A novel type epithermal neutron radiography detecting and imaging system

Transcription

1 Nuclear Instruments and Methods in Physics Research A 424 (1999) A novel type epithermal neutron radiography detecting and imaging system M. Balaskó, F.Kôrösi, E. Sváb *, I.Eördögh KFKI Atomic Energy Research Institute, POB 49, H-1525 Budapest, Hungary Agricultural University, Go( do( llol, H-2100 Go( do( llol, Hungary Research Institute for Solid State Physics of the Hungarian Academy, POB 49, H-1525 Budapest, Hungary Research Institute for Technical Physics and Material Science, POB 49, H-1525 Budapest, Hungary Abstract The transfer technique is widely used for epithermal neutron radiography (ENR) for making images upon the object to be investigated. We propose to use instead of the photosensitive film a gamma sensitive scintillation screen (NaCe single crystal), that is monitored by a computer controlled low light level TV camera. The exposure time has been reduced to a duration of only a short fraction of that needed for the conventional transfer process. The presented ENR images consist of electronic signals that are handled by an advanced image processing and analyzing program, the Iman 1.4 version, using a task oriented video grabber Elsevier Science B.V. All rights reserved. Keywords: Epithermal neutron radiography; Image acquisition; Image processing; Image analysis 1. Introduction * Corresponding author. Tel.:# /1418; fax:# ; svab@power.szfki.kfki.hu. The main property of the epithermal neutron radiography (ENR) is that it discriminates well between the elements among investigated materials [1]. Particularly, the lower scattering effects in materials having high hydrogen content [2] makes ENR sensitive to hydrogen determination in biological material [3]. So far ENR has been performed with the transfer technique. This is highly time consuming in view of the long and doubled exposition time of long duration [4] and the film imager cannot be easily connected to computerized on-line data-logging. In our case the epithermal images are temporally captured by an epithermal neutron activated indium foil of 0.1 mm thickness. This emmits gamma radiation that is converted by NaCe single crystal into light photons. The latter is visualized by a high sensitivity TV camera controlled by a computer. As a result the image processing time has been shortened considerably. The applied image processing subsystem has been developed in our institute by one of the authors (I.E.) and called Iman. It consists of frame grabber and a high speed image processing card. The input images are sampled in /99/$ see front matter 1999 Elsevier Science B.V. All rights reserved. PII: S ( 9 8 )

2 264 M. Balasko& et al./nuclear Instruments and Methods in Physics Research A 424 (1999) a real time mode and stored in a dual ported video frame buffer. The resolution is programmable from to picture elements (pixels). The default resolution in a standard CCIR 601 video mode makes up pixels, with a pixel aspect ratio of 1:1. The amplification and offset values of the input module can be adjusted by program instructions. In this work we describe the basic features of the applied method and the results of the quantitative analysis of the computer controlled pixel distribution of the epithermal neutron activated In foil. 2. Material and methods Fig. 1. The experimental scheme of the In foil activated epithermal neutron radiography image processing system. The measurements were performed at the dynamic neutron, gamma and X-ray radiography station at the Budapest Research Reactor (10 MW) [5]. The object to be investigated was kept close to the collimator where the neutron flux was 8 10 ncm s. The epithermal beam was obtained by placing a 1 mm thick Cd plate before the sample. By this way the thermal neutrons were filtered from the neutron beam obtained from the radial channel of the reactor. The sample was a 40 mm thick consumption meter made from Al. The epithermally activated In foil was read in an experimental set up scheme schematically shown in Fig. 1. Between the epithermal neutron activated In foil and the NaCe single crystal a 1 mm thick aluminum plate was placed to protect the surface of the single crystal and to absorb the β-radiation emitted by the activated In foil. The imaging time cycle of the low light level (LLL) TV camera from 40 ms was increased to 320 ms utilizing an appropriate software. Using the Iman 1.4 program, 8 frame time was 60 times integrated, grabbed and utilized for the qualitative and semi quantitative analysis. The frame grabber card, FRG2M/95 is equipped with a high speed, bi-directional video bus I/O socket for the linked convolution processor. This processor performs an enormous quantity of calculations by applying the so-called neighbourhood pixel operations up to 8 8 pixels in a real time mode. The software is capable of image acquisition, gray-colour-binary morphometry, image enhancement, transformations and segmentation. The used statistics involved feature histograms, integrated global specimen statistics and profile measurement. For the pixel count distributions spectrum analysis was applied. The In foil was activated by epithermal neutrons for 18 h and the image grabbing of decaying and converted γ photons was carried out at the 16th, 24th, 32nd and 107th min. 3. Results Fig. 2a indicates the beam area of γ photons emitted by the In foil after epithermal neutron activation. This was converted, according to the experimental set-up (Fig. 1) by NaCe single crystal into a gray level image with a shading correction to the NaCe single crystal, before applying In foil. On the right hand bottom corner of the image a shadow of the 40 mm thick consumption meter can be registered. After multiclass segmentation of this image we got an intensity distribution pattern presented in Fig. 2b. At the 16th min the total area of the epithermal neutron beam was as high as

3 M. Balasko& et al./nuclear Instruments and Methods in Physics Research A 424 (1999) Fig. 2. (a) The DNR image of epithermal neutron beam as assessed in the 16th min after In foil activation, on the right bottom corner of the image is the contour of the sample. (b) The surface of different epithermal intensities after multiclass segmentation of (a). (c) 3D intensity histogram of (a). (d) Pixel count distribution of (a). The inserted diagram of (d) is a periodogram of γ photon decay of In foil at the 16th min.

4 266 M. Balasko& et al./nuclear Instruments and Methods in Physics Research A 424 (1999) Fig. 3. (a) The DNR image of epithermal neutron beam as assessed in the 24th min after In foil activation. (b) The surface of different epithermal intensities after multiclass segmentation of (a). (c) 3D intensity histogram of (a). (d) Pixel count distribution of (a). The inserted diagram of (d) is a coherency of γ photon decay of In foil at the 16th min and 24th min.

5 M. Balasko& et al./nuclear Instruments and Methods in Physics Research A 424 (1999) Fig. 4. (a) The DNR image of epithermal neutron beam as assessed in the 32nd min after In foil activation. (b) The surface of different epithermal intensities after multiclass segmentation of (a). (c) 3D intensity histogram of (a). (d) 3D intensity histogram of DNR image as determined in 107th min. (e) Pixel count distribution of (a). (f) Pixel count distribution of DNR image as estimated at the 107th min. The inserted diagram of (e) is a coherency of γ photon decay of In foil at the 16th min and 32nd min. The inserted diagram of (f) a coherency of γ photon decay of In foil at the 16th min and the 107th min.

6 268 M. Balasko& et al./nuclear Instruments and Methods in Physics Research A 424 (1999) &84 cm, with the most intensive portion (dark area) being of &16 cm, i.e., &20% of the total one. The 3D diagram of the intensity distribution is displayed in Fig. 2c. The background intensity values correspond to the most light areas, while the high intensity values show the dark areas of Fig. 2b. The pixel count distribution of the gray scale value, as depicted in Fig. 2d, disclosed a very intensive pixel count distribution above 60 gray level value. In the inserted picture the periodogram of the frequencies, revealed in the pixel count of gray level values, is shown. The horizontal axis represents the frequency while the vertical axis depicts the corresponding periodograms. In tendency the periodogram values oscillate around a periodogram value of 10. The next image was taken at 24th min and the data obtained is presented in Fig. 3 with the same experimental set-up as was done at the 16th min. This time the whole γ photon area, delineating the epithermal neutrons, was compressed to &77 cm (Fig. 3a), with &11 cm intensive portion (Fig. 3b). The latter constituted &13% of the whole area, and decreased &31% in comparison with that one ascertained at the 16th min. The decay of In foil was verified by the values depicted in the 3D diagram (Fig. 3c). These pertained to the whole area and the intensity values. This trend was well reproduced in the intensity histogram (Fig. 3d), too. In the embedded diagram of Fig. 3d the coherency of the gamma photon decays is presented. The coherency values are between 0 and 1. The closer the value to 1, the bigger the coherency between the compared spectra. Between the spectrum as measured in 16th min and that of 24th min in a frequency range of a coherency value of &0.8 was noticed. Fig. 4a and b demonstrate the intensity distribution pattern of the γ photons emitted from the epithermal neutron activated In foil at the 32nd minute. The progression of the decay can be well traced by decreasing the whole surface of the γ photon beams from &84 cm to &70 cm. This intensity distribution of the epithermal neutrons, as detected by our imaging system, also revealed an abatement in the related intensity characterizing surfaces. The dark areas displaying the most intensive portions of the γ photon beam declined considerably compared to those observed at 16th minute after placing the In foil onto the surface of NaCe single crystal converter. It meant &62% reduction. A decline in the γ photon intensity can be scented from the 3D intensity distribution diagram (Fig. 4c). Both the intensity levels and the area depicting the beam of the epithermal neutrons are reduced, at the same time the background area is expanded. The peak of the pixel count distribution histogram was shifted to the lower gray values (Fig. 4e). However, a considerable amount of γ photons activity still remained, as indicated by the pixel count distribution at higher, e.g. above 60, gray levels. According to Fig. 4d by the 107th min the epithermal activation of the In foil totally decayed. The phenomenon was confirmed by the pixel count distribution histogram obtained at the 107th min. All pixel counts below gray level 30 represented background values (Fig. 4f). A good coherency appeared in the γ photon decay pattern of the epithermal neutron activated In foil at the 32nd min following the activation (see concerning inserted diagrams of Fig. 4e). The coherency vanished by the 107th min and had zero values as shown in the embedded diagram of Fig. 4f. 4. Conclusion According to the performed experiments the following conclusions can be drawn: 1. The duration of the imaging cycle of the ENR pictures applying real time reading can significantly be shortened by the outlined method 2. The existence of condition (1) makes the image postprocessing and archiving, using task oriented image handling computer program possible. 3. By extending this method and making use of the beam filters and foils, the radiography mapping of neutron beams, including even resonance neutrons, is postulated to be possible. Acknowledgements The presented work has been performed within the framework of the IAEA Project, No. 302-F1- HUN The authors are grateful to Károly

7 M. Balasko& et al./nuclear Instruments and Methods in Physics Research A 424 (1999) Szász, a programmer for IMAN, for the valuable contribution done while applying the β-version of Iman. References [1] H.V. Watts, Report ARF Armour Research Foundation, Chicago, IL, USA, 1962, p [2] H.L. Atkins, The 24th National Convention of the Society for Nondestructive Testing, Philadelphia, Pennsylvania, Session on Neutron Radiography published by the Society for Nondestructive Testing, Evanston, Illinois, [3] H.D. Kosanke, Trans. Am. Nucl. Soc. 14 (2) (1971) 533. [4] L. Moss, M. Beers, J.P. Barton et al., in: J.P. Barton (Ed.), Proc. 2nd World Conf. on Neutron Radiography, Paris, Reidel, Dordrecht, 1986, p [5] M. Balaskó, E. Sva b, Nucl. Instr. and Meth. A 377 (1996) 140.