel and aluminum, and the radiation thermometer is commercially available as a product.

A monochromatic radiation thermometer, which is a type of radiation thermometer, measures the temperature by detecting the brightness temperature (unit K) S from the intensity of light of a specific wavelength λ from the material to be measured 510 by the radiation sensor 520. It is a thing. From the spectral emissivity ε of the light having the wavelength λ on the surface of the material to be measured, the temperature near the surface can be obtained by the following equation 1 (C ₂ indicates a radiation constant).

[0005]

[Equation 1]

Here, the spectral emissivity ε is assumed to be a constant determined by the material and surface state of the material. As shown in FIG. 18, this constant is calculated by using Equation 2 from the temperature T ^* obtained by measuring the material 510 to be measured by contact temperature measurement with a thermocouple and the brightness temperature S of the monochromatic radiation thermometer 520. It has been decided (this is called offline work).

[0007]

[Equation 2]

On the other hand, there is a two-color thermometer which detects the temperature S with two wavelengths. The temperature measurement accuracy of these radiation thermometers does not pose a problem when the spectral emissivities at the two wavelengths to be measured are substantially equal or have a fixed proportional relationship. However, when the surface state of the heat object suddenly changes due to an oxidation reaction or the like and the spectral emissivity deviates from the above relationship, the measurement accuracy is significantly deteriorated. In particular, the error is even larger in a monochromatic radiation thermometer. Therefore, a temperature calculation method corresponding to the change in spectral emissivity is desired, and research is being made on this point, and an improved two-color thermometer that can be used even when the emissivity changes is considered.

[0009] As the technique, the method described in "Japanese Patent Publication No. 3-4855" and "Tanaka, DP Dewitt: Theory of a Ne"
w Radiation Thermometry Method and an Experimental
Study Using Galvannealed Steel Specimens (Proceedings of the Society of Instrument and Control Engineers Vol. 25, No. 10, 1031/103
Page 7 October 1989) "
E (Thermometry Re-established by Automatic Compen
sation of Emissivity) method. Both methods are basically the same, so the former will be explained.

The spectral emissivity ε of the radiant energy (light) emitted from the process material is obtained by using the Wien (Vienna) approximation rule, and is represented by the following equations 3 and 4 at wavelengths λ1 and λ2. The temperature T is deleted from the equations to obtain the equation 5 (the symbols used in these equations are shown in Table 1).

[0011]

[Table 1]

[0012]

[Equation 3]

[0013]

[Equation 4]

[0014]

[Equation 5]

The left side of the equation 5 is the ratio of "spectral emissivity to the power of the wavelength". For the sake of simplicity, the term emissivity exponentiation ratio will be used hereinafter.

Now, in Japanese Patent Publication No. 3-4855, the following equation 6
The correlation function "f" between the spectral emissivity ratio (ε 1 / ε 2) and the emissivity exponentiation ratio as shown in Fig. 4 is determined in advance by measurement, and when measuring the temperature, the correlation is calculated from the emissivity exponentiation ratio calculated by Equation 5 above. Obtain the spectral emissivity ratio by the function f,
The temperature T is also calculated by the following Equation 7.

[0017]

[Equation 6]

[0018]

[Equation 7]

[0019]

With the monochromatic radiation thermometer, highly accurate temperature measurement cannot be performed unless the spectral emissivity ε shown in FIG. 18 is actually measured. However, the spectral emissivity ε depends on the material type, temperature range, measurement wavelength, measurement angle of the radiation sensor, etc. For example, if the measurement angle of the radiation sensor is different, the spectral emissivity ε also takes different values. For this reason, the emissivity ε is rarely measured in every case, and it is rarely set based on it. It is appropriate to set the value listed in the data book only, or to rely on intuition and experience. Is set to a certain value. Therefore, the measured temperature may have a large error with respect to the true temperature.

In a high temperature process of a metal material or a silicon semiconductor, an oxide film is formed on the surface, and optical interference or the like is caused in the oxide film, so that the spectral emissivity ε is greatly changed depending on the oxide film, and the alloying process is performed. The same applies to changes in the material itself and changes in the surface roughness (roughness). These changes in the spectral emissivity ε are not detected by the above-described radiation thermometer (monochromatic type, two-color thermometer), and the presence or absence of the change in the spectral emissivity ε in the online process and its change width,
The temperature error range etc. due to it is hardly checked. Therefore, the temperature control error occurs in the product after the process is completed, and an unexpected defective product may appear.

On the other hand, the improved two-color thermometer is intended to solve the above-mentioned problem, but when it is applied to the surface temperature measurement in the process of forming the surface thin film, "f" of the formula 5 is determined in advance by off-line work. Therefore, the correlation function “f” is determined by the off-line work as shown in FIG. 18, like the emissivity measurement of the monochromatic thermometer described above. This improved type 2
Even the color thermometer does not output the temperature measurement error range.

As described above, the conventional radiation thermometer outputs a temperature measurement value, but does not output what degree of reliability the output temperature has, that is, a temperature measurement error. Therefore, when used for process control, preventive measures and adaptive control due to such temperature deviation could not be performed. In order to solve the above problem, the present invention provides a radiation thermometer that outputs an error prediction value.

[0023]

In order to solve the above problems, the radiation temperature measuring device of the present invention detects light emitted from the surface of the object to be measured at at least two different wavelengths and outputs it as a detection signal. A parameter corresponding to the emissivity power ratio is calculated from the radiation thermometer and the detection signal of at least two wavelengths, the temporal average value of the parameter is obtained, and whether the parameter is within a predetermined range with respect to the temporal average value is calculated. Judge,
When the parameter is within a predetermined range, the calculation means for calculating the temperature in the vicinity of the surface of the object to be measured from the detection signal based on the emissivity or the emissivity ratio of the surface of the object to be measured defined by the parameter is provided. Characterize.

The calculating means may be characterized by calculating the temperature in the vicinity of the surface of the object to be measured from the detection signal of at least one wavelength or the detection signal of at least two wavelengths.

The calculating means may be further characterized by further calculating an upper limit value and a lower limit value of the error near the surface of the object to be measured from the maximum value and the minimum value of the parameter corresponding to the emissivity power ratio.

When the parameter corresponding to the emissivity exponentiation ratio is out of a predetermined range, an alarm signal indicating that the emissivity is unstable may be output to the outside.

The parameter may be the difference of the reciprocal of the value of the detection signal of the wavelength of 2.

Further, the emissivity measuring apparatus of the present invention includes a radiation thermometer that detects light emitted from the surface of the object to be measured at at least two different wavelengths and outputs it as a detection signal, and a detection signal at at least two wavelengths. Calculate the parameter corresponding to the emissivity power ratio or the emissivity ratio from, calculate the temporal average value of the parameter, determine whether the parameter is within a predetermined range with respect to the temporal average value, and set the parameter within the predetermined range. In some cases, a calculating means for calculating the emissivity of the light emitted from the surface of the object to be measured from the parameter or the detection signal of at least two wavelengths is provided.

Here, the parameter may be the difference of the reciprocal of the value of the detection signal of the wavelength of 2.

Further, the present invention may be the following in addition to the above. That is, the radiation temperature measuring device of the present invention includes a radiation thermometer that detects light emitted from the surface of the object to be measured at at least two different wavelengths and outputs it as a detection signal, and an emissivity from the detection signal at at least two wavelengths. Calculate the parameter corresponding to the exponentiation ratio, based on the correlation between the parameter previously measured for the DUT and the emissivity or emissivity ratio of the DUT, the upper limit of the temperature near the surface of the DUT from the detection signal And a calculation means for calculating the lower limit value or the error range.

The correlation is defined by the parameter and the central value ε of the emissivity.
_It may be characterized by a correlation with any one of ₀ , the upper limit ε _U , and the lower limit ε _L of the emissivity. Alternatively, the correlation may be a correlation between the parameter and any one of the center value, the upper limit value, and the lower limit value of the emissivity of the two wavelengths.

The correlation may be determined in advance based on the data measured on the object to be measured off-line in advance.

As a method of determining the correlation, the emissivity exponentiation ratio and the emissivity or the emissivity ratio of a predetermined object to be measured are measured off-line in advance, and the measured data group is regressed for regression. When a regression function of the central value of emissivity or emissivity ratio is determined and there are data groups within a predetermined error range from the central value of emissivity or emissivity ratio by the regression function, the upper and lower limits of the error range are the upper and lower limits. As the regression function of the value,
The correlation between the emissivity exponentiation ratio and the emissivity of the DUT is determined.

[0034]

In the radiation temperature measuring apparatus of the present invention, the radiation thermometer detects the light radiated from the surface of the object to be measured, and the emissivity exponential ratio ([ε _i to λ _i power] / [ε _j λ
j]], where ε _i and ε _j are parameters corresponding to the emissivity of the DUT at wavelengths λ _i and λ j (for example, emissivity exponentiation ratio, difference of reciprocal of detection signal value of 2 wavelengths, etc.) Is calculated. This parameter takes a substantially constant value when the surface of the measured object is in a steady state. At this time, using the detection signal of one wavelength of the light emitted from the surface of the DUT and the wavelength thereof, or the detection signal of at least two wavelengths, the emissivity of the wavelength and the DUT are measured. It is possible to calculate the temperature near the surface of, and it can be obtained by a simple calculation process. That is, the emissivity can be easily obtained from the parameter corresponding to the emissivity exponential ratio or the detection signal of at least two wavelengths, and the temperature near the surface of the object to be measured can be easily calculated from the emissivity and the detection signal of at least one wavelength. Calculated by arithmetic processing.

Here, the error range of the temperature in the vicinity of the surface of the object to be measured can be obtained by a simple calculation process from the change in the parameter corresponding to the emissivity power ratio when the surface of the object to be measured is in a steady state.

On the other hand, when the parameter corresponding to the emissivity exponentiation ratio is not within the predetermined range, it is possible to output an alarm signal indicating that the temperature near the surface of the object to be measured is not accurately obtained.

Further, in the emissivity measuring apparatus of the present invention, the emissivity of the object to be measured can be obtained from the detection signals of the two wavelengths by a simple arithmetic processing, similarly to the above-mentioned radiation temperature measuring apparatus.

According to the radiation temperature measuring apparatus of the present invention, the temperature in the vicinity of the surface of the object to be measured and the upper limit value of this temperature error are calculated based on the correlation between the parameter measured in advance on the object to be measured and the emissivity of the object to be measured. And the lower limit value or the error range are calculated, the following is performed.

The light emitted from the surface of the object to be measured is detected by the radiation thermometer, and the parameter corresponding to the emissivity exponentiation ratio is calculated by the computing means. Then, in order to determine the temperature and the upper limit value and the lower limit value or the error range of this temperature error, the correlation between this parameter and the emissivity or emissivity ratio, the maximum value and the minimum value or the allowable range is measured in advance. On the basis of this correlation, the temperature near the surface of the object to be measured and the upper limit value and lower limit value or error range of this temperature error are obtained from the detection signal.

[0040]

Embodiments of the present invention will be described with reference to the drawings.
FIG. 1 shows a schematic configuration of an embodiment of the present invention. This embodiment is composed of a two-color radiation thermometer 130 and an arithmetic unit 140, and FIG. 1 shows a state in which a rolled steel plate or aluminum during the rolling process is used as the DUT 110.

The two-color radiation thermometer 130 is an existing one, and detects light emitted from the surface of the object 110 to be measured by a radiation sensor having at least two different wavelengths λ ₁ and λ ₂ built therein. The brightness temperatures S ₁ and S ₂ are output. The arithmetic unit 140 calculates the measured object 11 from the brightness temperatures S ₁ and S _2.
The temperature near the surface of 0 is calculated, and the data D _in and D _out including the measurement results are exchanged with the outside. As shown in FIG. 2, emissivity exponentiation ratio calculation block 210, determination block 220, emissivity exponentiation ratio time average calculation block 2
30, an emissivity calculation block 250, a maximum / minimum emissivity value calculation block 240, and a temperature calculation block 260. The arithmetic unit 140 is composed of a computer, its input / output interface and its software (this point is the same in other embodiments).

The emissivity exponentiation ratio calculation block 210 calculates the emissivity exponentiation ratio ([ε _i to λ _i ]] / from the brightness temperatures S ₁ and S ₂ .
The parameter Ku (t) corresponding to [ε _{j to} the power of λ j] is calculated. Here, the emissivity exponentiation ratio itself is used as the parameter Ku (t), and is calculated by the above-described equation 5.
The emissivity exponentiation ratio time average calculation block 230 sequentially weights and averages the parameter Ku (t) to obtain a temporal average value <Ku.
(T)> is calculated. This is shown in Equation 8 and the parameter Ku (t) sampled at a constant period (M / m)
Can also be stopped by averaging m.

[0043]

[Equation 8]

The decision block 220 determines the parameter Ku calculated by the time average calculation block 230 of the emissivity exponentiation ratio.
For the temporal average value <Ku (t)> of (t), it is determined whether the sampled parameter Ku (t) is within a predetermined range. For example, a temperature error of ± 5 near 1000 degrees
If it is in degrees, determine whether it is within ± 5%. This range setting will be described later. If the parameter Ku (t) is out of the predetermined range, an alarm signal is output to inform that the emissivity is unstable. Parameter Ku
When (t) is within a predetermined range, the temperature T near the surface of the DUT 110, the maximum value T _max , and the minimum value T _{min of} this temperature are obtained in other parts.

The emissivity calculation block 250 calculates the averaged emissivity ε (t) from the temporal average value <Ku (t)> (see equation 9). Further, the maximum value / minimum value calculation block 240 uses the sampled parameter Ku (t).
From the maximum and minimum values of the maximum emissivity ε (t) ε
_max (t) and minimum value ε _min (t) are calculated (Equation 10
a, 10b).

[0046]

[Equation 9]

[0047]

[Equation 10]

Emissivity exponentiation ratio calculation block 210, judgment block 220, emissivity exponentiation ratio time average calculation block 2
It is possible to configure the emissivity measuring unit 270 with 30, the emissivity calculation block 250, and the emissivity maximum / minimum value calculation block 240, and emissivity can be calculated in this part.

The temperature calculation block 260 includes the averaged emissivity ε (t), the detection wavelength λ ₁ and the brightness temperature S ₁ (or
The temperature T near the surface of the DUT 110 is calculated from the detection wavelength λ ₂ and the brightness temperature S ₂ ) to obtain the maximum value ε of the emissivity ε (t).
The maximum value T _max and the minimum value T _min are obtained from (t) _max , the minimum value ε (t) _min , the detection wavelength λ ₁ and the brightness temperature S ₁ (or the detection wavelength λ ₂ and the brightness temperature S ₂ ). Although these are obtained by the same calculation as the above-mentioned formula 1, these calculation formulas are shown in formulas 11 to 13 as a precaution. Here, Equation 11 is
Although λ ₁ and S ₁ are shown, the detection wavelength λ ₂ and the brightness temperature S ₂ can also be obtained by substituting the above-mentioned formula 1 in exactly the same manner.

[0050]

[Equation 11]

[0051]

[Equation 12]

[0052]

[Equation 13]

Next, the principle of this device will be described.

The spectral emissivity ε of the DUT 110 depends on the surface state thereof as described above. For example, when an oxide film is formed on the surface, light from the DUT interferes due to reflection of the DUT-oxide film, reflection of the oxide film-atmosphere, etc., and their reflectivity, oxide film thickness d, oxidation. It is known that the spectral emissivity ε changes according to the complex refractive index of the film, the wavelength, and the angle at which light is emitted (that is, the angle formed by the radiation sensor and the surface normal of the DUT (installation angle)). There is. When this example is shown, the change is as shown in FIG. 3 or FIG. In these figures, the vertical axis represents the emissivity ε and the horizontal axis represents the oxide film thickness d formed on the surface. Here, FIG. 3 shows a case where ordinary steel is used as the object 110 to be measured, and FIG. 5 shows a case where aluminum is used as the object 110 to be measured. The radiation sensor is on the surface normal to the object to be measured. The radiation sensor has Si (detection wavelength λ ₁ = 1 μm), Ge (detection wavelength λ ₂ = 1.6 μm), PbS (detection wavelength λ ₃ = 2 μm).
m) and PbSe (detection wavelength λ ₄ = 4 μm) can be used, and the emissivities ε _{1 to} ε ₄ indicate the emissivity for those wavelengths, respectively.

When no oxide film is formed on the surface (d
= 0), the emissivities ε _{1 to} ε ₄ indicate predetermined values,
It changes as the oxide film thickness d increases. This change is different for each wavelength λ _{1 to} λ ₄ , and this is the cause of the above-mentioned error in the monochromatic and two-color radiation thermometers. Further, the emissivity exponentiation ratio Ku _ij (Ku _ij = [ε _{i λ i}
Power] / [ε _{j to} the power of λ j]; where ε _i and ε _j are wavelengths λ _i
, Λj, the emissivity of the DUT also changes. Figure 4
6 shows changes in the emissivity power ratios Ku ₁₂ and Ku ₃₄ with respect to the emissivity in FIG. 3, and FIG. 6 shows changes in the emissivity power ratio with respect to FIG.

When the oxide film thickness d becomes large to some extent, the emissivities ε _{1 to} ε ₄ converge to a certain value. This is because the light from the DUT is absorbed by the oxide film (due to the complex component of the complex refractive index), and the oxide film becomes thick enough as a bulk oxide. Since the emissivity becomes a certain value, it becomes possible to measure the temperature with a monochromatic radiation thermometer. The monochromatic thermometer measurement assumes that the emissivity is stable, and this emissivity needs to be obtained in advance. There is no simple method for obtaining the emissivity in advance as described in the above-mentioned conventional technique.
It was according to the method of.

The present invention pays attention to this point, and the emissivity is temporal.
A steady state (oxide film formed on the surface
When not formed or the oxide film thickness d has become sufficiently large
When the temperature is detected online and the temperature and its
You can measure online by using the emissivity corresponding to the condition.
It is a scam. In this state, the emissivity power ratio is
Shows a constant value, especially the detection wave of the two-color radiation thermometer 130
Long λ₁, Λ₂Emissivity for ε₁, Ε₂(Or ε₃, Ε
_Four) Are equal as shown in FIGS. 3 and 5, the parameters are
If emissivity power ratio is used as Ku (t), the parameter
Ku (t) is shown in Expression 14. Emissivity ε₀(= Ε₁=
ε₂) Is calculated by Equation 15. Emissivity ε ₀In terms of time
The equalized emissivity ε (t) is shown in Equation 16. this is
Shows an analog average, sampled and averaged
In that case, the above equation 8 is obtained. Here, Equation 8 is the same as Equation 16
It is a rewriting of the strict average formula into a practical formula.

[0058]

[Equation 14]

[0059]

[Equation 15]

[0060]

[Equation 16]

In a transient state in which the emissivity changes with time, it is possible to measure the temperature with the improved two-color thermometer. This point can be easily realized by activating a routine for processing as the improved two-color thermometer by using the determination result (alarm signal output) of the determination block as a trigger. For this routine, the parameter Ku
Correlation block 29 between (t) and the emissivity ratio (ε ₁ / ε ₂ ).
The emissivity is obtained from 0 (see the dotted line portion in FIG. 2). Then, it suffices to add a routine for obtaining the temperature T from the emissivity ratio using the equation 7 to the temperature calculation block 260.

From the emissivity obtained as described above, Equation 1
The temperature is obtained by the calculation of (that is, Expression 11). Alternatively, it can be obtained by an operation equivalent to Expression 1. In Equation 7, ε ₁ = ε ₂
The temperature can be obtained by the equation (17). This equation 17 is ε
_{When 1} = ε ₂ , it has the same value as in Expression 11.

[0063]

[Equation 17]

Here, the temperature error ΔT is expressed by the equation 18 (ε ^* is an error with respect to the true emissivity ε). The relationship between this emissivity error ε ^* and the variation ΔKu of the emissivity exponential ratio with respect to the temporal average value <Ku (t)> is shown in Expression 19.
When the detection wavelength λ ₁ = 1 μm and the detection wavelength λ ₂ = 1.6 μm, the temperature error ΔT-emissivity exponentiation ratio fluctuation ΔKu is shown in the chart of FIG. 7. The determination block determines the range of the parameter Ku (t) using these relationships to make the determination. For example, A and A ′ in the figure show the limits of the range where the temperature error is ± 5 degrees (the range of the temperature error ± 5%) in the vicinity of 1000 degrees, and in the decision block, the parameter Ku within this range is approximately ±. It is judged whether it is within the range of 4%.

[0065]

[Equation 18]

[0066]

[Formula 19]

Next, the operation of this apparatus will be described.

From the DUT 110, light having a wavelength distribution corresponding to the temperature T, the reflectance of the oxide film, the oxide film thickness d, the complex refractive index of the oxide film, the wavelength, and the angle at which the light is emitted is emitted. .. Of this light, the light of wavelengths λ ₁ and λ ₂ is the two-color radiation thermometer 1
Measured at 30 and output as brightness temperatures S ₁ and S ₂ . Here, when Si and Ge are used for the radiation sensor,
“Detection wavelength λ ₁ = 1 μm, detection wavelength λ ₂ = 1.6 μm”
Is.

In the calculation unit 140, the emissivity exponentiation ratio calculation block 210 calculates the parameter Ku (t) from the brightness temperatures S ₁ and S ₂ , and the time average calculation block 230
Then, the temporal average value <Ku (t)> during the time M is obtained. In the decision block, it is decided from the temporal average value <Ku (t)> and the parameter Ku (t) whether it is a transient state or a steady state. When the oxide film thickness d increases with time, the parameter Ku (t) is large when the change in the parameter Ku (t) (change in Ku ₁₂ ) in FIG. 4 or 6 is large (for example, d = 0. The vicinity of 6 μm (FIG. 4) and the vicinity of 0.1 μm (FIG. 6) are determined to be a transient state, and when the change is small (for example, d = 2.2 μm).
In the vicinity (FIG. 4) and in the vicinity of 0.2 μm (FIG. 6), it is determined to be a steady state. After the time M after the parameter Ku (t) becomes a steady state value, a stable temporal average value <Ku (t)> is obtained.

In the steady state, the emissivity calculation block 2
The averaged emissivity ε (t) is calculated at 50, and the maximum value / minimum value calculation block 240 calculates the maximum value ε _max (t) and the minimum value ε _min (t) of the emissivity ε (t). It In the temperature calculation block 260, the temperature T near the surface of the DUT 110, the maximum value T _max , and the minimum value T _min are obtained from the brightness temperature S ₁ or S ₂ . When the oxide film is sufficiently thick (“d = 5 μm” for ordinary steel (FIGS. 3 and 4) and “d = 0.4 μm” for aluminum (FIGS. 5 and 6)), “Ku” for ordinary steel (T) = 0.789, ε (t) = 0.
789 ", for aluminum" Ku (t) = 0.27 "
2 to 0.2735, ε (t) = 0.272 to 0.273
5 ”, and these values are almost the same as the actual values.

On the other hand, in a transitional state, an alarm signal is output to give an alarm, and it operates as an improved two-color thermometer. When PbS and PbSe are used, “detection wavelength λ ₁ (λ ₃ ) = 2 μm, detection wavelength λ ₂ (λ ₄ ) = 4”
μm ”, and the change of the parameter Ku (t) becomes the change of Ku ₃₄ .

As described above, by determining whether or not it is in a steady state online, it is possible to perform measurement according to that state. In particular, since it operates as a monochromatic radiation thermometer in a steady state, the temperature is measured by a simple calculation and the operation becomes very fast. The measurement sensitivity is also improved. For reference, this point is explained below. The sensitivity n _i of the monochromatic radiation thermometer at the wavelength λ _i is obtained by partially differentiating the formula 11 (formula 20). On the other hand, in the case of two-color temperature measurement, it is approximately represented by Expression 21 (n
₁₂ is the difference between the detection wavelengths n ₁ and n ₂ of the radiation sensor ₂ ). As is clear from these equations, the sensitivity is better in the case of monochromatic temperature measurement.

[0073]

[Equation 20]

[0074]

[Equation 21]

In FIGS. 3 to 6, the installation angle is 0 degree, but if this angle is changed, the parameter Ku (t),
The emissivity ε (t) changes differently. Even in this case, these values are the brightness temperature S ₁ ,
It is similarly determined online using S ₂ and the detected wavelength value. Therefore, the temperature T near the surface of the object to be measured, the parameter Ku (t), and the emissivity ε (t) regardless of the installation state.
Can be measured. The same applies even if the material of the measured object changes.

Since good measurement is possible in this way, when this apparatus is used in an online process, the temperature controllability of the process is improved and the quality is greatly improved.

Next, a second embodiment of the present invention will be described.

In the above-described first embodiment, the emissivity power ratio itself is used as the parameter Ku (t) corresponding to the emissivity power ratio, but other mathematically equivalent parameters may be used. .. Table 1 shows, as an example, a modification of both sides (A in the table) of Equation 5 (B and C in the table).

[0079]

[Table 2]

Further, with respect to the temperature calculation block, mathematical equivalent conversion can be performed and can be obtained by the expression 22.

[0081]

[Equation 22]

In the second embodiment, the difference "1 / S ₁ -1 / S ₂ " of the reciprocal of the luminance signal is used as the parameter Ku (t),
This is a case where the temperature calculation block is based on Equation 22. Although it is almost the same as the above-described first embodiment, the configuration of the emissivity measuring unit 270 in the arithmetic unit 140 is slightly different, and FIG. 8 shows the arithmetic unit 1 of the second embodiment.
40 and the emissivity measuring unit 370 are shown (correlation block 290 is omitted). The differences are shown below.

The emissivity exponentiation ratio calculation block 310 obtains the parameter Ku (t) from the brightness temperatures S ₁ and S _{2 by} the calculation of “1 / S ₁ −1 / S ₂ ”, and the time average calculation block 23.
0 calculates the temporal average value <Ku (t)> of this parameter Ku (t) by the equation 8. The decision block decides whether the parameter Ku (t) is within a predetermined range. This range is calculated from the allowable temperature error ΔT by the above-mentioned equations 18 and 23.
It is decided based on.

[0084]

[Equation 23]

The emissivity calculation block 250 obtains the emissivity ε (t) from the brightness temperatures S ₁ and S ₂ by the expression 22, and the maximum / minimum value calculation block 240 uses the expressions 24a and 24b.
Maximum value of parameter Ku (t) sampled by
The maximum value ε _max (t) and the minimum value ε _min (t) of the emissivity ε (t) are calculated from the minimum value. Here, the emissivity calculation block 250 can also obtain the emissivity ε (t) from the parameter Ku (t) or the temporal average value <Ku (t)> in Equation 23.

[0086]

[Equation 24]

This device operates in the same manner as in the first embodiment described above except that the value of the parameter Ku (t) is different. However, the operation speed becomes faster because the exponential calculation is not performed.

Next, a third embodiment of the present invention will be described.

In the process line, there are measurement points that are suitable for a monochromatic radiation thermometer, and it is advantageous in terms of cost to install a monochromatic radiation thermometer at such a location. However, since the emissivity changes when the process material is changed, the emissivity is measured by the offline work as described above. A device for simplifying this work is a two-color radiation thermometer 13 as shown in FIG.
0, an emissivity measuring device including an arithmetic unit 440. The monochromatic radiation thermometer 120 is provided in the process line and has a wavelength λ ₀ radiated from the surface of the DUT 110.
It is an existing online radiation thermometer that detects the light of the above and outputs it as a brightness temperature S ₀ to a process controller (not shown).

The two-color radiation thermometer 130 is similar to that of the above-mentioned embodiment. The arithmetic unit 440 includes the emissivity measuring unit 270 or 370 of the above-described embodiment, and outputs the measured emissivity to a process controller (not shown). Here, for these detection wavelengths, "λ ₁ <λ
The relationship is ₀ <λ ₂ ”. The arithmetic unit 440 also has a storage unit (a disk device and its interface, etc.) for holding the measurement result, and holds the measurement result.

When measuring the emissivity, this device is temporarily installed near the monochromatic radiation thermometer 120 and connected to the process controller. At this time, the measurement angle of the monochromatic radiation thermometer 120 is made to match as much as possible. Then, the device is operated. The emissivity ε ₀ is calculated by the same operation as in the above-described embodiment, and is output to the process controller and held. Then, the temporarily installed device is removed. If there are other points to be measured, measure them in the same way. As described above, in the steady state, the wavelengths λ ₀ and λ
The emissivities of ₁ and λ ₂ are equal, and the measured emissivity is the emissivity of the monochromatic radiation thermometer 120. The process controller measures the temperature near the surface of the DUT 110 from the emissivity and the brightness temperature S _{0 of the} monochromatic radiation thermometer 120.

As described above, each time the material is changed, the emissivity of each of the monochromatic radiation thermometers provided in the process line is measured, and it is possible to respond to the change in the emissivity due to the difference in the material or the process online. It will be possible. The result of this measurement is stored in the process controller or disk, and a database of emissivity is created. When processing the same material next time, if the emissivity data is called from this database, process control becomes possible without re-measurement.

Since the database thus obtained contains data relating to measurement errors such as the upper and lower limits of the emissivity, a new radiation temperature measuring device can be constructed by using this database.
Even in the case of this radiation temperature measuring device, it is possible to obtain the measurement error of the measurement temperature, and the fact that the measurement error is obtained has important significance in the quality control of the process. Further, since the measurement temperature and the measurement error are obtained based on the correlation table using the database, the measurement can be performed at a very high speed. Therefore, it contributes to high-speed processing of the process and helps improve productivity. Hereinafter, this radiation temperature measuring device will be described as a fourth embodiment.

FIG. 10 shows a radiation temperature measuring apparatus according to the fourth embodiment of the present invention. This device has a two-color type radiation thermometer 130 similar to the above-mentioned embodiment and a calculation unit 470. In the calculation unit 470, the measured temperature and the emissivity are calculated from the correlation between the emissivity exponentiation ratio and the emissivity or emissivity ratio. The feature is that the measurement error is obtained. The processing is slightly different depending on the correlation with the emissivity and the correlation with the emissivity ratio, but first, as a result of the correlation with the emissivity,
Each part will be described.

The emissivity exponentiation ratio calculation block 410 is the same as in the above-described embodiment, and calculates the emissivity exponentiation ratio from the brightness temperatures S ₁ and S ₂ (shown again as equation 25; the same applies hereinafter). Also in this example, instead of the emissivity exponentiation ratio EPR itself, a parameter corresponding to the emissivity exponentiation ratio such as the above-mentioned parameter Ku (t) may be used. Here, the emissivity exponentiation ratio (or the corresponding parameter) is indicated by "EPR" for distinction.

[0096]

[Equation 25]

In the correlation data reference table 452, a regression function is obtained in advance from the correlation between the emissivity power ratio and the emissivity or the emissivity ratio, and these are stored. The regression function is a regression formula that is determined in advance based on the data actually measured on the DUT 110 offline, plots the measured values of the emissivity power ratio and the emissivity, and regresses these data groups. In the case of a monochromatic thermometer during actual measurement, the emissivity exponentiation ratio and the emissivity are correlated. The central regression function f, the upper limit regression function f _U , and the lower limit regression function f _L are the emissivity power ratio E
Correlation between PR and central value ε ₀ of emissivity, upper limit ε of emissivity
_It shows the correlation with _U and the correlation with the lower limit value ε _L of the emissivity, respectively, and the functions f, f _U , and f _L are expressed by Equation 26, Equation 27, and Equation 2
8 is shown.

[0098]

[Equation 26]

[0099]

[Equation 27]

[0100]

[Equation 28]

In the correlation data reference table 452, the ratios ε _U / ε ₀ and ε _L / ε ₀ of the central value to the upper limit value and the lower limit value are respectively set to the variation functions v _U , so that the temperature error can be easily obtained. It is defined as v _L , and the ratio of the center value to the upper limit value and the lower limit value is obtained from the emissivity exponentiation ratio (this point also applies to the emissivity ratio). Functions v _U , v
_L gives an estimated value necessary for obtaining the temperature error, and can be easily determined because the ratio of the measurement data may be taken.

The emissivity calculation block 450 uses the correlation data reference table 452 to measure the emissivity-to-power ratio EPR obtained in the emissivity-to-power ratio calculation block 410 from the actual measured value of the emissivity to power ratio corresponding to the central value ε ₀ and ratio ε of the emissivity. _U / ε ₀ , ε _L / ε ₀
To calculate. The temperature calculation block 460 determines the brightness temperature S ₁
And the measured temperature T ^* is obtained from the central value ε _{0 of the} emissivity. This calculation is similar to that of the monochromatic thermometer shown in Equation 1 above, but is shown in Equation 29 again.

[0103]

[Equation 29]

Further, the temperature error calculation block 462 determines the upper limit error ΔT _U and the lower limit error ΔT _U of the temperature error ΔT from the measured temperature T ^* and the ratios ε _U / ε ₀ and ε _L / ε ₀ based on the equation 30.
_{Find L.} Equation 30 is the upper limit value ε _U for Equation 1 above.
It is obtained by deleting the brightness temperature S from the equation for obtaining the upper limit temperature T _{U in} which the value of is substituted and the equation 29 of the measured temperature T ^* . The value of (ε ^* / ε) is calculated from (ε _U / ε ₀ ) to (ε _L / ε)
_{If it is changed to 0} ), the lower limit temperature T _L has the same relationship.

[0105]

[Equation 30]

Next, the functions f, f _U , f _L or the function v
_U and v _L will be described.

The dispersion functions v _U and v _L are expressed by the equations 31 and 32 (the functions f, f _U and f _L are represented by the equation 26).
.About.28 or any of formulas 34 to 36 described later).

[0108]

[Equation 31]

[0109]

[Equation 32]

If such functions v _U and v _L are determined, the emissivity or the emissivity ratio predicted value (regression estimated value by the offline functions f, f _U and f _L) with respect to an arbitrary emissivity exponentiation ratio EPR. ) Will give you an error margin. FIG. 11 shows such a state, the measured emissivity relative emissivity power ratio EPR ₀ or emissivity ratio f (EP
R ₀ ), its lower limit value f _U (EPR ₀ ), and its upper limit value f _L (E
It shows how PR ₀ ) is obtained. Function f,
f _U , f _L or the functions v _U , v _L are determined by the following procedure.

First, a predetermined object to be measured 1 is offline in advance.
The emissivity exponentiation ratio and the emissivity or the emissivity ratio for 10 are measured, and the measured values are plotted. A high-frequency portion of these data groups is regressed to obtain a function f.

The thus obtained "center value of emissivity ε _{0 with} respect to emissivity power ratio or center value of emissivity ratio (ε ₁ /
For the function f of ε ₂ ) ₀ ”, the central value ε ₀ or (ε ₁ /
Let ε ₂ ) ₀ be ± 15%, ± 10%, ± 5%, and other typical error values be the functions f _U and f _L. Here, the data case that becomes a rare case (3σ value) is excluded (plot A in FIG. 12). At this time, the functions v _U and v _L become constant values, and 1 ± 0.15, 1 ± 0.10, 1 ± 0.0
Takes a value such as 5. After confirming that the actual measurement data are contained in the area surrounded by the functions f _U and f _L , these are designated as functions v _U and v _L (FIG. 13).

In this case, as shown in FIG. 14, the measured data may vary in a certain range, and only the emissivity exponentiation ratio EPR in this range may greatly vary. This means that the emissivity fluctuates greatly when a specific surface condition is reached, and a large variation appears correspondingly. In this case as well, the functions v _U and v _L are shown, and the calculation error for adaptive emissivity fluctuation can be estimated. In this way, the correlation data reference table 452 holds the variation information.

Thus, the function v _U is the emissivity power ratio EP
The minimum value of the emissivity or emissivity ratio for R is given based on the offline data and takes a value of 1.0 or less. Further, the function v _L is given a maximum value of the emissivity or the emissivity ratio with respect to the emissivity exponentiation ratio EPR based on the offline data, and takes a value of 1.0 or more. For example, if the error is ± 10%, the functions v _U and v _L are as in Expression 33.

[0115]

[Expression 33]

Next, the operation of this device will be described.

The emissivity exponentiation ratio calculation block 410 obtains the emissivity exponentiation ratio EPR from the signals of the two-color emission thermometer 130, that is, the brightness temperatures S ₁ and S ₂ . This emissivity exponentiation ratio E
The PR is converted into the central value ε ₀ of the emissivity corresponding to the emissivity exponentiation ratio EPR using the correlation data reference table 452 in the emissivity calculation block 450. In addition, the ratio (ε _U / ε
₀ ) or (ε _L / ε ₀ ) is obtained. At this time, if the error has a typical value, the value is constant. In the case of FIG. 14, the ratio (ε _U / ε ₀ ) or (ε _L / ε ₀ ) according to the error range can be obtained.

In the temperature calculation block 460, the measured temperature T ^* is obtained from the central value ε _{0 of the} emissivity, and in the temperature error calculation block 462, the measured temperature T ^* and the ratios ε _U / ε ₀ , ε.
_The upper limit error ΔT _U and the lower limit error ΔT _L can be obtained from _L / ε ₀ . At this time, the ratios ε _U / ε ₀ and ε _L / ε ₀ are already given, and the error is obtained based on the equation 30 using this, so that this calculation is very simple.

In this way, the measured temperature of the surface of the object to be measured can be obtained, and the upper limit error ΔT _U and the lower limit error ΔT _L can be obtained at the same time. Therefore, it is possible to know how much the measured temperature T ^* has an error, and it is possible to effectively manage the temperature of the process. As a result, the quality control of the process can be favorably performed, which can be useful for improving productivity. Particularly, since the average value is not taken as in the above-mentioned embodiment, the response is very good.

When the object 110 to be measured is measured off-line in advance, if the two-color thermometer is used, the emissivity ratio (ε
₁ / ε ₂ ). The central regression function f, the upper limit regression function f _U , and the lower limit regression function f _L are obtained in advance and stored. These functions are the emissivity power ratio EP
Correlation between R and the central value of emissivity ratio (ε ₁ / ε ₂ ) ₀ , upper emissivity ratio (ε ₁ / ε ₂ ) _U , and lower emissivity ratio (ε ₁ / ε ₂ ) _L , respectively. , Equation 34, Equation 35, and Equation 36.

[0121]

[Equation 34]

[0122]

[Equation 35]

[0123]

[Equation 36]

In the correlation data reference table 452, similarly to the above case, the variation functions v _U and v _L are defined so that the temperature error can be easily obtained, and the ratio of the central value to the upper limit value “(ε ₁ / ε ₂ ) _U / (ε ₁ / ε ₂ ) ₀ ”, the ratio of the central value to the lower limit value“ (ε ₁ / ε ₂ ) _L / (ε ₁ / ε ₂ ) ₀ ”
Trying to get.

FIG. 15 shows an example of the functions f, f _U , and f _L , and the plots in the figure are those measured in advance offline with a predetermined material (Kawasaki Steel low carbon steel CAL material). .. This is an example of a case in which heat pattern changes, material changes, and surface state changes that cover all emissivity fluctuations that occur stochastically in the online process are assumed.
It covers all emissivity and emissivity ratio variations within a reasonable error range. The functions f _U and f _L indicate the boundaries between the upper limit and the lower limit, respectively, of the measured data obtained by removing the rare case (3σ value).

[0126] Emissivity calculation block 450, the center value of the emissivity ratio corresponding thereto from the measured value of emissivity power ratio EPR obtained in emissivity power ratio calculation block 410 using the correlation data reference table 452 (epsilon ₁ / Calculate ε ₂ ) ₀ , upper limit emissivity ratio (ε ₁ / ε ₂ ) _U , and lower limit emissivity ratio (ε ₁ / ε ₂ ) _L. The temperature calculation block 460 obtains the measured temperature T ^* from the central value (ε ₁ / ε ₂ ) ₀ of the emissivity ratio. This calculation is the same as that of the two-color type thermometer described above, but it is shown in Equation 37 again.

[0127]

[Equation 37]

In addition, the temperature error calculation block includes the brightness temperatures S ₁ and S ₂ and the center value (ε ₁ / ε ₂ ) _{0 of} the emissivity ratio, the upper limit emissivity ratio (ε ₁ / ε ₂ ) _U , and the lower limit emissivity. Ratio (ε ₁ / ε
₂ ) From _L, the upper limit error Δ of the temperature error ΔT is calculated based on the equation 38.
Obtain T _U and lower limit error ΔT _L. Expression 38 is the above Expression 3
4 is obtained by eliminating the brightness temperature S from the formula for obtaining the upper limit temperature T _{U in} which the upper limit value ε _U is substituted for and the formula 34 of the measured temperature T ^* (“λ ^* ” in Formula 38 is 2). Shows the apparent measurement wavelength in wavelength type temperature measurement). "(Ε ^* /
The value of (ε) / (ε ₁ / ε ₂ ) ”is set to“ (ε ₁ / ε ₂ ) _U /
(Ε ₁ / ε ₂ ) ₀ ”to“ (ε ₁ / ε ₂ ) _L / (ε ₁ /
ε ₂ ) ₀ ”, the lower limit temperature T _L has the same relationship.

[0129]

[Equation 38]

Also in this case, the measurement temperature T ^* , the upper limit error ΔT _U , and the lower limit error ΔT _L can be obtained in the same manner.

In the above embodiment, (ε ^* /
ε) or “(ε ^* / ε) / (ε ₁ / ε ₂ )” in Equation 38
Although the value of is directly obtained, the central value ε ₀ , the upper limit ε _U , the lower limit ε _L (or (ε ₁ / ε ₂ ) ₀ , (ε ₁
/ Ε ₂ ) _U and (ε ₁ / ε ₂ ) _L ) may all be obtained, or the measured temperature T ^* or the like may be obtained from any two.

FIG. 16 shows the upper limit value ε _U and the lower limit value ε of the emissivity.
_This is an example of a case where the measured temperature T ^{* and the} like are obtained from _L. In this device, the correlation data reference table 452
In addition, the upper limit value ε _U and the lower limit value ε _L of the emissivity are obtained from the actually measured values of the emissivity exponentiation ratio EPR. The emissivity median value calculation block 455 averages the upper limit value ε _U and the lower limit value ε _L to obtain the center value ε ₀ . From the center value ε _0, the measured temperature T ^* obtained by the temperature calculation block 460 and the upper limit ε _U
And from the measured temperature T ^* Minimum value calculation block 462a of determining the minimum value T ^* min of the measured temperature by a formula 30 described above, the measured temperature T ^* and from the lower limit value epsilon _L, the measurement temperature T ^* of the maximum value The calculation block 462b obtains the minimum value T ^* max of the measured temperature by the above-mentioned formula 30.

The emissivity exponentiation ratio calculation block 410 obtains the emissivity exponentiation ratio EPR from the brightness temperatures S ₁ and S ₂ , and the emissivity exponentiation ratio EPR is upper limit in the emissivity calculation block 450 using the correlation data reference table 452. The value ε _U and the lower limit ε _L are converted. Using these values, the measured temperature T ^* ,
The minimum value T ^* min and the minimum value T ^* max of this are determined.

FIG. 17 shows the upper limit error ΔT _U and the lower limit error ΔT
_L to obtain the ratio ε _U / ε ₀ , ε _L / ε ₀
(Operation block 464), the upper limit error ΔT _U and the lower limit error ΔT _L are obtained from this in the same manner as in the embodiment of FIG.

The present invention is not limited to the above-mentioned embodiments, but various modifications can be made.

For example, although the improved two-color thermometer is operated in a transient state, the improved two-color thermometer can be replaced by a cavity method, a TRACE thermometer, and other emissivity fluctuations. It may be configured with a sensor. The material of the object to be measured is not limited to steel plate or aluminum, and may be other material such as silicon.

It is also possible to switch the detection wavelength of the two-color radiation thermometer. That is, a plurality of radiation sensors (for example, Si, Ge, PbS, PbSe) may be provided, and two of them may be selected and output as the brightness temperatures S ₁ and S ₂ . At this time, the parameter Ku
The curve of (t) may change, and a steady state may be easily obtained. Furthermore, regarding the emissivity, the parameter Ku
Similarly, the parameter corresponding to this may be used.

The radiation temperature measuring device of the present invention calculates the temperature in the vicinity of the surface of the object to be measured based on the correlation between the parameter measured in advance on the object to be measured and the emissivity of the object to be measured. If it is, the temperature near the surface etc. can be obtained from the detection signal based on the correlation,
The calculation is very fast, and the temperature near the surface can be quickly obtained. Further, not only the temperature in the vicinity of the surface but also the upper limit value and the lower limit value of the temperature error or the error range can be quickly obtained.

[0139]

As described above, according to the radiation temperature measuring apparatus of the present invention, it is possible to detect whether or not the surface of the object to be measured is in a steady state. In this state, the temperature near the surface of the object to be measured can be obtained from the detection signal of the two-color thermometer by a simple calculation process, so that the temperature near the surface of the object to be measured can be detected quickly and with high sensitivity. It will be possible. Further, according to the emissivity measuring apparatus of the present invention, it is possible to detect whether or not the surface of the object to be measured is in a steady state.

According to the emissivity measuring apparatus of the present invention, it is possible to easily measure the change in emissivity. Further, since the variation of the emissivity of the material which is measured in advance is stored in the apparatus, it is possible to perform online prediction of the variation itself and online prediction of the temperature variation due to the variation. It becomes possible to apply process control to cope with such variations, and preventive measures due to out-of-temperature measurement, adaptive control, and extraction of caution parts due to out-of-temperature control of the product after product manufacture become possible.

[Brief description of drawings]

FIG. 1 is a configuration diagram of an embodiment.

FIG. 2 is a configuration diagram of a calculation block.

FIG. 3 is a diagram showing changes in emissivity.

FIG. 4 is a diagram showing a change in emissivity exponentiation ratio.

FIG. 5 is a diagram showing changes in emissivity.

FIG. 6 is a diagram showing a change in emissivity exponentiation ratio.

FIG. 7 is a chart diagram for determining a range.

FIG. 8 is a configuration diagram of a calculation block according to a second embodiment.

FIG. 9 is a configuration diagram of a third embodiment.

FIG. 10 is a configuration diagram of a fourth embodiment.

FIG. 11 is a diagram showing a correlation of an emissivity or an emissivity ratio, its lower limit value, and its upper limit value with respect to a measured emissivity power ratio.

FIG. 12 is a diagram showing a plot of a rare case.

FIG. 13 is a diagram showing an aspect of the functions f _U and f _L when the functions v _U and v _L have predetermined values.

FIG. 14 shows the functions f _U and f when a large variation is given.
_The figure which shows the aspect of _L.

FIG. 15 is a diagram showing an actual measurement example of the correlation between the emissivity exponentiation ratio and the emissivity ratio.

FIG. 16 is a configuration diagram of a fourth embodiment.

FIG. 17 is a configuration diagram of a fifth embodiment.

FIG. 18 is a configuration diagram of a conventional example.

[Explanation of symbols]

110 ... Object to be measured, 120 ... Monochromatic radiation thermometer, 130
… Two-color radiation thermometer, 140… Computing unit, 210…
Emissivity power ratio calculation block, 220 ... Judgment block, 2
30 ... Time average calculation block, 240 ... Maximum value / minimum value calculation block, 250 ... Emissivity calculation block, 260 ...
Temperature calculation block, 270 ... Emissivity measurement unit.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Shinichi Takechi Inventor, 1 Kawasaki-cho, Chuo-ku, Chiba, Chiba Prefecture Technical Research Division, Kawasaki Steel Co., Ltd. (72) Ichiro Maeda 1 Kawasaki-cho, Chuo-ku, Chiba, Chiba Prefecture Kawasaki Steel Corporation Technical Research Division

Claims (9)

[Claims] 1. A radiation thermometer for detecting light emitted from the surface of an object to be measured at at least two different wavelengths and outputting it as a detection signal, and a radiation thermometer corresponding to the emissivity power ratio from the detection signals of at least two wavelengths. The parameter is calculated, the temporal average value of the parameter is obtained, it is determined whether the parameter is within a predetermined range with respect to the temporal average value, and when the parameter is within the predetermined range, the parameter is determined from the parameter. A radiation temperature measuring device, comprising: an arithmetic means for calculating the temperature near the surface of the object to be measured from the detection signal based on the emissivity or the emissivity ratio of the surface of the object to be measured. 2. The calculation means calculates the temperature near the surface of the object to be measured from the detection signal of at least one wavelength or from the detection signal of at least two wavelengths. Radiation temperature measuring device. 3. The radiation temperature according to claim 1, wherein the calculation means further calculates an upper limit value and a lower limit value of the temperature near the surface of the object to be measured from the maximum value and the minimum value of the parameter. measuring device. 4. The radiation temperature measuring device according to claim 1, wherein when the parameter is outside the predetermined range, an alarm signal indicating that the emissivity is unstable is output to the outside. 5. A radiation thermometer for detecting light emitted from the surface of the object to be measured at at least two different wavelengths and outputting it as a detection signal, and a radiation thermometer corresponding to the emissivity power ratio from the detection signals of at least two wavelengths. The parameter is calculated, the temporal average value of the parameter is calculated, it is determined whether the parameter is within a predetermined range with respect to the temporal average value, and when the parameter is within the predetermined range, the parameter or at least An emissivity measuring apparatus comprising: an emissivity or an emissivity ratio of light emitted from the surface of the object to be measured from the detection signals of two wavelengths. 6. A radiation thermometer for detecting light radiated from the surface of the object to be measured at at least two different wavelengths and outputting it as a detection signal, and a radiation thermometer corresponding to the emissivity power ratio from the detection signals of at least two wavelengths. A parameter is calculated, and based on the correlation between the parameter previously measured for the object to be measured and the emissivity or emissivity ratio of at least one wavelength of the object to be measured, the vicinity of the surface of the object to be measured from the detection signal. Radiation temperature measuring device, comprising: an arithmetic means for calculating an upper limit value and a lower limit value or an error range of the temperature. 7. The radiation according to claim 6, wherein the correlation is a correlation between the parameter and any one of a center value, an upper limit value, and a lower limit value of the emissivity or emissivity ratio. Temperature measuring device. 8. The radiation temperature measuring device according to claim 6, wherein the correlation is determined in advance based on data measured off-line in advance for the object to be measured. 9. The emissivity exponentiation ratio and the emissivity or emissivity ratio of a predetermined object to be measured is measured off-line in advance, and the emissivity or the emissivity or the emission is regressed by returning to a frequently used part of the measured data group. When the regression function of the central value of the rate ratio is determined, and the data group is in a predetermined error range from the central value of the emissivity or the emissivity ratio by the regression function, the upper and lower limits of the error range are the upper and lower limits. The method of determining the correlation between the emissivity power ratio and the emissivity or the emissivity ratio of the DUT as a regression function of.