·Home ·Table of Contents ·Methods and Instrumentation

System of digital radiography for non-destructive testing in the radiation energy range 1-20 MeV Chakhlov V. L., Moskalyov Ju. A., Temnik A. C. Contact

The problem of development of high-efficient systems of digital radiography (SDR) for X-radiation energy range 1-20 MeV is primarily conditioned by the problems of radiation detecting and converting it into a visible image for the subsequent digital processing. The principal difficulty here is comparatively high penetration rate of X-radiation in the energy range of 1-20 MeV. On the one hand it makes possible to test objects of large thickness and on the other hand the radiation converter should be relatively thin in order to perform efficient detection.

Most of SDR employ the classic layout shown in Fig. 1. It consists of the luminescent screen, turning mirror, lens, CCD-camera, controller, and computer with software for image processing.

Fig 1:Scheme of the SDR

Since 1-20 MeV X-radiation has high penetrability and produces rather weak radiation contrast, the development of SDR with high contrast sensitivity presents a relatively complicated scientific and technical problem. For successful solution of this problem all the elements of SDR should provide maximum detection efficiency, radiation and light image contrast, and the minimum level of scattered radiation.

Usually the SDR is considered to consist of two principal units. The first is the X-ray image detector and the second is an image processing unit. The former is located in the irradiated region and includes a scintillation screen (converter), mirror, lens, shielded CCD-camera, and controller. The latter is located in a radiation-safe room and includes a digital interface and a computer with software for image processing.

Screens for converting X-ray images

Currently three types of screens are used for conversion of high-energy radiation to visible image. These are scintillation screens, based on single crystal CsI× Tl, polycrystalline luminescent screens on various phosphor compositions, and fiber screens made of luminescent glass doped with Tb.

However, according to [1] fiber screens have no advantages over other types of screens for energies above 1 MeV. Besides, high-energy X-rays cause radiation damage to the fiber scintillation material.

To determine the conversion efficiency of screens having various compositions the dependence of the fraction of X-ray quantum energy F deposited in a luminescent screen versus incident quantum energy E was calculated according to the technique described in a Ref. [2]. All calculations were made for the phosphor screen thickness of 500 m m and CsI× Tl single crystal screen thickness of 5 mm. The results are given in the Table 1.

Screen material	F (E)´104
	5 MeV	10 MeV	20 MeV
ZnCdS×Ag	0.32	0.26	0.20
BaSO4×Eu	0.37	0.30	0.22
CsI×Tl	0.50	0.40	0.30
Gd2O2S×Tb	0.80	0.65	0.50
PbWO4	1.10	0.90	0.70
CsI×Tl (single crystal)	46.00	38.00	29.00
Table 1:

From the table it follows that compositions which are most efficient in terms of deposited energy are CsI× Tl, Gd₂O₂S× Tb, and PbWO₄, with the energy deposited in PbWO₄ being twice as large as that in CsI× Tl for the quantum energy range of 5-20 MeV. However, taking into account that the phosphor PbWO₄ is still under development and has relatively low luminescence yield, the screens which are in most common use are those based on CsI× Tl and Gd₂O₂S× Tb. It should be noted that radiation conversion efficiency of a single crystal screen CsI× Tl is larger than that of phosphor screens by a factor of tens even if substrates made of such heavy metals as W and Ta are used. So maximum values of contrast sensitivity and signal/noise ratio can be currently achieved in SDRs with single crystal screens. The advantages of phosphor screens with heavy metal substrates manifest themselves when testing large objects because of the problems with the production of single crystal CsI× Tl screens with diameter larger than 200 mm. In this work the SDR with 5-mm-thick single crystal CsI× Tl screens with diameter of 200 mm and polycrystalline CsI× Tl screens with dimensions of 200´ 400 mm² and a 1-mm-thick Pb substrate has been developed.

X-ray sensitive unit and digital processing of images

To obtain the maximum image contrast and a low background rate of scattered X-rays a radiation sensitive unit with the following design features has been developed. The mirror consisted of a plastic substrate on which a layer of Al was sputtered. The objective `Avenir' 75/1.3 was employed. The CCD-camera was shielded with a 20-mm-thick Pb layer. The CCD-camera controller was mounted inside of an X-ray sensitive unit. The CCD-camera array (sensor) consists of one section serving both for accumulation of a photo- generated charge and transmission of a charge profile. It included a register performing line-by-line reading of the profile and the output device. The photo-sensitive array of the sensor consists of 1160 lines. Each of them includes 1040 photo-sensitive cells (pixels) with dimensions of 16´ 16 m m. The sensor is installed inside of a gas-filled metal housing of the DIP type with a Peltier micro-cooler that enables one to maintain the temperature at the level of -35⁰C. This allows one to decrease intrinsic sensor noise and to increase the accumulation time up to 15 min. This feature is very important when thick specimens are exposed and the dose rate on the surface of a scintillating screen is low. The controller sets the accumulation mode for the low-contrast images, maintains the preset temperature of the sensor, reads the accumulated charge, and transmits the obtained signal to the adapter via a 20-m-length cable. The adapter consequently receives the signals from every sensitive cell, converts them into a 12-bit code and provides its input into the computer via a parallel port using the EPP protocol. The control for the scanning process, as well as correction and digital image storage is performed using the Diada 2.0. software.

This software was developed using Microsoft Visual C++ 6.0 and Microsoft Foundational Classes Library. It requires Windows 95 or Windows NT. The following options are available in Diada 2.0.
Acquisition of images from CCD-camera SILAR (12 bits) or TWAIN-scanner (8 bits).
Open image files in standard formats: binary arrays, Windows bitmaps, Flexible image transport system (FITS), GIF, PCX, TIFF (8, 12 and 16 bits).
Display of images.
There are 4 different visualization types:

• Scan - show an image as color picture;
• Surface - presents an image as 3D surface plot;
• Tomogram - show 3 orthogonal slice planes for the image sequence;
• Insight - presents the image sequence as semitransparent body.
To enhance the contrast of an image one can use the lookup table operations and 9 color palettes.
To change the image visualization one can operate about 100 parameters.

Taking measurements of values in points, along lines or on polygonal regions. For polygons Diada calculates statistical estimations: minimum, maximum and mean. For lines it calculates only mean and standard deviations.

Enhancing and manipulating of images. Diada realizes the standard image processing operations: scaling, rotations, filtering, lookup-table transformation, correlation and autocorrelation, binary arithmetical and logical image operations. There are also some special operations for radiographic images: correction of CCD-errors, line equalization, delete background. All operations are related with the selected rectangular region on the images only.

Documentation of the results. One can use the usual Windows Clipboard functionality to transfer images and scan-lines to a word processor for writing reports. Use also the Save as command to store images in standard formats: BIN, BMP, FITS, GIF, PCX, SET, TIFF. Use Print command to print out the image. Use drawing tools to create overlays (marks like rectangles, ellipses, arrows, texts, pictures and so on) on the images. The internal image format of Diada 2.0 is 4096 gray-levels (12 bits).

Results

The experiments in which the 50-mm-thick steel specimen X-rayed with 10 MeV betatron radiation have shown that the usage of the SDR with a 5-mm-thick single crystal CsI× Tl screen allows the detection of a 0.2-mm-deep groove of a groove standard #2 (according to Russia State Standard GOST-7512-82). Grooves with a deep of 0.4 mm and 1.0 mm can be detected in steel specimens having thickness of 100 mm and 150 mm, respectively, yielding the sensitivity of the SDR of 0.4 to 0.7% in the steel thickness range of 50 to150 mm.

Using the SDR with the screen with the size of 200´ 400 mm consisting of the 0.4-mm-thick polycrystalline CsI× Tl on 1-mm-thick Pb substrate yields a contrast sensitivity of 1.0 % for 100-mm-thick steel specimens, i. e. 1-mm-deep grooves and wires of 1 mm in diameter could be detected in the standards.

The reported results on the contrast sensitivity of the SDR are preliminary and additional experiments showed that the sensitivity, resolution, and brightness of the screen can be improved by using new phosphors, increasing the phosphor layer density and applying heterogeneous amplifying screens made of heavy metals.

References

1. Cliff Bucno and Marion D. Barker, "High resolution digital radiography and three-dimensional computed tomography", SPJE, No 2009-21, pp. 1-12, 1997.
2. Ju.A. Moskalyov, A.Ja. Gurevich, and D.I. Svirjakin, "Energetic sensibility of luminescent transducers of X-ray radiation", Defectoscopy, No 9, pp. 45-49, 1988.

© AIPnD , created by NDT.net

|Home|    |Top|