Why Replace Film with Digital Radiography
The digital revolution that occurred in both amateur and professional photography was the direct result of increased recognition that—for most applications—digital works better. Digital radiography (DR) experienced similar rapid growth in clinical use, and has spilled over into the NDT arena. While DR technology has been in use for some time, recognition of its merits had been somewhat lacking among the NDT film radiography community. There are several types of DR but the two most commonly used in NDT--and the focus of this article--are Computed Radiography (CR), and Digital Detector Array Radiography (DDA).
Computed Radiography can be used with X-ray systems formerly used with film, but is much faster than film imaging. Computed Radiography (figure 1) uses a special phosphor plate or imaging plate (IP) to store the radiation signal not fully attenuated by the test object. The amount of radiation reaching the phosphor plate during exposure translates to more or less signal being stored in the phosphor. After exposure, the IP is scanned with special laser optics to release the hidden or ‘latent’ image in the phosphor. From there, photo-multiplication, signal amplification, and digitization bring the image into a pixel format – at a preselected resolution. As with any raw digital image it must be enhanced with further processing.
The IP is erased at the end of the scanning process and is ready for a new exposure. When used properly imaging plates can acquire thousands of exposures that, of course, leads to substantial savings in film and processing costs.
Digital Detector Arrays
DDA (digital detector array) images are captured with a solid state detector—or panel--that in as little as a few seconds converts and relays the information to the image processor and workstation for display. The signal conversion process that occurs in the multi-layered panel is shown in Figure 2.
Portions of the applied x-ray signal completely penetrate the part and interact with a phosphor or scintillator just inside the DDA input window (1). The phosphor/scintillator layer converts the x-ray signal to visible light. The efficiency of this first conversion directly affects image quality, and different types of phosphors or scintillators are used in the variety of DDA’s currently available. For example, use of a thicker, high quality scintillator of CsI (Cesium Iodide) would absorb more of the X-ray signal, resulting in SNR (signal to noise ratio) improvement.
The phosphor/scintillator layer is in contact with an amorphous silicon thin-film transistor array containing individual detector locations called pixels (2). High resolution detectors may have millions of pixels, depending on pixel dimension and array (active area) size. Each pixel translates the amount of light received from the phosphor/scintillator into an amount of electronic charge in the amorphous silicon. That charge is then rapidly removed from the transistor array in channels, synchronized, amplified, and digitized within the DDA electronics (3), and finally routed to the system image processor for eventual display at the workstation (4).
Most often, multiple images or “frames” are captured during the DDA imaging process. These frames are combined to average out noise and clean up an image. The final raw digital image must be digitally enhanced, typically with a digital filter, or by re-scaling, to improve feature or indication detection, characterization, and proper disposition.
One area of concern with DDA’s is ‘bad pixels’. These are pixels in the DDA that do not respond properly to the input signal, or respond independently of signal. All DDA’s have bad pixels, the extent of which will establish the grade (and eventual cost) of the DDA. Bad pixels that do not have enough neighbor pixels to be corrected are ‘cluster kernel pixels (CKP’s), that should remain in the production image and be treated as image artifacts. When inspecting high performance hardware, with tight material acceptance standards, the frequency and location of CKP’s, and the potential for masking a defect must be considered. There are several specific procedures and performance evaluations that are used to quantify and qualify various DDA metrics. These metrics include measurement of the detectors basic spatial resolution, contrast sensitivity, and retention of radiation signal, among others. Indications of performance degradation, or variation, necessitate the need for investigation and a determination for any inspection impact. Repeatability and reliability in a system must be demonstrated through periodic performance evaluations.
The dynamic range of an imaging system is its ability to detect an extensive thickness or density change in the test specimen, within a single view or exposure. Film radiographs are 8 bit images (8 bit = 28 = 256) that contain about 250 different shades of usable radiographic density (1.50 to 4.00 densities). This is a very limited dynamic range compared to most DR systems that, even on the low end, can produce 12 bit (4096 shades of gray) images, and on the high end up to 16 bit (65,000+) images. The high dynamic range of DR imaging systems, however, means greater sensitivity to scattered radiation, too much of which results in ‘noise’ in the image. There is noise in any imaging system, which must be limited or controlled to maintain the very low contrast sensitivity DR systems are quite capable of.
In radiographic inspection, contrast sensitivity is important, but spatial resolution (definition) is equally significant. Spatial resolution is the ability to see fine detail. This detail in a radiographic image is closely related to geometry of the technique, which includes distances and x-ray tube focus size. It is also highly dependant upon the detector itself. Radiographic film has photosensitive silver crystals or grains about 2-5 microns in size. Commonly used DR systems pixels are usually around 25 to 200 microns, or about 5 to 40 times more than film crystal size range. Therefore, film has inherently higher definition capability than any DR system. There are certain specialized types of DR detectors with limited active area size that feature pixel sizes down to 10 microns, but those are often limited to specific applications or radiation energies. CR systems currently are available with 25 micron scanning resolution, increasing resolution to nearly that of film, but often with a dramatic increase in image file size. DDA image definition is often improved with geometric magnification. Geometric magnification in a DDA technique always requires an x-ray tube with a reduced focal spot. A reduction in field-of-view is also a consideration.
Replacing Film with Digital
The incentives to switch from film to digital are to eliminate the purchase of commodity-based film, to forget the film processor and chemistry headaches, to speed up and improve the ‘drum-beat’ of the radiographic inspection process, and to gain more convenient image archives. Of course, the common concern of any radiography professional is, or should be: ‘will we find the same size or type of defect with DR that we would with film?’ There is no fixed answer to this question, as DR systems relate to film based imaging in only one manner – they both convert the primary signal, ionizing radiation, into readable information. That is where all similarity ends. So the question that must be asked is ‘Will this particular DR system meet the radiographic inspection requirement?’
There are DR systems that easily surpass the image quality of film. These systems are quite often custom built, using automation, geometric enlargement, and high resolution DDA’s within the system. Other systems are in use that do not have near film resolution, but fit the application handily. Any DR imaging system must be evaluated prior to implementation, with system resolution, material form, throughput, ease of operator interface, and budget all being considered in the final choice – the benchmark being the detection and accurate portrayal of any substandard discontinuities for a given radiographic inspection application.