Infrared Reflectance Spectroscopy And Thermographic . - Shroud

as measured with a moderate resolution laboratory instrument. Solid body speclra of common materials aregenerally broad featured, 4 and the spectral resolutionso obtained was adjudged adeq1)ate for the expectedspectral characteristics of the actual measurement.D.Experimental ProcedureT he focused blackbody source located rv40 cm fromthe target was positioned upon a preselected area ofinterest. The radiometer located 2 m from the clothwas focused on the area and adjusted for maximu111signal return. T he narrow FOV of the instrumentcontributed to positioning sensitivity. H owever, oncepositioned, the signal levels were stable. A spectrumwas recorded. The reforence surndard was then positioned directly over the area of interest, maintaininggeometry, and another spectrum was recorded. Signallevels in the 3- 5-µm band were considerably greaterthan in the 8-14-µm band, and corresponding attenuator changes were required.Of particular note during the course of the experiment was the relatively large fluctuations in atmospheric absorption, especially on those days when localprecipitation caused high relative humidity. Althoughsufficient data were taken to enable approximatespectral recovery, t he atmospheric fluctuations frommeasurement to measurement accompanied by inherent system noise were of such amplitude that the determination of absolute values of reflectance was unreliable. A further difficulty with t he circular variablefilter on the radiometer necessitated termination priorto completion, resulting in fewe r measurements thananticipated.E.Data Reduction and PresentationAs previously d iscussed, data reduction was attempted in accordance with Eq. (4). In practice, diffi cu lties arise, especially with spectra containing relatively narrowbanded features. It is obvious that if acertain spectral feature appears in both the sample andreference measurement, the ratio of these spectra in theregion of the feature yields a constant; however, if asmall shift in wavelength occurs in either measurement,an artifact is produced. The effect is similar if, for example, Lhe atmospheric absorption changes betweensample and standard measurements. These effects canbe a source of considerable error in spectral measurements and adequately justify dual beam instrumentaltechniques.Approximately fifty spectra we1·e taken in the courseof the experiment. Areas of interest were categori edas image, blood, scorch, and linen.'l'he image areas a re those parts of the cloth containing the anatomical attributes of the figure in Lhecloth. Generally, spectra were taken in those areaswhere the image was visually dense.The blood areas are those regions containing a lightcrimson stain resembling a common bloodstain to somedegree but devoid of its characteristic reddish-browncoloring.1924APPLIED OPTICSIVol. 19, No. 12 / 15 June 198009., 080 ll!53 07060 b 04§030201 0''-'sa.s9951010!51111s121.: s131JS 4WAVFLF NGTH (J m)Fig. 8. Absolute spectral reflectance comparisons of cotton 1rndwhole blood-on-cotlon in 8-14-µm band.Linen refers to those regions containing no visualfeatures and represents samp lings of the background ·or homogeneous base layer of the cloth.DiscussionAtmospheric flucLwiLions and noise precluded thedata from being reduced in accordance with Eq. (4). Asan alternative, a corrective spectral distribution wasdetermined by forcing t he normalized spectra of thefeatureless linen of the Shroud to agree with laboratorydata on linen and then applying the calculated conection factor to the remaining measurements after normalization. This method allows spectral comparisonswiLhin t he above approxima tion, but the absolutemagnitude of the reflectance is lost.As a point of reference several laboratory reflectancespectra are shown in Figs. 3-8. The spectral similaritiesof the samples are quite apparent. 'l'his result is largelytrue in both spectral bands with the exception ofwhole-blood-on-cotton in t,he 8- 14-µm band. Theseresults suggest that surface effects clearly dominate overknown chemical differences in this region of the spectrum. From the data, the fo llowing observations arenoted: Linen and cotton are spectrally similar in the3-5- and 8-14-µm bands, both exhibiting pronouncedfeatures at 4.7 and 1 J µmas shown in Figs. 3 and 6. InFig. 3 a Sadtler KBr standard a bsorption spectrum isshown in comparison with a reflectance measurement.The pronounced feature at 3.0 µ·m appears subdued inLh e cotton reflectance spectrum and is absent in thelinen spectrum. The feature that is present at 4.7 µmin reflectance is nearly absent in absorption. In general,caut,ion is required in spectral comparisons betweenconventional absorption spectra and reflectance spectrabecause of the additional com plications of refractiveeffects.Spectral comparisons of linen and a moderate scorchshown in Fig. 4 display similar features in the 3- 5- and8- 14-µm bands as shown in Fig. 7. In general, scorchF.

. .t- --r - -r.--.- - 1 1- -1. . . . ., ,, . , . ., , , , 1 t0.909oa:.:1001-2.6'2 33.234 3638442 44 46 48525405675885995WAVELENGTH 1ym1Fig. 9. Normalized spectral reflectance comparisons of scorchedlinen with averaged Shroud image and scorch areas in 3- 5-µm band.F/ a:s!::!w0 ,4 0 .3.w 13135I06Il!LOOD ON COTTON0 0 .5wN:JI04030.1013·3234 36384424414Ier'16 J\ .855.2511I 0567WAVELENGTH 1ym).;12.5T022.6·2 81208O?0115ul05 1109 0;'Nv\0.6105r'ig. 10. Normalized spectral reflectance comparisons of scorchedlinen with averaged Shroud image artd scorch areas in 8- 14-µniband.0.9SHROUO BLOOO10WAVELFNGTH {)"")7568 99510105II11512125 13135I WAVtLCN GTll ()"")F ig. 11. Normalized spectral reflectance comparisons of bloodon-colton with Shroud averaged blood areas in 3- 5-iim band.Fig. 12. Normalized spectral reflectance comparisons or wholeblood-on-cotton with averaged Shroud blood in 8- 14-µm band.spectra are invariant with respect to visual intensity,showing nearly identical absolute reflectances in bothspectral bands. Furthermore, there exists almostnegligible spectral variation between scorches and barelinen. I3lood-on-cotton in Fig. 5 has little effect oncotton spectral features in the 3-5-µm band but a pronounced effect in the 8-14-µm band as shoW!l in Fig.However, since the base material spectra of cotton andlinen were quite similar, it was reasoned that comparisons between blood-on cotton and blood on-linen werevalid. These results are mitigated somewhat by thepresumed age of the blood and by the possibility of itshaving undergone a rather largo temperature excursiondue to the fire. The effects of these factors on spectralfeatures are unknown.8.Intercomparison between laboratory data and measurements on the Shroud yielded the following observations: As shown in Figs. 9 and 10, laboratory observations of scorches on linen are similar to scorches onthe Shroud. Also shown is a marked similarity betweenimage and scorch areas in both spectral bands. Bloodcomparisons show marked differences in both bands,the disparity in the 8-14-µm band being quite pronounced as shown in Figs. 11and12. A whole-bloodon-linen laboratory measurement was 1 not available.G.ConclusionsDue to the uncertainties in the data it is not possibleto draw definitive conclusions. The spectral similarityof the image areas to known scorches is noted and isconsistent with observation in terms of color in thevisible region of the spectrum, however; this result is notwiLhout ambiguity since spectral similarities are characteristic of most areas examined as shown by the datain both spedral bands. Shroud blood comparisons with15 June 1980 /Vol. 19, No. 12IAPPLIED OPTICS1925

FLOOOLAMPSDSCANNINGHEADFig. 13. Experimentlll r.onfiguration for IR imaging experiments.known bloodstains show marked differences. lt is notknown if these differences are chemical or surface effects. With regard to the experiment in general, theJPany spedral similarities suggest that surface effectsd ominate over chemistry or composition in this regionof the spectrum and that if chemical differences discernible by spectroscopic Lechniques in these spectralbands exist, they lie well below the limits of sensitivityof the instrumentation described herein. Withoutdrastic improvements in instrumentation, a secondattempt is not recommended.Ill.Thermographic InvestigationsLargely nonquantitative in this application, thistechnique was employed to observe inhomogeneities inthe image attributable to differences in lR emissivitiesnot otherwise detectable in the visible region. In addition to imaging in t he lR spectral bands, it providesthe add it ional advantage of observations over a fairlywide FOV.Consider a generalized LR deteclor in a uniform envirnnment at a constant temperature T e and a bodyunder observation at temperature Tb. 'vVe seek thechange in equivalent temperature that results from agiven change in emissivity. The total flux incident onthe receiver can be written asAucaTt B(l - EnkTt.Ewhere 1Cb(5)J: ab(A)lo(.A)d . . ouTt.6E uA(Eb1To1 4 - Eb2To2·1) - Bu7' (Eo1 - Etz). ·6E uA {arbolbWe demand the condition4 A art - BuT,.iJEoEarbofbIt follows that(6)APPLIED OPTICS/ Vol. 19, No. 12 / 15 June 1980[c:rb1P. - at2(X))Jo(X)d .}where the radiation contrast between two bodies is related to the spectral absorptivity. If I o(A.) is large, thesecond term containing the background effect can bedropped. Furthermore, if 10 (A.) is limited to the visibleportion of the spectrum and a(ft.} is replaced by 1 p(ft.), where p(A) is the spectral reflectivity, Eq. (9) becomesfo"[Pb2(:\) - Pb1(X))fo(.\)d.\,(10)which is recognized as au expression describing contrastin I.he visible region of the spectrum bet.ween two bodiesof spectral reflectivity 1 1(.A) and P2(ft.). We concludethat the image observe'd in the IR region when illuminated with a source of strong visible radiation is approximately the image observed in the visible regionwith reversed contrast. We sh ow in the following section that this result is experimentally verified within thelimits of instrument resolution.A.fiE -6TB -6 b·J: (9)t:.E 1AiJ/ (8)Substitution of Eq. (7} into Eq. (8) yields constant;iJE 4A TEnTi,(7)where ab (ft.) spectral absorptivity of the body.If two such bodies are within the FOV, t he flux difference or relative contrast may be written from Eq. (5)as total ernis&ivity of body; andA,B factors associated with geometry.The first term in th e expression represents the selfemission term of the body and the second, the reflectedbackground radiation. Differentiation with respect to'l'b and Cb yields1926yielding the equivalent change in temperature due toa relative change in emissivity. IL is clear that for largedifferences in 7'1, and Te, the sensitivity to a cha nge inemissivity is correspond ingly increased.If the temperature of a body is raised by illuminatingwith a source of radiation of intensity IoCXJ, some of theradiation is thermalized, and the following equality isan expression of thermal equilibrium:ExperimentFigure 13 depicts the experimental configuration.Imaging was accomplished in the 3-5- and 8-14-µmbands with thermographic scanning cameras. Thesource of illuminat ion was two 1500·W photographic

Fig. H. 'F'ace region in 8-14-µm band. Features ohserved correspond closely to those observed in the visible v. ith reversedcontrast.Fig. 16. As in Fig. 15 with image expansion.Fig. 15.As in Fig. 14 with slighUy less contrast and scale lightsturned off.F'ig. 17. Back of head in 14-11m band. Irregularly shaped bright.erareas in upper part of photo correspond to red crimson stains invisible.15 June 1980 / Vol. 19, No. 12 /APPLIED OPTICS1927

Fig. 18. Hands in 8-14-µm band. Bright spot in upper left of photograph corresponds to red crimson stain on wrist in visible.Fig. 19. Foot area in 8-14-µm band.Fig. 20. Chest wound in 8-14-µm band. Bright patch correspondsto large crimson stain in visible. Irregular darker area correspondsto cloth patch sewn over a burned region from 1532 fire.floodlamps, which when f'ocused provided approximately uniform levels of illumination across the observed region. No contrast was discernible without thefloodlamp illumination indicating that the emissivitydifferences in various features on the Shroud were belowthe lirn it of sensitivity of the cameras at room temperature. Witli illumination considerable contrast wasnoted in the 8-14-µm band as shown in l 'igs. 14- 20. Ano inal temperature span from black to white levelswas l.75 C. No features were observed in the 3-5-µ.mband regardless of illumination. 'l'his result is attributed to differences in instrument apertures and hencebasic sensitivities rather than physical attributes of thesurface. By observing the reflected arc lamp illumination from a gold-plated diffuse standard it was ascertained that no detectable radiation in the 8- 14-µm1928APPLIED OPTICS/ Vol. 19, No. 12 I 15 June 1980

Fig. 21.Cold reflectance standard illuminated with floodlamps. Black level indicative of no detectable 8-14-µ m radiation fromfloodlamps.band was given off by the lamps. As shown in Fig. 21the standard appears totally black with nominal controlsettings. Regions observed included the face, sidewound, back of head, hands, and feet with contrastpolarity such that the warmer features appearbrighter.B.DiscussionThe lack of contrast with no illumination is attribu tedto a combination of (!::.E)/E « I and T e Tb [referencedto Eq. (6)) necessitating an increase in target temperature in excess of the background temperature to attainacceptable contrast levels. This results in similar features showing negligible differences in absolute reflectivities. The general character of the illuminated imagery reflects the validity of Eq. (IO) in that it appearsmuch like a black-and-white negative print. It is notedthat the featureless linen background appears black inthe IR rendition. The crimson stains evident on platesof the side wound, hands, face, and feet appear relativelybright as opposed to the visible appearance. Scorch andimage areas lie intermediate between the two. The ffiimagery is a reversed approximate replica of the imageobserved in the visible region with no inhomogeneitiesor artifact apparent to the authors. Since the inherentresolution of the IR camera is poorer than the highquality photographic imagery, this result can be questioned, however; if such an artifact exists, it lies belowthe limit of resolution obtainable with this instrumentation.C.ConclusionsWe have shown that emissivity differences in variousfeatures of the Shroud are too small to yield recognizable images in t he 3-5- or 8-14-µm band with instruments of temperature sensitivity on the order of !::.T 0.5 C or less. This result is consistent with compari-sons of laboratory reflect ance values of similar materials. · With artificial uniform visible illumination goodimagery was observed in the 8-14-µm band, however;this illumination scheme yields an image closely correlated to the image obtained in the visible region withreversed contrast. This result is consistent with theoretical considerations. With due regard to the limitsof instrument resolution and sensitivity, it is the authors' opinion that no significant anomalies exist;however, we leave the final interpretation to thosecompetent in these matters.The authors wish to thank P. Rinaldi, A. Otterbein,and L. Gonella for their efforts in making this projectpossible. We gratefully acknowledge the invaluableadvice and assistance from our many fellow investigators on this project and to those individuals and corporations whose financial contributions provided sorelyneeded support.This work was sponsored by Shroud of Turin Researc.h. roject, Inc. It was substantially accomplishedwhen both authors were with C.S. Air Force WeaponsLaboratory, Albuquerque, New Mexico. Affiliationwith the authors's present employers is coincidental,and no sponsorship or endorsement is implied.References1. E. J. Jumper and R. \V. '.\1etlem, Appl Opt. 19, 1909 (1980).2. For a rather vague reference to possible parallels see EncyclopedinBritannica (U. Chicago, 1979), Vol. 14, p. 1085.3. N. L. Alpert, W. E. Keiser, and H. A. Symanski, IR: Theory andPractice of J.r.:frared Spectroscopy (Plenum, New York, 1970), pp.7, 8.4. W. Wolfe, U. Arizona; private communication.5. W. Wolfe and C. Zissis, The Infrared Handbook (EnvironmentalResearch Institute of Michigan, Ann Arbor, 1978), pp. 3-84-3154.6. S. Pellicori, Appl. Opt. this issue 19, 191 (1980).15 June 1980 /Vol. 19, No. 12 /APPLIED OPTICS1929

Reflectance spectroscopy in the 3-5-and 8-14-µm bancls was attempted in situ using commercial equipment with mo l· crate success. Spectral comparison are made between laboratory reflectance dul.u mid selected Shroud fea tures. Infrared therruographic imaging was accomplished with an enhauced contrast technique using exter nal illumination.