Contributed paper OPTO-ELECTRONICS REVIEW 13(3), 253 257 Influence of sun radiation on results of non-contact temperature measurements in far infrared range H. MADURA * and M. KO ODZIEJCZYK Institute of Optoelectronics, Military University of Technology, 2 Kaliskiego Str., 00-908 Warsaw Non-contact measurements of an object temperature in IR range carried out in outdoor conditions can suffer from significant errors. An error of temperature measurement can be very high when sun radiation after reflection from an object propagates along optical axis of a measuring device (thermovision camera, pyrometer). Radiation beams reflected from an object and object s radiation itself are added and a final value of temperature is higher. The paper presents results of theoretical estimations of errors of temperature measurement and their comparison with experimental results. Calculations and measurements were made for objects of various emissivities in a spectral range of 8 9 µm. Keywords: pyrometry, thermography, sun radiation. 1. Introduction To make proper temperature measurement with pyrometers, emissivity of an investigated object should be known. Most frequently, a main source of measurement errors is improperly taken emissivity value. In some types of pyrometers, influence of object emissivity on a result of temperature measurement is significantly reduced [1]. If temperature measurements are carried out for an object illuminated with sun radiation, beams of sun radiation reflected from an object are added and the read value of the temperature is higher than actual one. In Refs. 2 and 3, IR pyrometers are described in which sun radiation is monitored but influence of this radiation on a result of temperature measurement was not estimated. Influence of sun radiation on a result of temperature measurement increases for low emissivity objects as well as for objects of low temperature. with spectral luminance of a source for the normal direction L n (l,t) as W Ln ( l, T ) 1 M (, T ) é ù» l ê ú. (2) p 2 ëcm mmsr û Knowing geometrical relations between a radiating object and a receiver of this radiation (Fig. 1), spectral distribution of intensity of radiation reaching input aperture of a receiver [4] can be determined E T L T S 1cos q1cos q2 ( l, ) = n ( l, ) ê ú, (3) 2 2 R ëcm mmû where S 1 is the area of an object s radiating surface, R is the distance between a radiation source and a measuring device in cm, q 1 is the angle between a normal to object 1.1. Used relations and notations According to the Planc s law, the dependence of spectral distribution of black body radiation is determined as M( l, T) = c1 c lt ê ú, (1) 5 l ( e 2 2-1) ëcm mmû where Plank constants c 1 = 37417.7107 ±0.0029 (Wµm 4 cm 2 ) and c 2 = 14387.752 ±0.025 10 2 µmk, is the wavelength, and T is the temperature of a black body. Spectral distribution of black body radiation is connected * e-mail: hmadura@wat.edu.pl Fig. 1. Geometrical relations between surfaces emitting and absorbing radiation and a camera or a pyrometer. Opto-Electron. Rev., 13, no. 3, 2005 H. Madura 253
Influence of sun radiation on results of non-contact temperature measurements in far infrared range surface and direction of radiation propagation, q 2 is the angle between optical axis of a measuring device and direction of radiation propagation. Assuming that q 2 =0, and substituting Eq. (2) into Eq. (3) we have 1 E T M T S 1cos q1 ( l, ) = ( l, ) ê ú. (4) p 2 2 R ëcm mmû Thus, total intensity of object radiation in the spectral range of l 1 l 2 wavelengths is l2 E = ò E ( l, T ) dt ê 2 ú l ëcm û. (5) 1 Total power irradiated from an object reaching input aperture of the optical system S 2 is given as P0 = S2E[ W], (6) where S 2 (cm 2 ) is the area of input aperture surface of a measuring device. 2. Sun radiation A temperature of the Sun surface is about 5900 K. Spectral distribution of sun radiation is the best approximated by a black body of the temperature of 5770 K, the size of which corresponds to the Sun s size [5]. This body emits uniform radiation in all directions. Total intensity of sun radiation measured outside the earth atmosphere is called solar constant. A value of a solar constant taken in 1981 by World Meteorological Organization (WMO) is 1367 Wm 2. The value of solar constant accepted in the year 2000 by American Society for Testing and Materials (ASTM) is 1366.1 ±0.08% Wm 2 [6]. This value has been obtained due to many-year registration of sun radiation in the range from 120 nm to 1000 µm [7]. Fig. 2. Solar spectral irradiance outside the atmosphere proposed by ASTM. Spectral solar irradiance outside the atmosphere E l, proposed by ASTM (Fig. 2), insignificantly differs from the distribution proposed by WMO, especially for the waves of wavelengths above 2 µm. These distributions are used in various simulation programs, e.g., MODTRAN, LOWTRAN or PCMODWIN. Using these programs, intensity of sun radiation reaching the Earth surface or coefficient of atmosphere transmittance can be determined. Because of the earth atmosphere properties, only part of sun radiation reaches the Earth surface. Sun radiation is absorbed and scattered by atmospheric gases and aerosols. A ratio of sun radiation intensity measured outside the earth atmosphere to sun radiation intensity reaching the Earth surface, for particular wavelengths, is called spectral coefficient of the earth atmosphere transmittance t E (l). Precise determination of t E (l) is difficult because it requires consideration of many factors and parameters describing the atmosphere characteristics. The most important factors are the content and condensation rate of water steam and influence of gases and aerosols in particular layers of the atmosphere. The t E (l) coefficient is also influenced by temperature and atmospheric pressure. When the value of t E (l) is determined, a thickness of the earth atmosphere should be considered depending on a position of a measuring point at the Earth surface in respect to the Sun. The power of sun radiation P S reaching, through the atmosphere, the object of the surface S 1 can be described as l2 PS = S1òt E( l) Eldl [ W]. (7) l 3. Model of object radiation 1 In real measuring conditions, a value of radiation reaching input aperture of an optical system of a thermovision camera depends on many factors. In majority of cases we can only estimate them because without using special apparatus we cannot determine them precisely. The main factor is radiant property of the object itself which is called the emissivity e 0. In measuring practice, often emissivity is not known and using available tables there is no certainty for its proper determination. Thus, in some measuring conditions the object s emissivity can be significantly different. It results from the fact that emissivity depends on such factors as, e.g., structure and ratio of surface oxidation, the object s temperature T 0, direction of observation or spectral range for which it is determined. Inaccurate determination of emissivity value of an object is a main reason of measurement errors. Other factors influencing a value of radiation reaching input aperture of optical signal of a thermovision camera are: radiation reflected from an object surface, absorption, dispersion, and radiant properties of the atmosphere. 254 Opto-Electron. Rev., 13, no. 3, 2005 2005 COSiW SEP, Warsaw
Contributed paper Fig. 3. Reflection of sun radiation from scattering (a) and mirror (b) surfaces. Total power reaching input aperture of optical system, for typical measuring conditions, can be written in form of expression used in majority of FLIR systems [8] P = e0t ap0 + ( 1- t a) Pa + ( 1-e0) t aprefl [ W], (8) where e 0 is the object emissivity, t a is the transmittance coefficient of atmosphere between an object and a camera, P 0 is the power of object radiation, P a is the power of atmosphere radiation, and P refl is the power of radiation reflected from an object surface. We can distinguish two basic radiation reflections from an object surface, i.e., scattered reflection and mirror one (Fig. 3). For intensively scattering surfaces, the Lambert s cosine law is used Prefl = PS cos acos b [ W], (9) where P S is the total sun radiation power incident on object s surface, a is the angle of incidence, and b is the angle of reflection. For smooth surfaces we can take, with a good approximation, that radiation reflection is of a mirror type, i.e., P refl» P S. In the calculation model, it was taken that sun radiation reflection from a plate surface is of a mirror type. Moreover, it was assumed that for short distances between an object and a camera t a = 1. Thus, finally, Eq. (8) is of the form P = e0p0 + ( 1-e0) PS [ W]. (10) 4. Estimation of measurement errors. Simulation results Solar irradiance reaching the Earth surface in spectral range of 8 9 µm was determined using PcModWin program (Fig. 4). It was calculated that for weather conditions of the following parameters: angle of deflection of the Sun of 50, metrological visibility of 23 km, cloudless sky, ambient temperature 285 K, atmospheric pressure 1027 hpa, relative air humidity 40%, the total sun radiation intensity on the Earth surface was about 24 mwcm 2. Fig. 4. Solar spectral irradiance E S,l, determined with PcModWin program for the angle of deflection 50 of the Sun from the zenith and Rural aerosol model and visibility of 23 km. Fig. 5. Algorithm used for calculation of a relative error of object s temperature reading. Opto-Electron. Rev., 13, no. 3, 2005 H. Madura 255
Influence of sun radiation on results of non-contact temperature measurements in far infrared range Fig. 6. Calculated relative error of temperature reading of the object emissivity e 0 = 0.2 for various values of sun radiation intensity. On the basis of mathematical description of a phenomenon of sun radiation reflection, the algorithm has been written which was used for simulations in MATLAB calculation environment. A relative error of temperature reading has been estimated for various values of sun radiation intensity (Figs. 6 and 7). Fig. 8. Scheme of a measuring set-up for investigations of sun radiation influence on measurement results of object s temperature. Fig. 9. Spectral emissivity of S1-S5 plates. Fig. 7. Calculated relative error of temperature reading of the object emissivity e 0 = 0.9 for various values of sun radiation intensity. 5. Experimental results A measuring set-up has been built to investigate sun radiation influence on an error of temperature measurement in the field conditions (Fig. 8). A temperature of S1-S5 metal plates having various emissivities was measured. The plates were subsequently situated inside a housing to have them in a field of view of a thermovision camera. A distance between a camera s objective and a plate was 120 cm. Sun radiation reaches the examined plate through a rectangular hole cut in a front wall of a housing to illuminate only a part of the investigated plate. Due to this, simultaneous measurement of a temperature of the plate s part reached by sun radiation (in Fig. 8 denoted with lighter colour) and the part without sun radiation (darker colour) was made. Sun rays, after reflection from the investigated plate propagate along an optical axis of a thermovision camera. Fig. 10. A thermogram of S1 plate of the temperature T 0 = 285 K and a profile of temperature distribution along the marked straight line. 256 Opto-Electron. Rev., 13, no. 3, 2005 2005 COSiW SEP, Warsaw
Basing on the results of experimental investigations, a relative error of the temperature reading d, caused by reflection of sun radiation from the object s surface, can be determined. Figure 12 presents comparison of experimental and calculation results. Insignificant differences between theoretical and experimental results testify on good consistence between a model and experiment. Theoretical values are little higher than experimental ones. So, it should be supposed that it is due to consideration of sun radiation in a model. 6. Conclusions Contributed paper Fig. 11. Spectrum of radiation intensity of non illuminated and illuminated object (S1 plate, T 0 = 290 K). In order to eliminate a heating effect of the investigated plates because of incident sun radiation, a rectangular hole was being opened only for measurement duration, i.e., for about 2 s. A temperature of the investigated plates was stated with a heater and monitored with a thermometer. Spectral emissivity of the plates was determined using Specord 71 IR spectrophotometer (Fig. 9). Emissivity averaged for a given range of camera s operation was introduced into a thermovision camera. During experimental investigations, the thermograms were obtained that are used for estimation of sun radiation influence on temperature measurement results. In a thermogram, Fig. 10, a distinct place can be seen from which the sun rays have been reflected. A temperature measured with a camera at this place significantly differs from the actual object s temperature which did not change during a measurement. Spectrum of the object s own radiation and spectrum of the object illuminated with sun radiation were determined using SR-5000 spectroradiometer of the set-up shown in Fig. 8. According to predictions, a value of sun radiation after reflection from an object was added to the object s radiation (Fig. 11). Fig. 12. Relative error of temperature reading of the investigated plates S3 and S4 for temperatures 285 K, 295 K, and 305 K. The intensity of incident sun radiation E S» 24 µwcm 2. The presented calculation and measurement results indicate that sun radiation, even in far infrared range, can cause significant errors in temperature measurements when thermovision cameras or pyrometers are used. It should be taken into account that when the temperature measurements of the objects illuminated with sun radiation are made, a measurement error increases for low-emissivity objects and low-temperature ones. An error of temperature measurement depends on a shape and kind of a surface of the investigated object and a position of a pyrometer (IR camera) in relation to a direction of propagation of main flux of the reflected sun radiation. Thus, new methods of temperature measurement and new measuring devices should be find. In new measuring devices, influence of sun radiation on the result of temperature measurement should be reduced or an user of a measuring device should be informed that the measurement is improper (with error). References 1. H. Madura and T. Pi¹tkowski, Emissivity compensation method in double band pyrometry, Infrared Physics & Technology 46, 185 189 (2004). 2. H. Madura, T. Pi¹tkowski, and E. Powiada, Multispectral precise pyrometer for measurement of seawater surface temperature, Infrared Physics & Technology 46, 69 73 (2004). 3. C.E. Everest and G.K.Walker, Infrared temperature monitoring apparatus having means for sky radiation compensation, United States Patent, No. 4 420 265, 1983. 4. K. Seyrafi and S.A. Hovanessian, Introduction to Electro-optical Imaging and Tracking Systems, pp. 56-61, Artech House, Boston, London, 1993. 5. J.G. Zissis, Sources of Radiation 1, in The Infrared & Electro-Optical Systems Handbook, p. 151, SPIE Press, Bellingham, 1993. 6. Standard solar constant and zero air mass solar spectral irradiance tables. Standard E490-00. American Society for Testing and Materials, West Conshohocken, PA, 2000. 7. C.A. Gueymard, The sun s total and spectral irradiance for solar energy application and solar radiation models, Solar Energy 76, 425 (2004). 8. ThermaCAM TM Researcher, User s Manual, FLIR Systems. Opto-Electron. Rev., 13, no. 3, 2005 H. Madura 257
258 Opto-Electron. Rev., 13, no. 3, 2005 2005 COSiW SEP, Warsaw