OSA's Digital Library

Optics Express

Optics Express

  • Editor: Andrew M. Weiner
  • Vol. 21, Iss. 7 — Apr. 8, 2013
  • pp: 8126–8134
« Show journal navigation

Sub-surface terahertz imaging through uneven surfaces: visualizing Neolithic wall paintings in Çatalhöyük

Gillian C. Walker, John W. Bowen, Wendy Matthews, Soumali Roychowdhury, Julien Labaune, Gerard Mourou, Michel Menu, Ian Hodder, and J. Bianca Jackson  »View Author Affiliations


Optics Express, Vol. 21, Issue 7, pp. 8126-8134 (2013)
http://dx.doi.org/10.1364/OE.21.008126


View Full Text Article

Acrobat PDF (4813 KB)





Browse Journals / Lookup Meetings

Browse by Journal and Year


   


Lookup Conference Papers

Close Browse Journals / Lookup Meetings

Article Tools

Share
Citations

Abstract

Pulsed terahertz imaging is being developed as a technique to image obscured mural paintings. Due to significant advances in terahertz technology, portable systems are now capable of operating in unregulated environments and this has prompted their use on archaeological excavations. August 2011 saw the first use of pulsed terahertz imaging at the archaeological site of Çatalhöyük, Turkey, where mural paintings dating from the Neolithic period are continuously being uncovered by archaeologists. In these particular paintings the paint is applied onto an uneven surface, and then covered by an equally uneven surface. Traditional terahertz data analysis has proven unsuccessful at sub-surface imaging of these paintings due to the effect of these uneven surfaces. For the first time, an image processing technique is presented, based around Gaussian beam-mode coupling, which enables the visualization of the obscured painting.

© 2013 OSA

1. Introduction

Summer 2011 saw the first use of a portable terahertz (THz) imaging system in an archaeological setting to try and image obscured wall paintings. Çatalhöyük in the Anatolia region of Turkey became a UNESCO World heritage site in 2012 [1]. The site contains Neolithic domestic dwellings, the walls of which were repeatedly re-plastered over their lifetime to remove soot damage and on occasion, between re-plastering, these walls were painted. Hundreds of years of human habitation can be chronicled in up to 10 cm of stratified wall plaster. The periodicity of the paintings within the plaster layers, and their subject matter, gives archaeologists insight into the life events of the occupants. However, there is often little or no indication that these paintings are present. Identifying a possible location of wall paintings and exposing them is time consuming, arduous and sometimes fruitless. In recent years THz technology has been developed and applied to the analysis of cultural heritage, and in particular the task of trying to image mural paintings that have for some reason been obscured by covering layers of plaster in their lifetime [2

2. J. B. Jackson, J. W. Bowen, G. Walker, J. Labaune, G. Mourou, M. Menu, and K. Fukunaga, “A survey of terahertz applications in cultural heritage conservation science,” IEEE Trans. THz Sci. Technol. 1(1), 220–231 (2011).

]. The development of portable THz technology has made the identification of obscured pigment layers and reconstruction of sub-surface images possible in the in situ environment [3

3. B. Sartorius, H. Roehle, H. Künzel, J. Böttcher, M. Schlak, D. Stanze, H. Venghaus, and M. Schell, “All-fiber terahertz time-domain spectrometer operating at 1.5 microm telecom wavelengths,” Opt. Express 16(13), 9565–9570 (2008). [CrossRef] [PubMed]

, 4

4. J. V. Rudd, D. Zimdars, and M. Wannuth, “Compact fiber-pigtailed terahertz imaging system,” Proc. SPIE 3934(3934), 27–35 (2000). [CrossRef]

]. This technology will radically change the working approach of archaeologists to the research and conservation of such paintings on site.

At Çatalhöyük, time-domain terahertz spectroscopy (THz-TDS) measurements were conducted on a 9000 year-old mural painting (Building 80-East Wall, Platform F.5014, 18918) decorating the wall above a burial site [5

5. S. Farid, Çatalhöyük 2011 Archive Report (2011), pp. 29–30.

]. Initially, the mural was completely covered by white clay plaster [6

6. I. Hodder, Çatalhöyük 2010 Archive Report (2010), pp. 3–4.

]. During its cleaning, traces of red paint were observed five to twenty layers deep, so the white plaster was painstakingly removed with scalpels to reveal the design. The geometric composition appears to be of such significance that from exposed sections it was evident that the same pattern was repeated through numerous cycles of the wall’s maintenance. Additionally, the exposed pattern resembles that of Çatalhöyük Shrine VIA.50, discovered in 1965, which was painted in red ochre and cinnabar [7

7. J. Mellaart, “Excavations at Çatal Hüyük, 1965: fourth preliminary report,” Anatolian Studies 16, 165–191 (1966). [CrossRef]

]. This made the still obscured sections of the painting intriguing test subjects for the use of THz imaging for the visualization of sub-surface wall paintings. However, as the original wall surface was uneven and non-uniform due to the mud brick support and its construction using primitive tools, all the subsequent refinishing left an increasingly uneven surface profile. This paper presents the first beam coupling-based algorithm developed to overcome the effect of non-uniform surfaces on sub-surface identification and imaging.

The excavating archaeologist identified a section of the wall painting where a single red ochre line was covered by plaster in the middle but exposed at each end. While it cannot be proved beyond doubt that the line continued under this section, it was the experience of the on-site archaeologist that it would. This made this section of the wall painting an ideal test area for the technique. The identified section is shown in Fig. 1(a)
Fig. 1 (a). A section of Neolithic wall with a line section covered with uneven plaster, imaged section highlighted. Figure 1. (b). The “pitch-catch” experimental configuration.
, where the dotted red line indicated the trajectory of the obscured paint layer. Figure 1(b) shows the Picometrix T-Ray 4000 system recording data on site using a “pitch-catch” configuration. This particular section of the wall should, in principle, have been an ideal test bed for the technique. Contrast in a terahertz image is determined by a change in the material properties across the sample being scanned, all other variables remaining constant across the image range. In this case a layer of pigment within the plaster reflects the THz signal, due to a difference in refractive index between pigment and plaster [8

8. J. B. Jackson, M. Mourou, J. F. Whitker, I. N. Durling III, S. L. Williamson, M. Menu, and G. A. Mourou, “Terahertz imaging for the non-destructive evaluation of mural paintings,” Opt. Commun. 281(4), 527–532 (2008). [CrossRef]

]. However, over this particular section of the wall, there was a raised surface profile which deflected the beam away from the detector in an unquantified way during the raster scan, making comparative spectroscopic analysis across the image surface impossible. Therefore traditional THz image reconstruction proved inadequate to retrieve the sub-surface line. A signal processing technique was developed that corrected for the defocusing of the beam as it was scanned over the uneven wall surface and the loss of signal at the detector due to deflections from the uneven surface deviating the beam. The image created following this correction is compared to that constructed using traditional THz imaging parameters, and is shown to reveal the obscured line painting.

2. Theory

In order to correct for the uneven wall surface deflecting the beam from the detector in the “pitch-catch” configuration (see Fig. 2(a)
Fig. 2 (a). A schematic diagram showing how the beam is deflected from the aligned detector position as the image is raster scanned. Figure 2. (b). A diagram showing the co-ordinate system used to calculate the correction.
), a three-dimensional surface profile for the wall was created from the THz time domain data. The first pixel was aligned to maximize the measured signal and this defined a zero z (depth profile) plane. This was assumed to be the focal plane of the experimental configuration. Deviation from this plane in the z direction was defined by the difference in the time delay between each pixel and the pixel defining the focal plane, enabling a depth map to be created for the wall surface. Differential gradients were determined in the x and y directions of the image through analysis of the time delay between adjacent pixels in each dimension, mapping the local angular displacement suffered by the beam across the wall surface. The relationship between the Cartesian co-ordinate system and the spherical coordinate system used to describe the deflection of the beam is illustrated in Fig. 2(b).

The incident THz beam was assumed to be a fundamental Gaussian beam-mode. For each pixel, using the surface profile data, the resulting propagation of the beam to the wall surface, either in advance of or beyond the theoretical focal plane, and its deflection from the optimum line of reflection were calculated. These misalignments were used to evaluate the overlap integral describing the coupling between this displaced reflected beam and the detector, based on the work of Kogelnik [9

9. H. Kogelnik, “Coupling and conversion coefficients for optical modes,” Proceedings of the Symposium on Quasi-Optics, Microwave Research Institute Symposia Series 14, (Polytechnic Press, 1964), pp. 333–347, (1964).

] and Joyce and DeLoach [10

10. W. B. Joyce and B. C. DeLoach, “Alignment of Gaussian beams,” Appl. Opt. 23(23), 4187–4196 (1984). [CrossRef] [PubMed]

], and this used to correct for the percentage of radiation deviated from the detector due to the uneven surface of the wall.

The beam displacements were calculated from the time domain data as follows. For a parallel displacement of the wall surface d behind the focal plane and for bistatic angle α between the source and detector, the lateral and longitudinal deflections of the beam were derived using the schematic diagram Fig. 3
Fig. 3 Schematic diagram illustrating the derivation of the longitudinal and latitudinal deflections.
., and are expressed in Eq. (1) and Eq. (2).

The lateral shift is given by:

Δlat=dcos(α/2)sinα
[1]

The longitudinal shift is given by:

Δlong=dcos(α/2)(1+cosα)
[2]

For a given tilt of the wall surface β in the φ direction (in the plane of incidence) and γ in the θ direction (perpendicular to the plane of incidence), similar considerations show that the overall round-trip time for the THz pulse to travel from the source to the detector, combining the effect of tilts β and γ with displacement d, is:
t=1c[d1+dcos(α/2)+d1cos2βcos2γ+dcosαcos2βcos2γcos(α/2)]
[3]
where c is the speed of light in free space and d1 is the focal distance from the source and detector lenses (here assumed to be identical) and the focal plane.

For an optimally coupled, undeviated beam, the round-trip time would be:

t0=2d1c
[4]

The tilts β and γ are obtainable directly from the local differential gradients at the wall surface. Therefore, for each pixel, the time difference tt0 can be solved using Eqs. (3) and (4) to find the displacement d, which, in turn, can be used in Eqs. (1) and (2) to find the lateral and longitudinal shifts.

In order to determine the image correction that is needed, these shifts and tilt angles are incorporated into a beam coupling integral [9

9. H. Kogelnik, “Coupling and conversion coefficients for optical modes,” Proceedings of the Symposium on Quasi-Optics, Microwave Research Institute Symposia Series 14, (Polytechnic Press, 1964), pp. 333–347, (1964).

,10

10. W. B. Joyce and B. C. DeLoach, “Alignment of Gaussian beams,” Appl. Opt. 23(23), 4187–4196 (1984). [CrossRef] [PubMed]

], which is evaluated to determine the amplitude coupling coefficient between the deflected beam and the optimal receive beam for the detector.

The form of the detector beam is defined by the radiation pattern of the detector antenna and its associated optics, which, from the reciprocity theorem, will be a time-reversed version of the beam that would be launched if the detector were replaced in its antenna by a radiating source. In the type of system considered here, both this detected beam and the beam emitted by the source are well approximated by fundamental Gaussian beam-modes; the dimensions of the terahertz optical system being such that it effectively operates as a mode filter, with higher order beam-modes lost through diffraction [11

11. D. H. Martin and J. W. Bowen, “Long wave optics,” IEEE Trans. Microw. Theory Tech. 41(10), 1676–1690 (1993). [CrossRef]

]. Although the Picometrix system uses antenna-based photoconductive generation and detection, similar considerations would apply to systems based around electro-optic generation and detection, with the pump and probe laser beam profiles effectively forming an edge taper across the source and detector crystals.

The coupling integral can be evaluated in any convenient cross-sectional plane in the region in which the deflected beam overlaps the detector beam. The value of the resulting coupling coefficient does not depend on the choice of plane over which it is evaluated. However, as the deflected beam suffers an angular displacement at the wall, it transpires that the mathematical expression for the resulting coupling coefficient is simplest if the coupling coefficient is evaluated in the cross-sectional plane which is normal to the axis of propagation of the deflected beam and contains the point where the axis of the deflected beam intersects the surface of the wall.

The amplitude coupling coefficient takes a value between 0 and 1, indicating by how far the beam has missed the aligned detector position; a coefficient of 1 indicating optimal alignment and 0 meaning the beam totally misses the detector. Because the beam parameters are frequency dependent, it is necessary to calculate the coupling coefficient for each frequency component of each pixel of the image.

The longitudinal and lateral displacements of the beam, and its angular tilt, can be seen as successive modifications of the beam, resulting in the overall beam coupling coefficient, <ΨDS>. Here, Ψ represents the field amplitude of the beam in the cross-sectional plane over which the coupling integral is evaluated, and the subscripts D and S refer, respectively, to the detector receive beam and the source beam.

For the configuration considered here, the emitted and detected beams are nominally identical, with a frequency dependent beam-waist size w0, and are assumed to be fundamental Gaussian beam-modes. In this case, evaluation of the coupling integral yields the amplitude coupling coefficient for the longitudinal shift Δlong:
ΨD|ΨS0={1+(Δlongkw02)2}1/2
[5]
where k = 2π/λ, at wavelength λ. Note that Eq. (5) has been written in terms of the beam-waist size and separation (which is equivalent to Δlong), rather than the beam widths and phase front curvatures in the plane over which the coupling integral is evaluated, resulting in an expression which is independent of this choice of plane, confirming the earlier assertion that the value of the coupling coefficient is independent of the choice of plane over which it is evaluated.

Introducing the simultaneous lateral shift Δlat modifies the amplitude coupling coefficient to become:
ΨD|ΨS1=ΨD|ΨS0.exp(Δlat2/de2)
[6]
where

de2=2{(k2w02)2+(kΔlong)2}k4w02
[7]

Finally, the effect of a small angular displacement of the beam 2β (corresponding to a tilt of the wall surface β) in the plane of incidence, and 2γ (corresponding to a wall tilt of γ) perpendicular to the plane of incidence, can be incorporated to give the overall amplitude coupling coefficient:
ΨD|ΨS=ΨD|ΨS1.|exp[(2β)2/θe2].exp[(2γ)2/θe2]|
[8]
with (from [9

9. H. Kogelnik, “Coupling and conversion coefficients for optical modes,” Proceedings of the Symposium on Quasi-Optics, Microwave Research Institute Symposia Series 14, (Polytechnic Press, 1964), pp. 333–347, (1964).

])
θe2=4k2[1wS2+1wD2+ik2(κSκD)]
[9]
where wS and wD are the beam widths, and κS and κD are the phase-front curvatures, in the cross-sectional plane which contains the point where the axis of the deflected beam intersects the surface of the wall. This point corresponds to the point where the axis of the angularly displaced beam would cross that of a beam displaced laterally by Δlat but without the angular displacement. The beam widths and phase front curvatures at this point are given by:
wS=w0[1+(2zSkw02)2]1/2
[10]
κS=1zS[1+(kw022zS)2]1
[11]
and similarly for the detector beam with the S subscripts replaced with D. The distance zS is the distance between the source beam-waist (i.e. in the focal plane) and the crossing point at the wall surface, along the axis of the incident beam:
zS=dcos(α/2)
[12]
The distance zD is the distance between the detector beam-waist projected onto the axis of the laterally displaced beam and the crossing point at the wall surface:
zD=dcosαcos(α/2)
[13]
The negative sign in Eq. (12) is important because, at the crossing point, one beam is diverging and one beam converging, modifying the sign of the phase-front curvature.

The overall amplitude coupling coefficient (Eq. (8)) was calculated frequency by frequency, pixel by pixel, across the scanned data and used to correct the measured spectra. A new terahertz image was created from the corrected data.

The above treatment does include some assumptions. Firstly, it is assumed that the wall is flat on the length scale of the focussed terahertz spot. Secondly, the effects of refraction at the wall surface have been ignored. Finally, small angle approximations have been made in the derivation of Eq. (8). These assumptions should be a good approximation for the typical small angular deviations of the wall surface that are found in practice. Moreover, it should be noted that the wall surface profile is mapped using the same terahertz beam that is used to produce the image and so the corrected image data automatically matches the surface profile of the wall averaged over the length scale of the terahertz spot.

3. Experimental results

All measurements were taken using a Picometrix T-Ray 4000® (TR4K) THz-TDS system. The major benefits of this system design are that the optical components are contained within a case of easy-to-transport size and weight, and the fiber-coupled antennas permit rapid modification of the measurement geometry, including in the off-axis reflection configuration as seen in Fig. 1(b). The mode-locked, two-stage, amplified, Ytterbium-fiber pump laser operated with a center frequency near 1064 nm, a 100 fs pulse width, a 50 MHz repetition rate and a maximum output power of 400 mW. The emitter and receiver were low temperature GaAs photoconductive antennas. The waveform acquisition rate was 100 Hz with a fixed 320 ps measurement window, with a 0.078125 ps time resolution. Under optimal conditions, the spectral bandwidth of the system in this geometry is 2 THz, with a dynamic range > 40 dB when averaging very few waveforms. However, in the summer months at Çatalhöyük, the daily temperature peaked at over 40° C, exceeding the manufacturer’s recommended operation temperature of 30° C. After this first occurrence, experiments were restricted to between early morning and early afternoon, and the system was stored in a room temperature environment the rest of the time. Transportation, unpacking and setup for measurement took under thirty minutes once the conditions were set. The environment, however, had a long-term impact on the system performance. Due to the heat, the laser power drifted lower during measurements as the site temperature increased, resulting in reduced dynamic range and noise in the time-domain signal. Although the optical components of the system were protected by the case and extra covering, it was impossible to avoid some exposure to the fine, dry dust. The pixel acquisition rate ranged from two to five pixels per second, depending on how much averaging was performed and the electronics’ response to the heat. Thus, measurement dimensions were kept small in order to maintain consistency within a single data set. The acquisition step size was 1 mm by 1 mm, and the dimensions of the section of the wall discussed here is 50 mm by 30 mm. It took approximately 90 minutes to record this image using 100 averages per pixel. Wavelet denoising [12

12. E. Berry, R. D. Boyle, A. J. Fitzgerald, and J. W. Handley, Computer vision beyond the Visible Spectrum (Springer, 2004), Chap. 9.

] was used in the post-processing algorithm, before the beam coupling correction was applied.

Figure 4(a)
Fig. 4 (a). A terahertz image of a wall section from Çatalhöyük calculated by integrating the in the frequency domain from 0.28 THz to 0.34 THz. Figure 4(b) A terahertz image of a wall section from Çatalhöyük calculated by integrating the in the frequency domain from 0.28 THz to 0.34 THz using the corrected data.
shows the traditional terahertz image, the image parameter calculated by integrating in the frequency domain from 0.28 THz to 0.34 THz, compared to the image corrected using the beam coupling algorithm in Fig. 4(b). The lower frequency range was used for a number of reasons, there was better signal to noise at this range due to less scattering through the plaster, additionally the beam waist is larger at the lower frequencies meaning the radiation, in part, is likely to still couple with the detector.

Figure 5(a)
Fig. 5 (a).A photographic image of the wall section imaged. Figure 5(b). The corrected THz image showing a line passing through the expected region highlighted in Fig. 5(a).
shows a photograph of the imaged section while Fig. 5(b) shows the corrected THz image. A line is visible in the corrected THz image in the location expected compared to the photograph.

4. Discussion

The construction of a THz image relies on the source of contrast changing pixel by pixel across the raster scanned region of the sample, and this being the only change across the image range. In the case of an embedded paint layer, the source of contrast should be the obscured paint layer, with reflections from this layer being the cause of change in the reflected signal. In the experimental conditions experienced at Çatalhöyük, straightforward visualization of the THz image was not possible because the embedded paint layer was painted on an uneven surface, and this was obscured by a covering layer, itself with an uneven surface. Signal variation resulting from the reflected beam being deflected from the detector and shifted away from optimum focus introduced an additional source of image contrast, making the reconstruction of an obscured paint layer unclear. This was corrected for in a unique way using a beam coupling algorithm that measured the beam deflection caused by the uneven covering surface by creating a surface profile from the THz time delay signal. The coupling coefficient between the deflected beam and the ideal beam for the ‘pitch-catch’ configuration was used to correct for the uneven signal loss in the experimental image data, on a pixel by pixel and frequency by frequency basis. Figure 4 shows a direct comparison between the traditional THz image and the corrected version. The region within which we are expecting to see a continuous line has a much lower contrast variation in the corrected THz image, and is the size order of magnitude expected and in the location indicated in Fig. 1(a) and Fig. 5(a). This is a real world experiment, with a real world result. We do not know if the line underneath the plaster was painted on a smooth surface or not and it is possible that the features seen in the light blue region below the red line in Fig. 5(b) are a result of surface variation in the subsurface plaster layers in this region. Nevertheless, the algorithm appears to have corrected for signal loss in this region through beam deflection and defocusing at the upper surface. A marked contrast in image quality of the obscured paint line is seen as a consequence in the THz frequency integration image as shown in Fig. 5.

5. Conclusion

The results of the first ever use of pulsed THz radiation to create an image of a sub-surface paint layer in an archaeological site are presented in this paper. Images of sub-surface paint layers are produced using a novel beam coupling algorithm that corrects for the aberrations induced in the unprocessed image as a consequence of an uneven covering plaster layer.

Acknowledgments

The authors would like to thank the AHRC/EPSRC Science and Heritage Programme and the European Commission's 7th Framework Programme project CHARISMA [grant agreement no. 228330].

References and links

1.

http://whc.unesco.org/en/list/1405.

2.

J. B. Jackson, J. W. Bowen, G. Walker, J. Labaune, G. Mourou, M. Menu, and K. Fukunaga, “A survey of terahertz applications in cultural heritage conservation science,” IEEE Trans. THz Sci. Technol. 1(1), 220–231 (2011).

3.

B. Sartorius, H. Roehle, H. Künzel, J. Böttcher, M. Schlak, D. Stanze, H. Venghaus, and M. Schell, “All-fiber terahertz time-domain spectrometer operating at 1.5 microm telecom wavelengths,” Opt. Express 16(13), 9565–9570 (2008). [CrossRef] [PubMed]

4.

J. V. Rudd, D. Zimdars, and M. Wannuth, “Compact fiber-pigtailed terahertz imaging system,” Proc. SPIE 3934(3934), 27–35 (2000). [CrossRef]

5.

S. Farid, Çatalhöyük 2011 Archive Report (2011), pp. 29–30.

6.

I. Hodder, Çatalhöyük 2010 Archive Report (2010), pp. 3–4.

7.

J. Mellaart, “Excavations at Çatal Hüyük, 1965: fourth preliminary report,” Anatolian Studies 16, 165–191 (1966). [CrossRef]

8.

J. B. Jackson, M. Mourou, J. F. Whitker, I. N. Durling III, S. L. Williamson, M. Menu, and G. A. Mourou, “Terahertz imaging for the non-destructive evaluation of mural paintings,” Opt. Commun. 281(4), 527–532 (2008). [CrossRef]

9.

H. Kogelnik, “Coupling and conversion coefficients for optical modes,” Proceedings of the Symposium on Quasi-Optics, Microwave Research Institute Symposia Series 14, (Polytechnic Press, 1964), pp. 333–347, (1964).

10.

W. B. Joyce and B. C. DeLoach, “Alignment of Gaussian beams,” Appl. Opt. 23(23), 4187–4196 (1984). [CrossRef] [PubMed]

11.

D. H. Martin and J. W. Bowen, “Long wave optics,” IEEE Trans. Microw. Theory Tech. 41(10), 1676–1690 (1993). [CrossRef]

12.

E. Berry, R. D. Boyle, A. J. Fitzgerald, and J. W. Handley, Computer vision beyond the Visible Spectrum (Springer, 2004), Chap. 9.

13.

S. Fowler, “Into the stone age with a scalpel - a dig with clues on early urban life,” The New York Times (September 7, 2011).

14.

K. Fukunaga and I. Hosako, “Innovative non-invasive analysis techniques for cultural heritage using terahertz technology,” C. R. Phys. 11(7-8), 519–526 (2010). [CrossRef]

OCIS Codes
(070.0070) Fourier optics and signal processing : Fourier optics and signal processing
(100.0100) Image processing : Image processing

ToC Category:
Imaging Systems

History
Original Manuscript: January 9, 2013
Revised Manuscript: March 7, 2013
Manuscript Accepted: March 10, 2013
Published: March 27, 2013

Citation
Gillian C. Walker, John W. Bowen, Wendy Matthews, Soumali Roychowdhury, Julien Labaune, Gerard Mourou, Michel Menu, Ian Hodder, and J. Bianca Jackson, "Sub-surface terahertz imaging through uneven surfaces: visualizing Neolithic wall paintings in Çatalhöyük," Opt. Express 21, 8126-8134 (2013)
http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-21-7-8126


Sort:  Author  |  Year  |  Journal  |  Reset  

References

  1. http://whc.unesco.org/en/list/1405 .
  2. J. B. Jackson, J. W. Bowen, G. Walker, J. Labaune, G. Mourou, M. Menu, and K. Fukunaga, “A survey of terahertz applications in cultural heritage conservation science,” IEEE Trans. THz Sci. Technol.1(1), 220–231 (2011).
  3. B. Sartorius, H. Roehle, H. Künzel, J. Böttcher, M. Schlak, D. Stanze, H. Venghaus, and M. Schell, “All-fiber terahertz time-domain spectrometer operating at 1.5 microm telecom wavelengths,” Opt. Express16(13), 9565–9570 (2008). [CrossRef] [PubMed]
  4. J. V. Rudd, D. Zimdars, and M. Wannuth, “Compact fiber-pigtailed terahertz imaging system,” Proc. SPIE3934(3934), 27–35 (2000). [CrossRef]
  5. S. Farid, Çatalhöyük 2011 Archive Report (2011), pp. 29–30.
  6. I. Hodder, Çatalhöyük 2010 Archive Report (2010), pp. 3–4.
  7. J. Mellaart, “Excavations at Çatal Hüyük, 1965: fourth preliminary report,” Anatolian Studies16, 165–191 (1966). [CrossRef]
  8. J. B. Jackson, M. Mourou, J. F. Whitker, I. N. Durling, S. L. Williamson, M. Menu, and G. A. Mourou, “Terahertz imaging for the non-destructive evaluation of mural paintings,” Opt. Commun.281(4), 527–532 (2008). [CrossRef]
  9. H. Kogelnik, “Coupling and conversion coefficients for optical modes,” Proceedings of the Symposium on Quasi-Optics, Microwave Research Institute Symposia Series 14, (Polytechnic Press, 1964), pp. 333–347, (1964).
  10. W. B. Joyce and B. C. DeLoach, “Alignment of Gaussian beams,” Appl. Opt.23(23), 4187–4196 (1984). [CrossRef] [PubMed]
  11. D. H. Martin and J. W. Bowen, “Long wave optics,” IEEE Trans. Microw. Theory Tech.41(10), 1676–1690 (1993). [CrossRef]
  12. E. Berry, R. D. Boyle, A. J. Fitzgerald, and J. W. Handley, Computer vision beyond the Visible Spectrum (Springer, 2004), Chap. 9.
  13. S. Fowler, “Into the stone age with a scalpel - a dig with clues on early urban life,” The New York Times (September 7, 2011).
  14. K. Fukunaga and I. Hosako, “Innovative non-invasive analysis techniques for cultural heritage using terahertz technology,” C. R. Phys.11(7-8), 519–526 (2010). [CrossRef]

Cited By

Alert me when this paper is cited

OSA is able to provide readers links to articles that cite this paper by participating in CrossRef's Cited-By Linking service. CrossRef includes content from more than 3000 publishers and societies. In addition to listing OSA journal articles that cite this paper, citing articles from other participating publishers will also be listed.

Figures

Fig. 1 Fig. 2 Fig. 3
 
Fig. 4 Fig. 5
 

« Previous Article  |  Next Article »

OSA is a member of CrossRef.

CrossCheck Deposited