For decades in the Nondestructive Testing (NDT) industry, film has been the Radiographic imaging method of choice. Its evolution and industry competition has driven film into a high quality-level with many options that cover virtually anything requiring radiography. Slow film is used for fine resolution options while faster speeds cover larger or less critical materials. Today, we have many new competitors attempting to dethrone film as the industry king. Many of these options share a single result: digital imaging.
The transition from traditional to digital images has been a gradual one with very small leaps and bounds. We’ve all, likely, been presented with it in one form or another, but recent NDT industry changes have put digital squarely in many of our future business plans. Projections of better quality along with promises of a much higher production are sparking the interest of service providers and fabricators alike. Before we understand the new technology we’re up against, let’s understand what digital imaging really is.
What is Digital Imaging for NDT?
If we were to take a digital technology and strip away the hands-on portion, we are left with Digital Imaging. Computed Radiography (CR), Digital Detector Arrays (DDA), Linear Diode Arrays (LDA), and other “Radiographic Techniques” result in digital imaging. The “Imaging Medium” is the receiver of the radiation dose which gives each modality its own “personality”.
While digital imaging is a constant to these processes, the images produced have their own characteristics based on practical applications. A CR image tends to have a high resolution, though can also suffer from noise. A DDA image, inversely, will tend to have a lower resolution, but should excel at noise reduction. Alternative applications may also have an effect.
Signal, Noise and Radiography
Computed Radiography on a pipe will have a different noise level than CR on a steel casting. This brings me to our first two topics: signal and noise. Signal is the primary radiation beam transmitted to the imaging medium. Noise, however, is a combination of any extra energy reaching that same imaging medium. Light, scattered radiation, electronic noise, and haloing name just a few sources of noise. Any technique will have signal AND noise characteristics based on the properties of that equipment. Ideally, zero noise with 100% signal would be our target but, because noise is driven by signal and vice versa, that target is not only impractical, but impossible. All Radiography has noise regardless of our best intentions. A well-developed process should give an optimum signal to noise ratio, not eliminate it completely.
The Best Resolution for NDT Digital Imaging
Another factor we should understand, at least a little bit: resolution. Consider the television market as our example. Television screens are composed of a grid of tiny pixels making up the picture. Once high-definition TV’s took the market, we started discussions on resolution. 720P, 1080P, and now 4K are all what we would consider a total resolution, or the total of pixels in one line or one panel. In a 1080P television, the vertical lines of the screen are 1080 pixels long. There are then 1920 of those vertical lines, meaning that we have a resolution of 1920 x 1080, or 2.07 million pixels, in the panel.
If we considered a television with the exact same resolution, but a larger physical size, the pixels themselves would have to be larger. It still has 1080 pixels in each line with 1920 of those lines in the screen, but each pixel needs to cover more physical area to make up the larger screen size. This forces the industry to format in ‘pixel pitch’ rather than total pixels.
Pixel pitch describes the measurement from the center of one pixel to the center of the next, eliminating the total pixel count as our metric. That standardizes the resolution without an image size, making a simple resolution comparison of differently sized images. Pitch is measured in ‘microns’ or 1/1000 of a millimeter, and a common pixel pitch would be 50-200 microns. Ideally, you would want as many pixels as reasonably possible in an image. Too small of a pitch, however, will result in more noise, so the smallest isn’t always the best. A strong technician should know how to negotiate an acceptable resolution.
4096 Shades of Grey: Determining Radiographic Bit-Depth
We’ve established that pixels are laid out in a two-dimensional grid but haven’t discussed how to fill them. That’s where bit-depth comes into play. Each tiny pixel in an image has a ‘grey value’ represented by a number. The potential range of that number, known as the ‘dynamic range’, depends on the bit-depth of the image.
One bit is a piece of information containing either a 1 or 0 representing ‘on’ or ‘off’. If we take a second bit and string the two together, we can get the on/off positions, as well as two more in between. The two bits (2 bit-depth) could represent 00 or 11 but could also be 01 or 10. If we added another bit (3 bit-depth), the options could be 000, 001, 010, 100, 011, 110, 101, or 111. The amount of options doubles each time we add one bit.
Now, let us consider these shades of grey from white to black rather than an on/off scenario. We’ll also be discussing this at a much higher bit-depth. A bit-depth of 12 would give us 212 options or 4096 shades of grey while 16 bits would give us 216 or 65,536 shades of grey. A set of all zeros would result in no exposure, or white, and a set of all ones would result in complete saturation, or black. The remaining options in between would be shades of grey.
If we were to radiograph a casting with many thickness changes, 16 bit would give us the range to put them all on one image. A weld with minimal thickness changes would be much better at 12 bit since the high range wouldn’t benefit that application. With a higher bit-depth comes a higher radiation exposure and a chance at more noise, the same as higher resolution. Many applications will see no benefit to the higher range so 12 bit is generally plenty. Radiography requiring great detail will likely need a 14-16 bit setting, regardless of the extra noise. It’s a determination the technician needs to make.
Digital Imaging is the Future of NDT
Clearly, there are many options and features built into digital images and this may be the most many of us will ever read about something as common as a picture. However, the simple truth is: digital imaging is our future and we need to prepare for it.
Film will be obsolete eventually and an ounce of prevention is worth a pound of cure. The earlier the industry gets ahead of the curve, the more time and money our clients will save later. Moreover, training an NDT technician correctly can go a long way to develop a sound program in the long run.
Computed Radiography Webinar
Interested in learning more about the future of NDT and digital imaging? Click the link below to sign up for a webinar with the author, Wesley Soape.