koorio.com
海量文库 文档专家
当前位置:首页 >> 物理 >>

2004Development of Coded-Aperture Imaging with a Compact Gamma Camera_图文

Development of Coded-Aperture Imaging with a Compact Gamma Camera
M. Gmar, O. Gal, C. Le Goaller, O. P. Ivanov, V. N. Potapov, V. E. Stepanov, F. Lainé, F. Lamadie

Abstract--In the 1990’s, CEA has developed one of the most compact operational gamma cameras in the world, CARTOGAM, capable of mapping the surrounding radioactivity and representing it superimposed onto a visible-light image. Since, the camera has been successfully industrialized and commercialized. In the same time, the Kurchatov Institute has developed a codedaperture gamma camera. The two teams have joined their efforts to develop a compact coded-aperture gamma camera. This paper presents the first results obtained with a coded mask adapted to the existing CARTOGAM camera. According to preliminary simulations compact coded-aperture gamma camera could enable Cs-137 or Co-60 source imaging. Innovative masks have been achieved that lead to the following results: a significant increase in the sensitivity (up to 10 times), an acceptable angular resolution (2° to 3°) and a large field of view (about 30°). The tungsten-alloy masks were produced and the adaptation of the camera was carried out. The laboratory results of Cs-137 and Co60 imaging, including shadowgrams and reconstructed images for point sources, are presented.
I. INTRODUCTION
Gamma imaging is of interest in nuclear industry for decontamination or decommissioning purposes, because it enables to locate sources remotely and rapidly. Hence, it can lead to a significant reduction of the dose received by the operating personnel, as well as a reduction of the operating costs. Over the last few years, gamma cameras have been developed in different laboratories for these applications [1-5]. Prototypes have been qualified and operated on various sites.
The simplest means to form a gamma image on the detector is to use a pinhole collimator. But, recently, some teams have demonstrated the possibility of improving the sensitivity of such devices by using a coded aperture of small size [6].

In the 1990’s, CEA has developed one of the most compact operational compact gamma cameras in the world, CARTOGAM (Fig. 1), capable of mapping the surrounding radioactivity and representing it superimposed onto a visiblelight image [7]. Since, the camera has been successfully industrialized and commercialized.
In the same time, the Kurchatov Institute has developed a coded-aperture gamma camera [8]. The two teams have joined their efforts to develop a compact coded-aperture gamma camera by adapting a coded mask to the existing compact pinhole camera CARTOGAM. This paper presents the first results obtained with this camera.
II. GAMMA IMAGING WITH A CODED APERTURE
A. CARTOGAM: a Compact Gamma Camera
The CARTOGAM portable camera (Fig. 1) has been developed for gamma imaging applications in nuclear facilities. Its most remarkable characteristics are its mass – only 15 kg for the detection head, including the shielding – and its size – 8 cm in diameter and 40 cm in length. This compactness allows an easy handling.
The detector comprises a scintillator (CsI:Tl, 4 mm), an image intensifier and a CCD. In the current configuration, two pinholes are available, with 30° and 50° fields of view. A more detailed presentation of the camera and its performance can be found in [9] and [10].

Manuscript received October 29, 2003. This work was supported by the B01-5.12 Common-Interest Program between CEA and COGEMA, and by the INTAS Grant No. 01-0401.
M. Gmar, O. Gal, B. Dessus, and F. Lainé are with the CEA-DRT/LIST, CEA/Saclay, Gif-sur-Yvette Cedex 91191, France (e-mail: mehdi.gmar@cea.fr).
C. Le Goaller is with the CEA-DEN/DDCO, CEA/Valrh?, BP 171 Bagnols-sur-Cèze Cedex 30207, Fance (e-mail: christophe.legoaller@cea.fr).
O.P. Ivanov, V.N. Potapov, and V.E. Stepanov are with Kurchatov Institute, 1 Kurchatov Sq., Moscow 123182, Russia (e-mail: olegivanov@mail.ru).

Fig. 1. The CARTOGAM gamma-imaging system, with its detection head placed on a photographic tripod, in front of the control PC (laboratory configuration; photo CEA/Gonin).

0-7803-8257-9/04/$20.00 ? 2004 IEEE.

1052

B. The Coded Masks
A set of three masks has been designed and manufactured. The pattern of each mask is a hexagonal uniformly redundant array (HURA). The innovative points are the size and the hexagonal shape of holes (Table I). By construction, the geometrical transparency of each mask is about 50%. The open area of the central pattern is 1.9 cm?. Fig. 2 shows the masks, one of rank 6 and one of rank 9. In this paper, we report the first results obtained with these masks, which thicknesses are 6 and 12 mm.

Fig. 4(a) and 5(a) show the reconstructed images of 137Cs and 60Co point sources with short exposure times (respectively
2.2 s and 4.4 s). As shown in Fig. 4(b) and 5(b), increasing the
exposure time reduces the background level.

(a)

(b)

Fig. 4. Reconstructed images of a 600 MBq 137Cs source seen from a distance of 2 m (13 ?Gy/h), with two exposure times: (a) 2.2 s and (b) 30.8 s.

(a)

(b)

Fig. 2. Photos of HURA coded masks (a) of rank 6 (12 mm-thick) and (b) of rank 9 (6 mm-thick).

TABLE I CHARACTERISTICS OF THE MASKS

Mask Thickness Distance between
holes Number of holes in
central pattern Opening area

R6 12 mm 1.85 mm
64 1.9 cm?

R9 6 mm 1.26 mm 136 1.9 cm?

(a)

(b)

Fig. 5. Reconstructed images of a 54 MBq 60Co source seen from a distance of 2 m (4 ?Gy/h), with two exposure times: (a) 4.4 s and (b) 30.8 s.

B. Signal-to-Noise Ratio
An important figure of merit for image quality is the signalto-noise ratio (SNR) of the image for given dose rate, exposure time and spatial resolution. In a first method, this ratio is estimated by the peak height divided by the standard deviation of the background surrounding the peak.

40
C. The Decoding Software

A specific software has been developed by the Kurchatov

30

Institute team to decode the raw images. A complete

description of algorithm can be found elsewhere [11]. Fig. 3

20

shows a raw image and the corresponding reconstructed image.

10

signal-to-noise ratio

0

0

200

400

600

exposure time (s)

(a)

(b)

Fig. 3. Image of a 137Cs point source obtained with the mask of rank 6: (a) raw image and (b) corresponding reconstructed image.

III. EXPERIMENTAL RESULTS
A. Sensitivity Obviously, the first improvement that we could expect is an increase in the camera sensitivity, because the opening area of the coded aperture is greater than the opening of the pinhole.

Fig. 6. Signal-to-noise ratio versus exposure time, measured by the first method (see text) for a 13 ?Gy/h 137Cs point source (mask of rank 9).
Fig. 6 shows the evolution of the signal-to-noise ratio versus exposure time, measured by this first method for a 13 ?Gy/h 137Cs point source (mask of rank 9). This practical measurement shows a correct increase in √texp with increasing exposure time texp up to about one minute, afterwards saturation is observed. This saturation is due to the systematic local variations of the background level introduced by the decoding procedure, which become predominant in the

0-7803-8257-9/04/$20.00 ? 2004 IEEE.

1053

standard deviation when the statistical fluctuations are low. The SNR law is in the form

( ) S N = α texp 1+ α 2 β 2 ? texp ,

(1)

where α accounts for the slope of the initial square-root law and β for the asymptotic limit.

We tested a second method to estimate the SNR, where the noise is defined as the standard deviation of the difference between two identical successive images, divided by √2. As shown in Fig. 7, systematic contributions in the noise are almost removed by this latter method, and a square-root law is observed until long exposure times.

D. Field of View Fig. 9 shows the linearity of the peak position in the decoded image versus the angular position of the source. This curve gives the angular magnification of our imaging device. As we know the limit in pixels of the field of view in the reconstructed image, we can deduce the angular field of view. We obtain a field of view of 26° for the mask of rank 6, and 27° for the mask of rank 9.
800
600

pixels

160

signal-to-noise ratio

120

80

40

0

0

200

400

600

exposure time (s)

Fig. 7. Signal-to-noise ratio versus exposure time, measured by the second method (see text) for a 13 ?Gy/h 137Cs point source (mask of rank 9).

C. Sensitivity versus Angle
The sensitivity of the detector is not uniform throughout the image area. This is due to the photocathode response and, secondly, to the fiber-optic tapers. This non-uniformity can be corrected in the decoding procedure by dividing the raw image by a reference image obtained with the bare detector under flood exposure. Otherwise it may have a contribution in the background level of the decoded image, independently of the source angular position. Nevertheless, it can be observed that the response of the camera is not uniform with angular position of the source (Fig. 8). This is due to a simple geometrical effect that reduces the photon flux passing through the holes when the source angle is increased.

normalized peak height

1

0,8

0,6

0,4

R6 60Co

0,2

R6 137Cs

0

-15 -10 -5

0

5

angle (°)

10 15

Fig. 8. Sensitivity of the camera versus angular position of a point source, for two gamma energies (mask of rank 6).

400

200

y = 22,297x + 392,58

R2 = 0,9992

0 -15,0 -10,0

-5,0 0,0

5,0

angle (°)

10,0 15,0

Fig. 9. Position of the peak in the decoded image versus angular position of the point source.

E. Angular Resolution
A section of the peak through the maximum is shown in Fig. 10 for the image of a 137Cs point source. We use the full width at half maximum (FWHM) of the peak to calculate the resolution. As a first approximation we consider that two point sources can be separated only if the distance between the two peak maxima is equal or greater than the FWHM. Table II shows the FWHM values for the two masks and two different sources.

300

Normalized intensity

250

200

150

100

50

0

0

200

400

600

800

pixels

Fig. 10. Reconstructed image of a 137Cs point source and corresponding

profile (mask of rank 6).

TABLE II
RESOLUTION OF THE IMAGE EXPRESSED AS THE FWHM OF THE PEAK
(IN PIXELS AND IN DEGREES) FOR MASKS OF RANK 6 AND 9 AND FOR A 137CS OR 60CO POINT SOURCE

Rank 6 Rank 9

137Cs

60Co

pixels degrees pixels degrees

66±5 3 64±5 3

49±3 2.2 60±8 2.6

0-7803-8257-9/04/$20.00 ? 2004 IEEE.

1054

F. Background Reduction
A very interesting property of a HURA mask is the fact that it can be made antisymetric by 60° rotation. By difference of two raw images, one in the mask position and the other in the antimask position, we can, in principle, subtract the background, i.e. the part of the signal that is not modulated by the mask.
To test this property, we placed a 137Cs point source at 7 m from the camera and a 60Co point source, laterally, to create a high background, as shown in Fig. 11. The resulting decoded image is shown in Fig. 12, either without or with the maskantimask procedure. In the latter case, a reduction of the background level is observed since the contrast passes from 83% to 98% and the SNR from 6.9 to 28.1, for the same total exposure time.

G. Low Dose Rate and Limit of Detection
We carried out first attempts to estimate the detection limit of the camera with a coded aperture. A source of 137Cs was placed at 20 m in front of the camera. The dose rate at the camera location was 120 nGy/h. Fig. 15 demonstrates that this dose rate greatly exceeds the detection limit of the camera in ten minutes.

1 ?Gy/h

~7m

16 ?Gy/h ~1m

60Co 53 MBq

137Cs 600 MBq

Fig. 11. High-background experimental setting, with a weak, on-axis 137Cs source and a strong, lateral 60Co.

Fig 14. Reconstructed image of a 600 MBq 137Cs source at 20 m from the camera (120 nGy/h) obtained in 10 min (mask-antimask procedure used, mask of rank 6).
We tried to meet the detection limit by reducing the time of exposure. The image of Fig. 16, obtained in 2 min, is close to the detection limit (120 nGy/h in 2 min).

Fig. 12. Reconstructed images of a 600 MBq 137Cs source at 7 m from the camera in a 60Co high background, obtained in 6 min (total exposure time), without (left) or with (right) mask-antimask procedure (mask of rank 6).
As shown in Fig. 13, in the same conditions, the pinhole camera did not detect the source.
Fig. 13. Image obtained with the 30° pinhole collimator in the same conditions as Fig. 12 (10 min of exposure). The 137Cs source is not detected.

Fig. 15. Reconstructed image of a 600 MBq 137Cs source at 20 m from the camera (120 nGy/h) obtained in 2 min (mask-antimask procedure used, mask of rank 6).
IV. CONCLUSION
In this article, we described the first results obtained with the CARTOGAM compact gamma camera fitted with a coded mask. The experimental measurements are still in progress, but we can draw the first conclusions as follows. The sensitivity of the camera is greatly increased and the next step will be to precisely quantify this increase by comparison with the pinhole configuration. The angular resolution is good (2° to 3° for a 30° field of view). We also tested the background reduction by mask-antimask procedure. This reduction is efficient and allows to image low-dose-rate sources in short time.
Now, we are intending to carry out a systematic and quantitative comparison between coded aperture and pinhole performance. During the year 2004, we have also planned to perform on-site image acquisition in a nuclear facility at Mol, in Belgium, in the framework of a collaboration with SCKCEN.

0-7803-8257-9/04/$20.00 ? 2004 IEEE.

1055

V. REFERENCES
[1] C. Le Goaller, G. Imbard et al., “The development and improvement of the Aladin gamma camera to localise gamma activity in nuclear installations,” European Commission, Nuclear Science and Technology EUR18230 EN, 1998.
[2] A.N. Sudarkin et al., “High-energy radiation visualizer (HERV): A new system for imaging in X-ray and gamma-ray emission regions,” IEEE Trans. on Nucl. Sci., vol. 43(4), pp. 2427-2433, 1996.
[3] AIL, GammaCam - http://www.ail.com/page13_gammacam.htm [4] RMD, RadCam - http://www.rmdinc.com/production/RadCam.html [5] A.N. Sudarkin, O.P. Ivanov, V.E. Stepanov and L.I. Urutskoev,
“Portable gamma ray imager and its application for the inspection of the near-reactor premises contaminated by radioactive substances,” Nucl. Instrum. Meth., vol. A 414, pp. 418-426, 1998. [6] P.T. Durrant, M. Dallimore, I.D. Jupp and D. Ramsden, “The application of pinhole and coded aperture imaging in the nuclear environment,” Nucl. Instrum. Meth., vol. A 422, pp. 667-671, 1999.

[7] O. Gal, C. Izac, F. Jean, F. Lainé, C. Lévêque, A. Nguyen, “CARTOGAM – a portable gamma camera for remote localization of radioactive sources in nuclear facilities,” Nucl. Instrum. Meth., vol. A 460, pp. 138-145, 1999.
[8] O.P. Ivanov, A.N. Sudarkin, V.E. Stepanov and L.I. Urutskoev, “Portable Digital X-Ray and Gamma-Ray Imaging with Coded Mask Performance Characteristics and Methods of Image Reconstruction,” Nucl. Instrum. Meth., vol. A 422, pp. 729-734, 1999.
[9] O. Gal, F. Jean, F. Lainé, C. Lévêque, “The CARTOGAM portable gamma imaging system,” IEEE Trans. Nucl. Sci., vol. 47, no. 3, pp. 952956, June 2000.
[10] O. Gal, B. Dessus, F. Jean, F. Lainé, C. Lévêque, “Operation of the CARTOGAM portable gamma camera in a photon-counting mode,” IEEE Trans. Nucl. Sci., vol. 48, no. 4, pp. 1198-1204, Aug. 2001.
[11] O.P. Ivanov, “Control and image decoding software for portable gammaray imaging system with coded aperture,” IEEE NSS-MIC, conference record, Seattle (WA), October 26-28 1999.

0-7803-8257-9/04/$20.00 ? 2004 IEEE.

1056


网站首页 | 网站地图
All rights reserved Powered by 酷我资料网 koorio.com
copyright ©right 2014-2019。
文档资料库内容来自网络,如有侵犯请联系客服。3088529994@qq.com