APPLICATION IN NDT OF COMPUTER DIGITAL X-RAY REAL TIMEIMAGE SYSTEM
Abstract Nondestructive testing (NDT) is one of the most commonly used scientific means to detect the internal defects
without damaging the tested objects, therefore it is found widely applications in the industries of mechanical, petrochemical, aerospace etc. Radiographic Testing (RT) by X-ray is commonly used NDT method at present, its disadvantages are unavoidable for its long test period, high cost and environment pollution. With the development of computer technology, an emerging NDT method with digital X-ray real time image (DXRTI) is found increased application in industry. This paper presents the fundamentals and applications of computer numerical image technique (CNIT) in NDT with DXRTI technique, and introduced the new developed DXRTI software ‘XCX2000’ and its applications and economic benefit in weld defects detecting for LPG Cylinders.
computer, software, Real Time Image, digital images processing, X-ray,
1. Introduction The principle of NDT with DXRTI may be briefly expressed by the two conversions: After penetration of metal, X-ray signals was captured and converted into the visualized images by the Image Intensifier, which is called photoelectric conversion. The visualized images was shot by very fine resolution video camera and then was converted into digital image by Image Acquisition Card , which is called A/D conversion. The digital image was processed with computer and display the properties, dimension and location of the inside defects on screen. Finally, the defect grade was assessed according to the relevant codes and standards. The quality and effect of DXRTI is acceptably good compared with radiographic films, so that the later can be replaced by the former method. Moreover, with its much higher efficiency, lower cost, better inspection effectiveness, remoter transmission and much more convenience, DXRTI has prosperous future in the field of NDT. 2. System requirements DXRTI NDT system is composed by X-ray source, image intensifier, optical channel, video camera, computer, Image Acquisition Card, image display and storage equipment(see Fig.1). The X-ray source is different from the conventional one in its constant potential, small focus and forced convection cooling system. 2.1 The Computer’s Requirements Intel MMX166 CPU or faster, 32MB RAM, 2.0G Hard disk free space,2D Graphic Accelerate Card(4M video memory ), 17” SVGA Monitor, CD-R Writer. Operating System: Microsoft Windows 95 /98.
Fig. 1 Block Diagram for NDT with DXRTI
2.2 System Management For convenient operations, two computers can be connected to the system to form a peer to peer network, one for image acquisition and the other for image assessment, and each computer works with identical purpose. The result images are stored in the hard disk for assessment and transmitted for very long distance via internet. 3. Digital image processing by computer The processing of digital images by computer includes three parts: image acquisition, images treatment and image display。 3.1 Image Acquisition X-ray from radiographic machine penetrates the test object and converted into the visible image on the screen by the image intensifier. Then the images are changed into video signals by the video camera and sent to the Image Acquisition Card for digital conversion. Finally the digital images are to be processed by the computer. There are two important parameters in the image acquisition: gray scale and
resolution of the images. Commonly, the A/D converter of the image acquisition is 8 bits with colors of 256 gray scales. Higher resolution of the image acquisition will effectively improve the defects resolving power. The resolution of the image acquisition should be no less than 768X576 dot lines. 3.2 Image Processing The main purpose of image processing is to extract specific information from the images for comprehensive computer analysis and the identification of different images. 3.2.1 Image Superposition for noise reduction It is noticed that random noise information is unavoidable in image acquisition, which is one of the important factors affecting the quality of the images. For the case of stable image acquisition, one of the effective and practical ways to eliminate the random noise is to superpose the input images continuously. Theoretically, the time dependent noise can be removed or filtered completely if sufficient frames of the images superposed. Therefore, frame superposition is commonly used in the processing of images. This method is expressed by the following equation:
g ? x, y ? ?
? fk ?x, y ?
Where, g(x, y) stands for the dot gray scale of the image, M is the total numbers of the superposed image frames, fk(x,y) is the gray scale value for the Kth frame image. The sampling test shows that the results will be satisfactory when M≥16. 3.2.2 Gray Scale Expand
g ? x, y ? ? 255 ? f ?x, y ? ? A1? ? A2 ? A1?
Gray scale conversion technique is commonly used. With this technique, the image quality may be improved by adjusting the gray scale of the image following some
specified rules. For most of the cases, the image gray scales vary within very narrow regimes, resulting into the image with very poor contrast and vague margins. By linear conversion, the gray scales of the image can be expanded to 256 (0～255), so that the contrast and the resolving power of the image will be improved. Suppose the original gray scale regimes of a image is A1～A2, the relevant conversion equation is
? ? ? ? ?b?x ? ? a?x ??? ? g ?x, y ? ? 127?1 ? sin ? f ?x, y ? ? ? 2?b?x ? ? a?x ?? ? ? b? x ? ? a ? x ? ?? ?
Moreover, with proper nonlinear conversion of the image gray scales, it is possible to highlight special portion of the image with certain gray scales. For example, the S-T conversion may effectively improve the contrast of the welding defects and therefore the defect margin become clear. Its conversion equation is written as per following: Where, b(x), a(x) stands for the maximum and minimum gray scale value of the relevant dot column respectively. 3.2.3 Brims Sharpening To sharpen an image is to highlight its brims with better contrast for easier identification of the internal defects of the welds or objects. Image sharpening is substantially a technique of high pass filtration realized by a special model operator. This operation has obvious characteristics of simple and efficient and therefore is applicable for brim sharpening, boundary detecting, characterizing and smooth filtering for the defects. 3.2.4 Image Reversion
g ?x, y ? ? 255? f ?x, y ?
For the convenience of the technicians, it is necessary to reverse the resultant image and display it in the similar way as the negative image of X-ray by the equation below:
Where, f(x,y),g(x,y) stands for the original and the reversed dot gray scale values respectively. 3.2.5 Pseudo-color Treatment For the numbers of the colors is larger than the image gray scales, the gray scales image can be mirrored to the relevant colors which improves the identification or the sensitivity of human eyes. Such processing is called pseudo-color treatment. 3.2.6 Miscellaneous The other image processing approaches include histogram change, digital filter etc. 3.3 Image Display and Recording The processed images are displayed on the screen of the monitor for further evaluation and assessment. 3.3.1 Weld Defects Identification by Computer The weld defect characteristics such as boundary length, longitudinal diameter, transversal diameter, and areas are calculated by tracing the defect boundary with computer. Furthermore, based on the above parameters, the name of the weld defect will be identified and the dimension calculated by the computer automatically as per the specified empirical formula. For the computer identification model of the defects is very complicated, the defect assessment is still processed manually at present, but the defect quantitative analysis and classification work can be carried out with computer. 3.3.2 Image Compression coding Because the digital images take huge amount of space, it’s necessary to have the image data compressed for efficient transmission and storage. The DXRTI system is coded with non-anamorphic means and applicable for the normal compression coding such as RLE8 and LZW.
3.3.3 Image Storage and File Format For the future checkout and search, the image data of the DXRTI system should be stored in the format of image file with large capacity storage media such as CD-R and kept in the place free of magnetic field, humidity, dust and mechanical press. TIFF is more suitable for the image files in DXRTI system. 3.3.4 Image Storage Device Weld detecting images normally stored by the hard disk of the computer at first, and then are transmitted and written onto CD-R disc by a CD-Writer. One may store 4200 frames of images with its capacity of 650MB and a shelf life of more than 30 years. Therefore, it’s the predominant storage media for the detected image information. 4. Programming for DXRTI detecting system Between the end of 1998 and the beginning of 1999, the technicians in the plant developed a new package of the DXRTI software named ‘XCX2000’ based on the principle of computer numerical imaging and the previous imaging technique. The software composed of two parts: the image acquisition program and image assessment program shown by the block diagrams below. 4.1 Image Acquisition Program (see Fig. 2) 4.2 Image Assessment Program (see Fig. 3)
Fig. 2 Block Diagram for Image Acquisition Program
Fig. 3 Block Diagram for Image Assessment Program
(a) Original image
(b) Image after Brims Sharpen
(c) Image after linear Gray Scale Expand
(d) Image after S-T conversion
(e) Image after histogram change
Fig. 4 Sample of image processing effect
1) Lin Piyuan, Pu Hepin: Computer images processing technology. Publishing House of Electronical Technology University, 1998. 2)Zhang Hua, Some remarkable problems in popularizing ISO9000 quality systems for software engineering. Symposiums of 5th Chinese computer application congress. Publishing House of Electronics Industry, 1999. 3)inshaw D.A., Dobbins III J.T.. Recent progress in noise reduction and scatter correction in dual-energy imaging[J].1995，SPIE-2432：134～142. 4)D. A. Stubbs,L. P. Zawada. Detection of porosity in glass ceramic matrix composites using anult rasonic multiple-gate c-scan technique,Materials