WO2003087885A2

WO2003087885A2 - Apparatus and method for true temperature estimation

Info

Publication number: WO2003087885A2
Application number: PCT/IL2003/000302
Authority: WO
Inventors: Yvgeni Glazman
Original assignee: D.G.Y. Infrared Ltd.
Priority date: 2002-04-10
Filing date: 2003-04-10
Publication date: 2003-10-23
Also published as: AU2003225518A1; WO2003087885A3; AU2003225518A8

Abstract

A method of optical pyrometery, including: making at least one measurement of a brightness temperature of an object (104), generating at least one model of the temperature of the object, testing the model under a plurality of different conditions (108), and accepting the model if it performs in an expected manner under said different conditions (122).

Description

APPARATUS AND METHOD FOR TRUE TEMPERATURE ESTIMATION

RELATED APPLICATION The present application claims the benefit under 35 USC 119(e) of US provisional application filed on April 10, 2002 by the same inventor and having serial number 60/371,088, the disclosure of which is incorporated herein by reference.

FIELD OF THE INVENTION The field of the invention is optical pyrometry, for example passive optical pyrometry for process monitoring.

BACKGROUND OF THE INVENTION In many manufacturing processes, control of temperature to within a narrow range is critical for the quality of the product, and/or the efficiency or even the safety of the process.

Typically, in metallurgy, and in glass and ceramic manufacture, it is desirable to control or know the temperature to within ±1% or ±2%. During manufacture of semiconductor wafers with Rapid Thermal Control, the temperature should be controlled to within ±0.5% or ±0.25%. Precise control of temperature requires accurate measurement of temperature.

Temperature is measured accurately by contact thermometry. But in many manufacturing processes, contact thermometry is not practical, for example because the process involves highly corrosive materials, or temperatures that are too high for contact thermometry, or moving bodies that are likely to damage or be damaged by a temperature sensor that they come into direct contact with.

For this reason, non-contact optical methods are often used to measure temperature.

These methods are based on Planck's law for blackbody radiation, which gives the emission intensity of a perfectly black body as a function of temperature and wavelength. Planck's law states that for a black body of temperature T, the radiation intensity / emitted per wavelength λ per steradian is

where c\=2πc^h, C2=^:ch/kB, c is the speed of light, kg is Boltzmann's constant, and h is

Planck's constant. These constants may depend on the units used.

Real bodies are not perfectly black, but emit less radiation at any given wavelength and temperature than Planck's blackbody law predicts. The ratio of actual emission to the blackbody emission is called the emissivity, and in general it depends on wavelength and temperature, as well as on the composition of the body and its surface characteristics. If the emissivity of a body is known, then its temperature can be determined by measuring the intensity of radiation it emits at any given wavelength. If, in a particular manufacturing process, the emissivity is known for the bodies whose temperature is being measured, and it does not change, then accurate temperature measurements are possible using passive brightness pyrometry at a single wavelength. If the emissivity is not known, and varies in a unpredictable way, then temperature can be determined if the emissivity is measured in some way. In active optical pyrometry, light of a known wavelength and intensity is reflected from an object, and the reflectivity of the body is determined by measuring the reflected light intensity. The emissivity of an opaque body is equal to 1 minus the reflectivity, at the same wavelength and temperature. Active optical pyrometry has the disadvantage that the measured reflected light intensity may depend not just on the reflectivity of the body, but also on the geometry and surface smoothness of the body (specular vs. diffusive reflection), and on unintended changes in the intensity of the light aimed at the body. If these other factors vary in an unknown way from body to body on an assembly line, for example, then it may be difficult to make accurate measurements of the emissivity. In practice, active optical pyrometry is used mostly at relatively low temperatures, since at higher temperatures, where the emitted radiation is largely in the visible range for example, the emitted light intensity is so great that it may be difficult to detect reflected light from an external light source.

Among passive methods of optical pyrometry, emission may be measured at two or more different wavelengths. If there is a known functional relationship between the emissivity at a given wavelength, and the ratio of that emissivity to the emissivity at a different wavelength, and if this relationship holds, or varies in a known way, over the temperature range of interest, then multi-wavelength methods can be used to determine emissivity. Although multi-wavelength passive optical pyrometry can be used somewhat successfully, it has the disadvantage that it is generally necessary to calibrate the apparatus for each new type of material and temperature range. Other passive optical pyrometry methods involve increasing the effective emissivity of a body by reflecting back to the body some of the radiation it emits. If almost all of the emitted radiation is reflected back, for example by almost completely surrounding the body by a cavity, then the effective emissivity will approach 1, and the temperature may be determined directly by Planck's blackbody law. Even if only some of the emitted radiation is reflected back, information about emissivity of the body may be obtained by comparing the radiation with and without reflecting it back, if the geometry of the body and the reflector, and the reflectivity (specular and diffusive) of the reflector, are all known. For example, refractory plates, or hemispheres, or other shapes, may be used. As with active pyrometry, these methods may be difficult to use if the geometry and surface characteristics of the bodies vary unpredictably. In addition, it may be difficult to bring such reflectors close enough to the body to reflect back a significant fraction of the emitted radiation, if the body being measured is highly corrosive, or very hot, or moving and in danger of accidentally hitting the reflector. Some formulas for brightness temperature are presented here, as background for the detailed description of the invention. The brightness temperature of a body is the temperature it would need to have, if its emissivity were equal to 1, in order to produce the same emission intensity, at a given wavelength, as it actually produces, with emissivity less than 1, at its true temperature. Thus brightness temperature is a function of wavelength, and is always less than true temperature.

For a given optical pyrometer sensitive to a given range of wavelengths, narrow enough so that the right hand side of Eq. (1) and the emissivity are nearly constant over that range, the signal u generated by the pyrometer is proportional to 7, so we may write

^' where ε(λ,T) is the emissivity, K is a aperture constant, dependent on wavelength, on the sensitivity of the pyrometer at that wavelength, and on the bandwidth of wavelengths. K may be found by calibrating the pyrometer with a body of known emissivity (for example a blackbody cavity) and known temperature (determined, for example, by contact thermometry). Then the temperature T of a real body is related to the signal u generated by the pyrometer by = ^2 (3) λ\og(κ/u -l) where c₂ = hc/kβ = 14388 micro-kelvin-meters. The brightness temperature T_b is given by the same expression, and, taking emissivity into account,

When ε = 1, of course, 7^ = T.

SUMMARY OF THE INVENTION

An aspect of some embodiments of the invention relates to a method of pyrometry in which information about an object is estimated by generating a model of the information and checking the model under various conditions to see if it behaves as if it is correct. In an exemplary embodiment of the invention, the model is of emissivity and true temperature. In an exemplary embodiment of the invention, the checking of the model is by testing the behavior of the model as compared to one or more observations, under different detection conditions, for example, different signal amplification conditions. The determination of whether the behavior was as expected is optionally performed using one or more statistical tests.

In an exemplary embodiment of the invention, the model is generated by generating a set of models, for example covering a model space and testing all of the models. Alternatively, other search methods known in the art may be used. The inventor has realized that amplification by a detector is equivalent to changing of an object's emissivity. This realization is used for some types of processing according to the present invention. Another realization by the inventor, which is used in some embodiments of the invention is that, when the amplification is the reciprocal of the object's emissivity, the actual viewed brightness temperature is the same as the true temperature. In some embodiments of the invention, use is made of the non-linear relationship between emissivity and brightness temperature (for a particular true temperature) to detect when and if the emissivity of the object is correctly estimated. Another realization by the inventor which is used in some embodiments of the invention is that the sensitivity of the calculated true temperature to changes in brightness temperature is greater at low emissivities. This is a feature which is corrected for in some embodiments and/or utilized in some embodiments.

In some embodiments of the invention, the pyrometer detects and uses small fluctuations of brightness temperature of an imaged object, for example spatial and or temporal variations.

In an exemplary embodiment of the invention, a pyrometer according to the present invention is manufactured to have very narrow and precise wavelength calibration and also to have a very precise detection and amplification ability and/or high signal to noise ratio, for example, over 50 or 100. Such precision may be of use for detecting small measurement fluctuations (rather than detector or system noise). Optionally, only a single wavelength and/or single measurement is used to obtain a temperature estimation. In an exemplary embodiment of the invention, digitization of a temperature/emissivity space is used to generate and/or find models of the object. Optionally, the digitization steps are chosen to be larger than the noise level of the device, and, optionally, smaller than the fluctuations being measured.

In an exemplary embodiment of the invention, a single wavelength pyrometer is used as part of a ID or 2D mapping pyrometer, for example using a sensor array or using a mechanical scanner.

In an exemplary embodiment of the invention, no calibration is required beyond calibration of the detector in general. Alternatively, some process dependent information may be provided, for example a temperature range or information about statistical properties of the process. Optionally, this information is used to speed up the temperature estimation process or to reject substantially impossible solutions (e.g., very low emissivity), rather than as a required part of the process.

Optionally, a pyrometer in accordance with the present invention is used for hot industrial processes, such as metal manufacturing and forming and silicon processing. Alternatively, with suitable sensors, a lower temperature process is provided for, for example imaging of a landscape.

In an exemplary embodiment of the invention, the estimated temperature is exact to within 7%, 5%, 2%, 1%, 0.5%, 0.25% or any intermediate or smaller percentage of the true temperature. Similar precision may be provided for the emissivity. In general, however, a precision of greater than 5% or 1% is desired. In some embodiments, a certain failure rate, for example, of l%-5% may be allowed, in determining a correct or any seemingly correct temperature.

There is thus provided in accordance with an exemplary embodiment of the invention, a method of optical pyrometery, comprising: (a) making at least one measurement of a brightness temperature of an object;

(b) generating at least one model of the temperature of the object;

(c) testing the model under a plurality of different conditions; and (d) accepting the model if it performs in an expected manner under said different conditions. Optionally, generating a model comprises generating a model responsive to the measured brightness temperature of the object. Alternatively or additionally, making at least one measurement comprising reconstructing a brightness temperature using a pyrometer calibration factor. Alternatively or additionally, performing in an expected manner comprises generating a series of true temperatures under the different conditions and said series bear an approximation of an expected relationship to the true temperature under the different conditions. Alternatively or additionally, performing in an expected manner comprises generating a series of emissivities under the different conditions and said series bear an approximation of an expected relationship to the emissivities under the different conditions.

In an exemplary embodiment of the invention, accepting the model comprises checking the model under a further set of conditions. Alternatively or additionally, said different conditions comprise different signal amplification conditions of a signal received from the object, which are expected to yield a same true temperature for all conditions and emissivities linked by the amplification conditions.

In an exemplary embodiment of the invention, the different conditions comprise different real conditions. Alternatively or additionally, the different conditions comprise different calculated conditions. Alternatively or additionally, the expected manner comprises having an expected distribution property of measurements. Optionally, said property is maximum width. Alternatively or additionally, said property is skewedness.

In an exemplary embodiment of the invention, said generating comprises generating using a discrete set of values for true temperature and emissivity. Optionally, the method comprises increasing the minimum emissivity in the range as an effective emissivity of the model is increased. Alternatively or additionally, the method comprises extracting a true temperature from the model, taking into account the effective emissivity.

In an exemplary embodiment of the invention, the method comprises changing at least one parameter of said trying or of said accepting if no suitable model is found.

In an exemplary embodiment of the invention, the method comprises increasing an emissivity parameter of said model and repeating said testing. Optionally, the method comprises increasing said emissivity to near a value of 1. Alternatively or additionally, the method comprises using said emissivity parameter to calculate an expected brightness temperature. In an exemplary embodiment of the invention, the method comprises using said emissivity parameter as a limit for finding a new model.

There is also provided in accordance with an exemplary embodiment of the invention, an optical pyrometry method for estimating at least one of a true temperature and emissivity of a body of unknown emissivity, comprising: a) making a plurality of measurements of an emission intensity of the body at a same substantially single wavelength; and b) doing a statistical analysis of the plurality of measured emission intensities to obtain an estimate of the true temperature of the body, without using prior knowledge of a precise value of emissivity for the body. Optionally, the plurality of measurements are made with different degrees of amplification or attenuation of the emission intensity. Alternatively or additionally, doing the statistical analysis comprises calculating a measured brightness temperature for each measurement.

In an exemplary embodiment of the invention, doing the statistical analysis comprises: (a) generating a plurality of candidate models for true temperature and emissivity;

(b) selecting a model which appears to best fit a behavior profile expected of a correct model. Alternatively or additionally, doing the statistical analysis comprises: a) choosing a discrete set of emissivities and a discrete set of true temperatures; b) calculating an expected brightness temperature for each combination of emissivity and temperature in the discrete sets; and c) for each measurement, finding a tentative emissivity in the discrete set of emissivities, and a tentative true temperature in the discrete set of true temperatures, which tentative temperature and emissivity minimize the absolute difference between the expected brightness temperature and the measured brightness temperature. In an exemplary embodiment of the invention, doing the statistical analysis comprises obtaining the estimate of the true temperature by averaging the tentative true temperatures found for a plurality of the measurements. Optionally, the method comprises obtaining an estimate of the emissivity of the body by finding an emissivity which would produce the measured brightness temperature at the estimated true temperature. In an exemplary embodiment of the invention, doing the statistical analysis comprises choosing a discrete set of emissivities and a discrete set of true temperatures. In an exemplary embodiment of the invention, at least one of choosing the discrete set of emissivities and choosing the discrete set of true temperatures comprises choosing a range of the set that depends on the measured emission intensity.

In an exemplary embodiment of the invention, making a plurality of measurements of the emission intensity of the body is repeated for a plurality of different spots on the surface of the body, and doing a statistical analysis of the plurality of measured emission intensities comprises doing a statistical analysis of the plurality of measured emission intensities at each spot in the array, to obtain an estimate of the true temperature for each spot in the array.

There is also provided in accordance with an exemplary embodiment of the invention, an optical pyrometry apparatus for estimating at least one of a true temperature and emissivity of a body, comprising: a) a radiation detector configured to measure the emission intensity of the body at a substantially single wavelength; and b) a control module configured to direct the radiation detector to make a plurality of measurements of the emission intensity of the body at the same substantially single wavelength and at different amplifications, and to use the plurality of measured emission intensities to obtain an estimate of the true temperature of the body, without using prior knowledge of a precise value of emissivity of the body. Optionally, the radiation detector is configured to measure the emission intensity at a plurality of locations on the body, and the control module is configured to direct the radiation detector to make a plurality of measurements of the emission intensity at each of the locations, and to use the plurality of measured emission intensities to obtain an estimate of the true temperature at each of the locations. Optionally, the radiation detector successively scans the array of locations. Alternatively or additionally, the radiation detector is a compound detector comprising a set of individual detectors, each of which is configured to measure the emission intensity at a different one or more locations in the array.

BRIEF DESCRIPTION OF DRAWINGS Exemplary embodiments of the invention are described in the following sections with reference to the drawings. Fig. 1 is a general flow chart, showing a method of true temperature estimation, according to an exemplary embodiment of the invention;

Fig. 2 is a flow chart of a method of processing and analyzing data in the method of Fig. 1, in accordance with one embodiment of the invention; Figs. 3A-3C are flowcharts of various details of Figs. 1 and 2, in accordance with exemplary embodiments of the invention; and

Fig. 4 is a schematic block diagram of a scanning pyrometer in accordance with an exemplary embodiment of the invention, viewing one or more bodies on an assembly line. DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

Fig. 1 shows a general flowchart 100 for a method of estimating a true temperature of an object, using passive pyrometer techniques, in accordance with an exemplary embodiment of the invention. In an exemplary embodiment of the invention, the method includes making at least one measurement, treating the measurement in a manner that includes generating one or more models, checking if any of the models are good enough, and if not, modifying various parameters. While the following describes a particular technique, in which emissivity is increased monotonically, this is not essential for all embodiments of the invention. Instead, a person skilled in the art will appreciate that various search techniques which apply the "treatment" (or other treatments, for example as detailed below), may be used instead. At 101, the device used is optionally calibrated. It is a feature of some embodiments of the invention, that a very precise pyrometer detector is used in which very high precision is provided in one or both of wavelength and light flux detection. In addition, very high quality A/D converters and amplifiers are optionally provided.

At 102, a signal

at a wavelength λ is detected from an object whose temperature is to be estimated.

At 104, a brightness temperature Tj, (marked as Ttøi ) is calculated from the signal ui , based on the calibration factors determined at 101.

At 106, the brightness temperature is treated, in a manner to be described below, and which results in a set of results. At 108, the results are checked to see if they are suitable. If so, the current state is stored (110), which includes an estimated T^g and/or an estimated £_ruQ.

At 112, a parameter of the treatment, a minimum ε value used for searching, is increased. In an exemplary embodiment of the invention, the increase is the same as the range provided for within the treatment (as described below, this can be an increase by a factor of (amplification factor)20, if 20 steps are made in the treatment. It should be noted that this ε value also comprises an effective ε value (e.g., for storing in a state), for some embodiments of the invention. If the results were suitable, this increase is used for checking the results. This is optional and the process may jump instead to 122 (where the results are extracted). If the results were not suitable, this increase is used to look for a better set of results.

The inventor has found that as the emissivity approaches 1, more solutions (e.g., stable models) can generally be found, due to the non-linearity of the relationship between T, ε and

T\₎. In some embodiments of the invention, only the effective ε used in a model for calculating the brightness temperature is modified, while the range of emissivities used by the treatment to search for stability is not or is differently modified. It should also be noted that as ε approaches 1, Ty, approaches T^g, meaning that the effective amplification caused by using the new effective ε is equivalent to the reciprocal of the emissivity. In some embodiments of the invention, the effective ε is used instead of calculating a new pair {Tj, εj} for the new brightness temperature.

At 114, a check is made to see if the increased ε is not too high, for example lower than a threshold of, for example, 0.8, 0.9, 1.0 or some other threshold, which is optionally based on a range of amplification used in the treatment. If not, at 116, a new brightness temperature is calculated based on the new effective ε, for example using the new emissivity and the previous true temperature of the model or by amplifying the input signal using the factor of increase of ε or by multiplying ε by the factor inside the equation connecting T, Ttø and ε. Then, (106)-(114) are repeated. In some embodiments of the invention, the new brightness temperature is calculated inside the treatment.

At 118, if no good results were discovered so far, one or more treatment parameters are changed (120) and (106)-(118) are repeated, optionally from the original value of ε. In an exemplary embodiment of the invention, what is modified is an amplification factor used in the treatment to emulate changes in object emissivity. Act 118 may change the parameters also if a good result was found, but this result did not maintain itself over changes in effective ε, for example, over 3 or 4 cycles of increasing ε. Other methods may be used to determine if "good" results are actually a statistical artifact.

At 122, one or both of ε and Tt _e are extracted from the first saved "good state", after which all or most other states were good. Other statistical tests may be applied as well to determine which states are to be considered "good".

Optionally, at 124 and 126, the method is repeated for multiple wavelengths and/or multiple times. A better estimate of E and/or T may be, for example, an average of such results and/or a most common result and/or using the results the reject some results and/or any other statistical processing method known in the art to estimate a signal in the presence of multiple, noisy samples.

Fig. 2 is a flowchart 106 of a method of treatment of results in accordance with an exemplary embodiment of the invention. Briefly, the treatment method comprises generating a model that includes an estimated temperature T[ and an estimated emissivity ε[ and confronting that model with one or more changes and analyzing the results. An incorrect model is expected to fail in matching an expected behavior profile of a correct model. In an exemplary embodiment of the invention, the effective emissivity of the viewed object is modified by changing (e.g., amplifying) the gain of the detector (in reality or by calculation). This results in a new brightness temperature Ttø. New estimated true temperatures and estimated emissivities are generated for this new brightness temperature. If the model being confronted is a correct model, it will yield consistent results (e.g., a same estimated true temperature) when used in multiple different gain situations. Alternatively, one of the new estimated pairs of temperature and emissivity may prove to be a model that behaves consistently over a range of amplifications.

At 202, an amplification factor of n = 1 is selected, i.e., no amplification relative to the measurement conditions of the original brightness temperature.

At 204, a model relating the measured brightness temperature to a pair of Tj and ε[ is found. An exemplary such method is described below in Fig. 3 A.

At 206, the amplification factor n is multiplied by an emissivity amplification factor, for example, 1.02 or 1.01, in general, a small number.

At 208, 204 and 206 are repeated multiple times. A parameter k may be used to count the number of repeats, which may be, for example, a preset value k, a value depending on the quality of the results (e.g., Fig. 1, act 108) or modified (e.g., Fig. 1, act 120). Parameter k may also be a function of the current effective emissivity.

The result is a set of rows (if a table arrangement is used), each corresponding to one amplification factor and each of which contains one pair of {T[, εj} and possibly other parameters. At 210, one or more optional filtration procedures are applied, to delete one or more rows. Exemplary such filters are described in Fig. 3B below. Referring back to Fig 1, act 108, the remaining rows are expected to have the following properties, if the original model is correct. First, the estimated true temperatures Tj for all the rows should be the same to within a small error factor. Second, the estimated emissivities εj should match a geometrical series starting at the original emissivity ε, with the amplification factor being the series quotient. In an exemplary embodiment of the invention, it is desired that for at least 70% or 80% of a consecutive set of 10 rows the estimated temperature be within 3% or 5% of the average of temperatures. Some small spaces of missing rows may be allowed in some embodiments of the invention. A similar constraint may be placed on the emissivity, for example 80% of the emissivities be within 10% of the average emissivity. Other constraints may be placed in other embodiments of the invention, for example, different required numbers of consecutive rows (e.g., 5, 8, 15, 20 or intermediate or larger or smaller numbers), different amounts of deviations (e.g., less than 20%, 10%, 7%, 2%, or intermediate numbers) and/or different percentages (e.g., 60%, 85%, 95% or intermediate or larger numbers). Alternatively or additionally, a constraint may be set on various statistical properties of the results, for example, their range, standard deviation, shape of distribution and/or skewedness. A statistical test may be expected to be definite to at least 75%, 85%, 90%, or intermediate or larger numbers. Optionally, the first row is dropped.

Fig. 3 A is a flowchart of one method of finding a model for Fig. 2, act 204, in which a set of discrete values for T and ε are used. At 302 a range of temperatures T{min,max} is selected, as well as a step ΔT. At 304, a range of emissivities ε{min,max} is selected as well as a step Δε. It should be noted that in one embodiment of the invention, ε{min} is increased at act 112. Iterating over the ranges of T and ε yields a series of pairs {T[, εj}. Each such pair describes a model.

At 306 a test brightness temperature Ttø_est is calculated as a function of T[ and E . This is repeated for all the pairs {Tj, εj} .

At 308, a brightness temperature Ttø_n for the current amplification factor is provided, for example by measurement or by calculation (can be pre-calculation, e.g., at 116). In an exemplary embodiment of the invention, the calculation is by multiplying the current effective emissivity by the amplification factor n. Another method is to take a new measurement with a new amplification setting. Another method is to store the original detected signal and amplify it, using analog or digital means. At 310, a model is selected for which the difference between the brightness temperature calculated by the model and the "real" brightness temperature for the current amplification factor (either a measured one or an artificially amplified one), is minimized:

Δ=min |Tbtest(^Ti> εi)-Tbnl It should be noted that due to the use of digitization, Δ will generally not be zero for any of the models. It should be noted that if the model is correct, T[ should remain the same for all the models with minimum difference Δ.

In some embodiments of the invention, ε{min} is increased for some situations. In one example, as the number of amplifications increase, the ε{min} is increased, for example, by the same amplification factor. This may reduce noise levels in the lower emissivity cases.

Fig. 3B is a flowchart of optional applying of filters (210) to delete amplification factor rows. One or more of these filters is optionally applied. Alternatively, other filters may be applied.

At 322, rows where the model uses an emissivity of greater than 1, is dropped. In this context it should be noted that an effective emissivity of exactly 1 is reached when the amplification factor is the inverse of the real emissivity of the object. In some embodiments of the invention, emissivities slightly (e.g., 0.5%-5%) greater than 1 may be allowed, for example to compensate for the fluctuations, noise and/or digitization effects of the system. Alternatively or additionally, ε is allowed to vary to be greater than 1, to prevent edge effects. At 324, rows that have extreme values are rejected. For example, rows in which the estimated true temperature is different by more than 5% (or some other threshold, such as a function of a standard deviation of the values) from an average value, are rejected. A similar statistical method maybe used to reject rows based on the distribution of emissivity values.

At 326, rows in which the estimated emissivity is not monotonically increasing as compared to the other rows, are deleted. For example, a majority rule method may be used to select which are the rows that do not conform to the increase and which rows set the increase.

Fig. 3C is a flowchart of a method of extracting the temperature and/or emissivity from the results (122), in accordance with an exemplary embodiment of the invention.

At 342, it is decided which are the saved set of results to use, to provide the brightness temperature (e.g., measured or calculated) and true temperature (e.g., approximately the same for all values in a result set). For example, the first saved results set is optionally used.

At 344, the emissivity is calculated as ε=exp(c2/λ(T_true-l-Tb-l)) If a brightness temperature other than a measured one (in which a real amplification is known) is used, correction is optionally provided for the factor caused by mathematical amplification of the brightness temperature.

Optionally (346), the average values (or some other statistical processed values) are used instead of the value of a particular model. In an exemplary embodiment of the invention, the various ranges, steps sizes and/or thresholds used are selected for a particular process. For example, in an aluminum extruding process, T{min,max}={300 °C, 600 °C}, ε{min,max}= {0.1,1}, ΔT=1 °C, Δε=0.01 and the thresholds are as above. In an exemplary embodiment of the invention, ΔT is selected to be over the system noise level. In an exemplary embodiment of the invention, the system is built to have the following properties. A brightness temperature precision (as calibrated against a black body) is, for example, better than 0.1%, 0.05% or 0.01%. An effective wavelength precision is, for example, better than 0.5% or even 0.1% (measured at half amplitude), for example, 70 nm for a 2.2 μ wavelength. A detector stability is, for example, better than 1%. An amplifier linearity is, for example, better than 1%. An A/D linearity, is, for example, better than 0.01%. While such components are well within the art, apparently, to date, there has not been felt a need for such precision. In an exemplary embodiment of the invention, the input signal is chopped, with a plate having apertures with one or more narrow wavelength filters, and the dark times between apertures is used to correct a base current. In an exemplary embodiment of the invention, the detector is heat controlled using a thermostat to within ±0.5 degrees Kelvin.

It should be noted that in the relationship between emissivity and brightness temperature shown in equation (4,5), there are two expressions that include ε, and the smaller one is ignored. However, it is optionally taken into account and the equations used adjusted appropriately. It should be noted that once a correct temperature and emissivity model are found, various correction and precision increasing methods may be applied, for example, as described above, averaging multiple results, which allow models that are not exactly matching to digitization points, to be generated.

It should also be noted that while the above method is useful for passive pyrometry, the above method can also be used for active pyrometry, with suitable modifications in the model being detected and in how it is expected to behave.

In the above description, a simple search method has been described. It should be appreciated that a different search method may be used, for example, one which follows a corridor surrounding the brightness temperature curve (relationship between ε and T for a given brightness temperature). In another example, rather than starting from a low emissivity, a high emissivity is used as a starting point and the effective emissivity reduced, rather than increased, at 112. In another example, a plurality of {T, ε} pairs are generated and each such pair is tested under the various conditions and no further search is made in the various conditions. In this context it is noted that the relationship between T, ε and Ttø is non-linear and that change sin amplification correspond to proportional change sin emissivity of the viewed object. If an initial ε is assumed and the amplification is changed by a known step then a calculation of the true temperature based on a new measurement of Ttø should result in a same T. This does not require the searching described above, per se. In any case, a best fit for a set of measurements can be used.

The method of Fig. 3A may be varied in other ways as well. For example, in one embodiment of the invention, once an initial guess {T^ε } is made, a repeated measurement

(or calculation) under a different amplification factor is made and what is compared is this measured brightness temperature and what the initial guess shows the model should be. This difference is expected to be very small for a set of measurements, as described above with regard to checking matching of a series of temperatures to a true temperature. This may be applied to another one of the calculated "rows", instead of finding new models for each row, so that a mix of the two methods is optionally applied. In another variation, for each amplification factor a new guess {T[,s[} is made and compared to the brightness temperature that the current effective emissivity would guess at using the correction factor of the amplification. Optionally, in some embodiments, each amplification factor is checked to see if T is stable under various amplification and/or measurement conditions. Then a new model {Ti,εi} is generated for that amplification. Optionally, rows that do not stay stable are dropped and rows that stay may be used directly for an estimate of the correct emissivity.

In some embodiments of the invention no "search" in a space of models is required. Instead, a statistical equation(s) relating the expected results and their expected statistical variable, is generated. This equation (or set of equations) can then be solved, using methods known in the art, including analytical and numerical methods. For minimum based methods (e.g., finding a minimum skew in fluctuations relative to a model of the fluctuations), such equations may be differentiated. Also, it should be noted that the above statistical processing can be varied, for example, using various T tests or checking the actual shape of the distribution and/or taking standard deviation into account, rather than how many point fall within a certain range.

Referring to the above described possibility of various statistical tests. In an exemplary embodiment of the invention, some basic statistical properties of the process are assumed, for example, that the fluctuations have a certain distribution or that most of them fall within an expected width. It should be noted that some of this information (e.g., true or brightness temperature fluctuations and distribution statistics) can be detected using contact pyrometery, for example using a thermocouple and/or using expensive and fragile devices, such as active pyrometers. Making multiple measurements typically provides a distribution of the measurements that follow the real distribution. In an exemplary embodiment of the invention, a first emissivity at which the distribution width of the values is smaller than an expected range is looked for, this emissivity will generally be close to the real emissivity. In another exemplary embodiment, when the emissivity is wrong, the resulting distribution of temperatures is incorrect, as the relationship between true temperature, brightness temperature and emissivity is not linear. Other statistical properties of fluctuations distribution may be searched for as well. Optionally, for a new process a series of possible statistical properties are searched for to select an appropriate property and/or test. It should be noted that the distribution of such properties is not generally expected, for some processes. Possibly, the fluctuations are caused by statistical variations in surface geometry, small temperature currents, cooling air currents, smoke and steam clouds and/or changes in surface chemistry.

In one embodiment of the invention, the real temperature fluctuations may be assumed to be smaller than the pyrometer self- fluctuations. If the Emissivity is incorrectly estimated, however, this may cause calculated fluctuations in temperature greater than expected. Optionally, this is used to provide a rough estimate for ε. Fig. 4 shows an assembly line with a sheet of aluminum 408, or another metal, extruded between rollers 407. Other assembly lines may be used instead. A pyrometer 401, for example as described above, views a spot 405 on sheet 408, as it goes past the radiation detector, measuring the intensity of radiation emitted by spot 405 within a desired range of wavelengths. In an exemplary embodiment of the invention, pyrometer 401 comprises a scanner for example, for ID or 2D scanning of aluminum 408. In an exemplary embodiment of the invention, as a generally short distance and possibly wide angle view is used for industrial processes, a vibrating mirror scanner is used. For example, a mirror 410 is vibrated in one axis by a first motor 412, which is itself vibrated in a second axis by a motor 414. A controller 430 described below may control the vibrations and optionally synchronize it to the rotation of a chopper 418 described below. A detect signal is optionally focused by a lens 416 and optionally chapped by a chopper 418. In an exemplary embodiment of the invention, the dark current of the chopper is used to correct the signal detected by a detector 422. Optionally, this is done using analog means and/or by controller 430. The chopper is rotated by a motor 420 and may include multiple narrow wavelength filters. An amplifier 428 is optionally provided to amplify the signal detected by detector 422 for general use and/or for carrying out the method described above. It should be noted that attenuation may be used instead of positive amplification for some embodiments of the invention. Optionally, the detector is relatively large, for example 3mm square, however this is not essential. In other embodiments a one or two dimensional detector anay may be used instead of a scanner. Optionally, detector 422 is temperature controlled by a thermostatic controller 426 using a heater and/or cooler 424.

The signal is converted to a digital signal by an A/D and controller 430 and then transmitted to a general purpose computer 432, where further processing is provide before display of a true temperature estimate and/or an emissivity estimate on a display 434. Of course the above is a general description of a particular embodiment and should not be construed as overly limiting. For example, controller 430 may complete the processing. Further, the scanning system and/or detection system may be varied, as well as or instead of, the amplification system.

Alternatively or additionally to using a display 434, the pyrometer may be used to directly or indirectly control the industrial process, for example, controlling temperature, speed, raw materials and/or other industrial variable. Alternatively or additionally, the results are conveyed to a process control computer or to a monitoring station. Optionally, information about the spatial distribution of brightness temperature, found by observing the set of spots, is used to put limits on the emissivity of the body. For example, information about surface roughness, which is often positively correlated with emissivity, is optionally obtained by statistically analyzing the spatial variations in brightness temperature. Deep narrow pits in the surface of the body being measured, which may approximate black body cavities, are optionally detected as local spots of higher brightness temperature than their surroundings.

In an exemplary embodiment of the invention, pyrometer 401 is calibrated using a black body. In an exemplary embodiment of the invention, the calibration is performed after the black body has stabilized for a relatively long period of time, for example 30 minutes. Alternatively or additionally, the calibration (which, for example, finds a pyrometer calibration factor K, related for example to the sizes of the lens and detectors and transmission of filters) is also used to determine to a high precision the exact wavelength being measured. An exemplary calibration process is described in a US provisional application, of same title and filed on a same date as this PCT application and having attorney docket number 375/03496.

Although Fig. 4 shows an exemplary embodiment of the invention being used in an industrial assembly line, the invention is not limited to such applications, but is optionally used in any situation, industrial or otherwise, where it is desired to monitor or control temperature. For example, it is optionally used in a real-time, two-dimensional infrared imaging array, with applications such as night-vision goggles for military use, medical imaging, and detection of poor thermal insulation in buildings.

It will be appreciated that the above described methods may be varied in many ways, including changing the order of steps and/or performing some steps in parallel. In addition, different apparatus arrangements may be used. For example, one or more sensors may be used and various calculation circuitry, software and/or hardware may be used. Further, an existing pyrometer may be modified to use the above method. It should also be appreciated that the above described description of methods and apparatus are to be interpreted as including apparatus for carrying out the methods and methods of using the apparatus. It should be understood that features and/or steps described with respect to one embodiment may be used with other embodiments and that not all embodiments of the invention have all of the features and/or steps shown in a particular figure or described with respect to one of the embodiments. Variations of embodiments described will occur to persons of the art.

It is noted that some of the above described embodiments may describe a best mode contemplated by the inventors and therefore may include structure, acts or details of structures and acts that may not be essential to the invention and which are described as examples. Structure and acts described herein are replaceable by equivalents which perform the same function, even if the structure or acts are different, as known in the art. Therefore, the scope of the invention is limited only by the elements and limitations as used in the claims. When used in the following claims, the terms "comprise", "include", "have" and their conjugates mean "including but not limited to".

Claims

CLA S

1. A method of optical pyrometery, comprising:

(a) making at least one measurement of a brightness temperature of an object; (b) generating at least one model of the temperature of the object;

(c) testing the model under a plurality of different conditions; and

(d) accepting the model if it performs in an expected manner under said different conditions.

2. A method according to claim 1, wherein generating a model comprises generating a model responsive to the measured brightness temperature of the object.

3. A method according to claim 1, wherein making at least one measurement comprising reconstructing a brightness temperature using a pyrometer calibration factor.

4. A method according to claim 1, wherein performing in an expected manner comprises generating a series of tme temperatures under the different conditions and said series bear an approximation of an expected relationship to the true temperature under the different conditions.

5. A method according to claim 1, wherein performing in an expected manner comprises generating a series of emissivities under the different conditions and said series bear an approximation of an expected relationship to the emissivities under the different conditions.

6. A method according to claim 1, wherein accepting the model comprises checking the model under a further set of conditions.

7. A method according to any of claims 1-6, wherein said different conditions comprise different signal amplification conditions of a signal received from the object, which are expected to yield a same true temperature for all conditions and emissivities linked by the amplification conditions.

8. A method according to any of claims 1-6, wherein the different conditions comprise different real conditions.

9. A method according to any of claims 1-6, wherein the different conditions comprise different calculated conditions.

10. A method according to any of claims 1-6, wherein the expected manner comprises having an expected distribution property of measurements.

11. A method according to claim 10, wherein said property is maximum width.

12. A method according to claim 10, wherein said property is skewedness.

13. A method according to claim 1, wherein said generating comprises generating using a discrete set of values for tme temperature and emissivity.

14. A method according to claim 13, comprising increasing the minimum emissivity in the range as an effective emissivity of the model is increased.

15. A method according to claim 13, comprising extracting a tme temperature from the model, taking into account the effective emissivity.

16. A method according to any of claims 1-6, comprising changing at least one parameter of said trying or of said accepting if no suitable model is found.

17. A method according to any of claims 1-6, comprising increasing an emissivity parameter of said model and repeating said testing.

18. A method according to claim 17, comprising increasing said emissivity to near a value of l.

19. A method according to claim 17 or claim 18, comprising using said emissivity parameter to calculate an expected brightness temperature.

20. A method according to claim 17, comprising using said emissivity parameter as a limit for finding a new model.

21. An optical pyrometry method for estimating at least one of a true temperature and emissivity of a body of unknown emissivity, comprising: a) making a plurality of measurements of an emission intensity of the body at a same substantially single wavelength; and b) doing a statistical analysis of the plurality of measured emission intensities to obtain an estimate of the tme temperature of the body, without using prior knowledge of a precise value of emissivity for the body.

22. A method according to claim 21, wherein the plurality of measurements are made with different degrees of amplification or attenuation of the emission intensity.

23. A method according to claim 21, wherein doing the statistical analysis comprises calculating a measured brightness temperature for each measurement.

24. A method according to claim 21, wherein doing the statistical analysis comprises: (a) generating a plurality of candidate models for true temperature and emissivity;

(b) selecting a model which appears to best fit a behavior profile expected of a coreect model.

25. A method according to claim 23, wherein doing the statistical analysis comprises: a) choosing a discrete set of emissivities and a discrete set of tme temperatures; b) calculating an expected brightness temperature for each combination of emissivity and temperature in the discrete sets; and c) for each measurement, finding a tentative emissivity in the discrete set of emissivities, and a tentative true temperature in the discrete set of true temperatures, which tentative temperature and emissivity minimize the absolute difference between the expected brightness temperature and the measured brightness temperature.

26. A method according to claim 25 wherein doing the statistical analysis comprises obtaining the estimate of the tme temperature by averaging the tentative tme temperatures found for a plurality of the measurements.

27. A method according to claim 26, also comprising obtaining an estimate of the emissivity of the body by finding an emissivity which would produce the measured brightness temperature at the estimated tme temperature.

28. A method according to any of claims 21-23, wherein doing the statistical analysis comprises choosing a discrete set of emissivities and a discrete set of tme temperatures.

29., A method according to any of claims 24-27, wherein at least one of choosing the discrete set of emissivities and choosing the discrete set of tme temperatures comprises choosing a range of the set that depends on the measured emission intensity.

30. A method according to any of claim 21-27, wherein making a plurality of measurements of the emission intensity of the body is repeated for a plurality of different spots on the surface of the body, and doing a statistical analysis of the plurality of measured emission intensities comprises doing a statistical analysis of the plurality of measured emission intensities at each spot in the array, to obtain an estimate of the tme temperature for each spot in the array.

31. An optical pyrometry apparatus for estimating at least one of a true temperature and emissivity of a body, comprising: a) a radiation detector configured to measure the emission intensity of the body at a substantially single wavelength; and b) a control module configured to direct the radiation detector to make a plurality of measurements of the emission intensity of the body at the same substantially single wavelength and at different amplifications, and to use the plurality of measured emission intensities to obtain an estimate of the tme temperature of the body, without using prior knowledge of a precise value of emissivity of the body.

32. Apparatus according to claim 31, wherein the radiation detector is configured to measure the emission intensity at a plurality of locations on the body, and the control module is configured to direct the radiation detector to make a plurality of measurements of the emission intensity at each of the locations, and to use the plurality of measured emission intensities to obtain an estimate of the tme temperature at each of the locations.

33. Apparatus according to claim 32, wherein the radiation detector successively scans the array of locations.

34. Apparatus according to claim 32 or claim 33, wherein the radiation detector is a compound detector comprising a set of individual detectors, each of which is configured to measure the emission intensity at a different one or more locations in the array.