JPH0515201B2 - - Google Patents

Description

[Detailed description of the invention]

[Industrial Application Field] The present invention relates to a method of calibrating an infrared thickness meter in advance according to the type of object to be measured, and particularly relates to a method of processing a detection signal in this case. [Prior Art] In general, an infrared thickness gage places an object to be measured between a light source and a photodetector, and when the light from the light source passes through the object, it is attenuated by absorption and scattering. The thickness of the object to be measured is measured by utilizing the fact that this amount of attenuation is a function of the thickness of the object to be measured. In this case, since the amount of attenuation of the transmitted light described above varies depending on the characteristics of the object to be measured, the instrument is calibrated prior to measurement. This calibration is performed by setting a reference line, that is, a calibration curve, determined by the characteristics of the object to be measured. This calibration curve L (third
(see figure) has a known thickness, as will be explained in detail later.
It is set by measuring the amount of transmitted light Ia and Ib of samples Ta and Tb of the object to be measured, which have ta and tb, and is a semi-logarithmic graph, that is, the horizontal axis is the thickness t, and the vertical axis is the logarithm value of the detection signal I. It is displayed as a straight line on the Totsuta graph (the graph shown in FIG. 3). Then, once this calibration curve L is set, the object to be measured
By measuring the amount of transmitted light Ii of Ti,
The thickness ti′ of Ti is automatically measured. this thickness
ti′ is usually displayed on a CRT included in the thickness gauge or recorded and output from a printer. By the way, the function, or physical law, regarding the attenuation of the amount of transmitted light and the thickness of the object to be measured is conventionally expressed by the following equation (5): I = I ₀ e ^- 〓 ^t ... (5) where I = the object to be measured. Transmitted light amount detection signal value of the object to be measured _I0 =light amount detection signal value when the optical path is open μ=absorption coefficient of the object to be measured t=thickness of the object to be measured was used. That is, the amount of attenuation of the amount of transmitted light has been determined only by the absorption coefficient μ of the object to be measured. However, in the measurement method based on the above equation (5), an error occurs in the measured thickness of the object to be measured, for example, for objects that reflect a considerable amount of light, such as plastic sheets or films. Therefore, the applicants adopted the following equation (2) I=I ₀ (1-r)e ^- 〓 ^t ...(2) in place of the above equation (5) as the above-mentioned physical law. The formula is the same as formula (2) shown in claim 2], and based on this, a measurement method for measuring the thickness of the object to be measured (hereinafter referred to as the first measurement method) was developed, A patent application was filed (Japanese Unexamined Patent Publication No. 176508/1983). According to this, the influence of reflected light, which conventionally causes errors, is removed, so accurate measurement can be performed even on objects to be measured with high reflectance. Here, the setting of the above-mentioned calibration curve L will be explained in relation to the instrument calibration method in this first measurement method. The calibration curve L is based on two samples Ta and Tb.
It is set by measuring , that is, by calculating the above-mentioned physical law equation (2). That is, in the measured value Ia=I ₀ (1-r)e ^- 〓 ^ta Ib=I ₀ (1-r) e ^- 〓 ^tb , the detection signals Ia, Ib, and I ₀ are the values obtained by measurement, respectively. The thicknesses ta and tb are known values, and the two unknown values are the extinction coefficient μ and the reflectance r. Therefore, from the above two measured values, the extinction coefficient μ and the reflectance r can be calculated. is calculated.
That is, the calibration curve L is set. This calibration curve L is obtained by logarithmically converting the above equation (2) using the following equation log _e I = log _e I ₀ (1-r) ^- 〓 ^t , where the extinction coefficient μ and reflectance r are was assumed to be constant, so everything except the thickness t is a constant. Therefore, the logarithm of the detection signal I and the thickness t are linear equations, and are displayed as a straight line on the graph (Figure 3). Ru. Then, the measured thickness of the object to be measured Ti is measured using such a calibration curve L.
Since ti' includes the factor of reflectance r in the calibration curve L, conventional measurement errors based on reflected light are removed. However, even in the first measurement method, errors occur in the measured thickness of the object to be measured, especially when the measured thickness range of the object to be measured is wide. As a result of extensive research, the present applicants found that this error is due to the fact that the extinction coefficient of the object to be measured is not constant with respect to the thickness of the object. We then developed a measurement method (hereinafter referred to as the second measurement method) that could eliminate measurement errors based on this by setting multiple calibration curves, and filed a patent application (patent application). 62-65766). Next, this second measurement method will be briefly explained below. Note that, as in the first measurement method, the physical law defined by equation (2) is used in this measurement method. In this measurement method, when calibrating the instrument, that is, when setting the calibration curve, an appropriate number of samples Ta, Tb, and Tc with known thicknesses are selected depending on the thickness range of the object to be measured.
Tg is selected. Then, after measuring and calculating each of these samples with a thickness gauge,
As shown in Figure 4, the CRT installed in the thickness gauge
Graphically displayed above. In this case, if the sample thickness range is large, the sample Tc
~Tg are not located on the calibration curve L, and they define a downwardly convex curve M with respect to the calibration curve L. This means that in the physical law equation (2), the reflectance r does not change, so the extinction coefficient μ changes with respect to the thickness t, and this fluctuation gradually decreases as the thickness t increases. That's what it means. Therefore, in this measurement method, the calibration curve is replaced with the line segment L and the samples Ta, Td; Te, Tf; Tg,
Set multiple line segments Na, Nb, and Nc passing through Tb. When setting this calibration curve Na, Nb, Nc,
Observe the graphic display and select samples that define each of the calibration curves Na, Nb, and Nc so as to approximate curve M. Therefore, according to this measurement method, the measured object
By measuring the amount of transmitted light Ii of Ti, the calibration curve Ma is calculated.
Even if the thickness ti'' of the workpiece Ti measured via the curve M has an error with respect to the true thickness ti that should be measured via the curve M, the error Δti'' is suppressed to a very small value. Ru. This error Δti'' is significantly improved compared to the error Δti′ generated by the first measurement method also shown in FIG. 4. Note that the error due to reflected light is eliminated. Of course. As described above, according to the present measurement method, even if the thickness of the object to be measured is wide, the thickness can be measured with high accuracy. [Problems to be solved by the invention] In this way, according to the second measurement method described above,
Even objects to be measured with a wide thickness range can be measured with high accuracy. However, such a measurement method has the disadvantage that calibration of the instrument requires relatively complicated labor and judgment. In other words, when setting the calibration curves Na, Nb, and Nc, measure and calculate a large number of samples, display them graphically, observe this graph, set the required number of calibration curves, and then This requires work such as determining each boundary value of these calibration curves. Furthermore, judgment is required to set each calibration curve. For example, as shown in Figure 4, the calibration curve
Calibration curve determined by samples Ta and Tc instead of Na
By setting Na′, the measurement error increases from Δti″ to Δti. Therefore, an object of the present invention is to simplify and easily calibrate the instrument, especially for objects to be measured with a wide thickness range. An object of the present invention is to provide a detection signal processing method for an infrared thickness gauge that can be performed with high accuracy. The signal processing method sets the rate of decrease of the transmitted light amount detection signal with respect to the thickness of the measured object based on the extinction coefficient of the measured object as a function of the magnitude of the transmitted light amount detection signal, and calculates the calibration by calculating this function. In the detection signal processing method of an infrared thickness gauge by setting a line, the above function is expressed by the following formula (1) dx/dt=-αx ⁿ ...(1) where x=lnI I=transmitted light amount detection signal value t=thickness of the object to be measured α=parameter n=power number, and the transmitted light amount detection signal value I in equation (1) is
The following formula (2) I=I ₀ (1-r)e ^- 〓 ^t ...(2) Here, when I ₀ = thickness of the object to be measured t is 0,
In other words, the light intensity detection signal value when the optical path is open r = reflectance of the object to be measured μ = extinction coefficient of the object to be measured. ,... is obtained by integrating the above equation (1) using the following equation (3) t=-1/(1-n)α{x ^(1-n) −x ₀ ^(1-n) }... (3) Here, when x ₀ = measured object thickness t is 0,
In other words, the logarithmic value of the light intensity detection signal when the optical path is open. However, n≠1, and the following equation (4) applied when n=1: t=-1/α(ln x - ln x ₀ )... (4) Calculated thickness ta′, tb′,
... is measured, and the value of the power number n that maximizes the value of the correlation coefficient m between the calculated thickness and the true thickness is selected by successive least squares calculations. [Function] The setting of a calibration curve in instrument calibration is achieved by sequentially processing a predetermined function by a computer, and is therefore easily and quickly performed. Moreover, since the function includes the variation of the extinction coefficient with respect to the thickness of the measured object and each factor of the reflectance, it can be used even for a measured object with a wide thickness range and high reflectance.
Accurate measurements are taken. [Example] Next, an example of the detection signal processing method for an infrared thickness gauge according to the present invention will be described in detail with reference to the accompanying drawings. First, prior to explaining the processing method, the configuration of the infrared thickness gauge according to the present invention will be briefly explained. In FIG. 2, the infrared thickness gage consists of a measuring section 10 and an operator console 12.
is equipped with an infrared light source 14, a holder 18 for holding an object to be measured or a sample 16 of known thickness, and a transmitted light detector 20 having a photoelectric conversion element.
The operator console 12 has an A/D converter 2
2, a CPU 24, an internal storage device 26, an operation keyboard 28, a CRT 30, a printer 32, and an external storage device 34 are provided, and the measurement unit 10 and operator console 12 are connected via a conversion amplifier circuit 36, and the CPU 24 has a A remote controller 22 is provided. When measuring the object to be measured, projection light La is projected from the infrared light source 14, and transmitted light Lc transmitted through the object to be measured 16 is
It is detected by a transmitted light detector 20 and converted into a detection signal I. Next, this detection signal I is logarithmically converted into a detection signal V by a conversion amplifier circuit 36, and the obtained detection signal V is sent to the object under test 16 by an A/D converter 22.
is converted into a signal W indicating the thickness of the CRT3 as appropriate.
0 or stored and output from the printer 32. In this case, the thickness gauge is operated via the keyboard 28 or remotely via the operating device 38. Next, a method for calibrating an infrared thickness gauge according to the present invention will be explained with reference to FIG. When calibrating, first open the optical path and proceed to step S1.
Detection signal I ₀ is detected and input and stored in internal storage device 26 . Next, samples Ta, Tb, . Next, in step S3, a calibration curve as shown in FIG. 4 is graphically displayed on the CRT 30 using the stored data. In this case, if the calibration curve is close to the line segment L, you can immediately start the step as shown by the dashed line.
The process proceeds to S8, and if it deviates from the line segment L, the process proceeds to a correction or linearization process. In the linearization process, the above equation (3) and t=-1/(1-n)α{x ^(1-n) −x ₀ ^(1-n) } ...(3) When n=1 The calculation process proceeds based on the applied formula (4) t=-1/α(ln x-ln x ₀ )...(4), but the power number n
The calculated thickness values ta′, tb′,
... and the true thicknesses ta, tb, ... and the correlation coefficient m is selected to be the maximum value. In this linearization process, the power number n is calculated sequentially as n=0, 1, 2,
The first step is to input integer values sequentially starting from 0 and select the value of n' that maximizes the above m, and the selected value n' is selected in a range of ±0.5 in steps of 0.1. Input the numerical values and find n'± where the above m becomes the maximum value.
The second step is to select the value of Δn'. That is, in the first step of linearization, n values 0, 1, 2, . . . are sequentially input in step S4, and the corresponding m values are calculated and compared. Table 1 shows the calculation results in this step for a certain sample and shows that m has a maximum value of 0.99992 at n=4.
Next, in step S6 of the second step of linearization, the values of n' 3.6, 3.7, ... are sequentially input.
Their corresponding values of m are calculated and compared.
Table 2 shows the calculation results and shows that m has a maximum value of 0.99997 when n=4.2.
Note that the value of the correlation coefficient m never exceeds 1 due to its nature. Furthermore, these linearizing steps are automatically performed by a computer and are usually completed in about 10 seconds.

【table】

〔Effect of the invention〕

As explained above, the detection signal processing method for an infrared thickness meter according to the present invention is capable of converting the rate of decrease in the transmitted light amount detection signal with respect to the thickness of the object to be measured based on the variation in the extinction coefficient of the object to be measured into the transmitted light amount detection signal. The calibration curve is set as a function related to the magnitude of , and the calibration curve can be set by arithmetic processing of this function.By performing this arithmetic processing on a computer, the calibration of the instrument can be easily and easily made without any judgment. This can be accomplished without any effort.
Moreover, the calibration error can be suppressed to a small level. Therefore, the thickness of an object to be measured whose thickness range is wide and whose extinction coefficient fluctuates can be easily and accurately measured. Furthermore, since the function includes a factor related to the reflectance of the object to be measured, the thickness of the object can be accurately measured even if the object has a relatively high reflectance.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing an embodiment of the instrument calibration method in the detection signal processing method for an infrared thickness gauge according to the present invention, and FIG. 2 is a block diagram showing the configuration of the infrared thickness gauge according to the present invention. Figure 3 is a graph explaining a calibration curve set during instrument calibration of a conventional infrared thickness gauge, and Figure 4 is a graph explaining a calibration curve set during instrument calibration of an infrared thickness gauge, which is the basis of the present invention. This is a graph explaining. 10...Measurement section, 12...Operator console,
14... Infrared light source, 16... Measured object or its sample, 18... Holder, 20... Transmitted light detector, 2
2...A/D converter, 24...CPU, 26...internal storage device, 28...operation keyboard, 30...CRT,
32...Printer, 34...External storage device, 36...Conversion amplifier circuit, 38...Remote controller, T...Sample or object to be measured, I...Transmitted light amount detection signal, t...True thickness of sample or object to be measured, Δt...measurement error,
t′, t″…calculated or measured thickness of the sample or object to be measured.

Claims (1)

[Claims] 1. Setting the rate of decrease of the transmitted light amount detection signal with respect to the thickness of the measured object based on the extinction coefficient of the measured object as a function of the magnitude of the transmitted light amount detection signal, and arithmetic processing of this function. In the detection signal processing ^method for an infrared thickness gauge in which a calibration curve is set by Signal value t = Thickness of the object to be measured α = Parameter n = Power number, and the transmitted light amount detection signal value I in equation (1) is
The following formula (2) I=I ₀ (1-r)e ^- 〓 ^t ...(2) Here, when I ₀ = thickness of the object to be measured t is 0,
In other words, the light intensity detection signal value when the optical path is open r = reflectance of the object to be measured μ = extinction coefficient of the object to be measured; ,... is obtained by integrating the above equation (1) using the following equation (3) t=-1/(1-n)α{x ^(1-n) −x ₀ ^(1-n) }... (3) Here, when x ₀ = measured object thickness t is 0,
In other words, the logarithm of the light intensity detection signal when the optical path is open. However, n≠1, and the following equation (4) applied when n=1: t=-1/α(ln x - ln x ₀ )... (4) Calculated thickness ta′, tb′,
... and selects the value of the power number n that maximizes the value of the correlation coefficient m between the calculated thickness and the true thickness by successive least squares calculations. detection signal processing method.