DoaWise

2026-07-09 · EN

Digital radiography (DDA) for fast field inspection

In digital radiography, "fast field inspection" now comes down to one piece of hardware: the flat panel detector. The digital detector array (DDA) turned radiography from a process paced by film chemistry into something close to real-time imaging. The beam hits the panel, an image appears on screen within seconds, and the inspector can change a parameter and see the effect on the next shot almost immediately. That is a step change in field time, required exposure and archiving load. But that speed demands discipline: in DDA work, image quality is no longer governed by developing time — it is governed by signal-to-noise ratio (SNR), basic spatial resolution and correct exposure. And "fast radiography in the field" is never purely an imaging question; radiation safety and the licensing framework are inseparable from the job. This article looks at how a DDA works, what its quality metrics are, how it differs from film and CR, what the regulatory duties are, and which standard clauses apply — all from a field perspective.

1. What a DDA is and how it works

A digital flat panel detector is a detector array that converts incident X-ray or gamma radiation into an electronic signal. Two basic architectures exist. Indirect panels first convert radiation into visible light with a scintillator layer — usually GOS (Gd₂O₂S, gadolinium oxysulfide) or CsI (cesium iodide) — and an amorphous silicon (a-Si) photodiode matrix underneath turns that light into electrical charge. Direct panels use a photoconductive layer such as amorphous selenium (a-Se) to generate charge from the radiation without an intermediate step. In both cases the panel is a matrix divided into millions of pixels; each pixel reads the charge it has collected, and the result becomes a digital image built from gray values.

One distinction matters in the field: in high-energy industrial radiography — thick-section welds and castings shot with Ir-192 gamma sources or high-kV X-rays — indirect scintillator panels (CsI/GOS + a-Si) dominate in practice. Direct a-Se panels are mostly used at lower energies, especially in medical imaging. The two architectures are not interchangeable for every job.

CsI-based columnar scintillators channel light through needle-like structures instead of spreading it sideways, which gives both high efficiency and good sharpness; that makes them the preferred choice for heavy-section castings and welds. Pixel pitch (typically 100–200 µm) sets the physical ceiling on the resolution a panel can reach.

2. Real-time imaging and field workflow

The most visible advantage of a DDA is that the image forms on screen immediately after the shot. In film radiography the expose–develop–dry–evaluate cycle usually runs in tens of minutes; with a DDA that cycle drops to a few seconds per shot. This has three consequences in field practice:

  • Setup verification is fast. Beam angle, source-to-detector alignment and IQI position are visible on the first shot; when the geometry is wrong, a repeat shot costs seconds instead of a lost cycle.
  • Exposure optimization is iterative. The inspector can change kV, mA and the number of integrated frames and watch how SNR responds.
  • Decisions come faster. On a line with many welds, the inspector makes the image quality call on site, and the report takes shape before anyone gets back to the office.

In real-time work the image is built by summing multiple frames (frame averaging or integration). As the frame count rises, random noise falls and SNR improves — which is why "real time" usually means "an integrated exposure of a few seconds".

This live, iterative workflow has one critical precondition: it essentially requires an X-ray tube. An X-ray source can be switched on and off, so the beam can be cut and a parameter changed while personnel stand at a safe distance. Gamma sources (Ir-192, Se-75, Co-60) cannot be switched off; live, repeated exposure with personnel nearby would mean continuous irradiation, which no procedure allows. With gamma sources, trigger the exposure remotely and safely, and never expose before the controlled area is cleared. In other words, the appeal of "live optimization" is bounded by the source type and by radiation safety rules.

3. SNR — the heart of digital radiography

In film, image quality is broadly set by film density and contrast. In DDA work, signal-to-noise ratio (SNR) takes the front seat. SNR is the ratio of the mean gray value in a region to the standard deviation of the noise in that same region. High SNR means a discontinuity can separate itself from the noise floor; low SNR can hide a fine crack or a low-contrast slag inclusion entirely.

The key point is this: in a DDA, you can raise SNR by increasing exposure time — that is, the frame count. In film, improvement stops beyond a certain density; in digital imaging, noise keeps falling statistically as more frames are collected. Measured SNR depends on the detector's actual spatial resolution (its unsharpness), so standards work with normalized SNR (SNR_N) to compare systems of different resolution fairly. EN ISO 17636-2 normalizes measured SNR to a reference basic spatial resolution of 88.6 µm: SNR_N = SNR × 88.6 / SRb. Acceptance thresholds are tabulated directly as SNR_N against material thickness. Without that normalization, comparing a sharp but noisy system with a soft but clean one would be misleading.

4. Basic spatial resolution

The second governing parameter in DDA work is basic spatial resolution (SRb). It represents the smallest detail the system can resolve and is directly related to pixel size in practice; it is normally measured with a duplex wire image quality indicator (duplex wire IQI, EN ISO 19232-5). Read the image profile across the wire pairs of the duplex indicator: the first wire pair whose profile dip falls below 20% counts as unresolved; take the preceding pair, which is still resolved as two separate wires. SRb is half the total unsharpness (uT) of that pair: SRb = uT / 2 (in µm).

Why does this matter? Film radiography offers very high resolution, since the film grain carries sub-micron detail; in a DDA the pixel pitch imposes a physical limit. Catching sharp-edged discontinuities such as fine cracks or narrow lack of penetration requires a small enough SRb. That is why standards specify the largest permitted SRb by material thickness and class. In practice you improve SRb by selecting a panel with a smaller pixel pitch, using geometric magnification, reducing focal spot size, and setting source-to-object and object-to-detector distances correctly to limit geometric unsharpness (Ug).

SNR and SRb pull against each other constantly: magnify to gain resolution and you shrink the covered area; integrate longer to raise SNR and you extend the shot. A good DDA procedure finds the window that holds both parameters at the thresholds the standard demands, at the same time.

5. Comparison with film and CR: speed, exposure, archiving

Compare the three technologies from a field perspective.

Speed. Film is the slowest because of chemical processing; every shot is developed separately. Computed radiography (CR) reads a phosphor plate in a scanner — no chemistry, but plate handling, scanning and erasing still take minutes per shot. A DDA brings the shot-to-image cycle down to seconds and allows real-time work. On jobs with high weld counts, the throughput difference is dramatic.

Exposure and dose. A DDA has wide dynamic range; because it tolerates very different thicknesses in a single shot, it reduces the repeat shots that film and CR would need. The high quantum efficiency of CsI panels often makes it possible to reach adequate SNR with less exposure. Note the distinction here: panel efficiency lowers the required exposure — the radiation reaching the panel per shot — but it does not directly lower personnel effective dose. Personnel dose is governed by distance, time and shielding, by controlled area management, and by the ALARA principle. The contribution of higher efficiency is indirect: less exposure means less scattered radiation and fewer repeat shots, which helps limit total irradiation — but protection measures are what actually control dose. And "less exposure" is not always right: the correct exposure is whatever is needed to hit the target SNR.

Archiving. With film, the archive is a physical burden and a degradation risk; humidity, light and density loss over time are real problems. Both CR and DDA are digital from the start: images are stored in DICONDE format (the industrial NDT version of DICOM) together with exposure parameters and traceability records. A digital archive makes search, re-evaluation, remote second opinion and audit trail far easier. In return, it demands its own discipline for data integrity, backup and long-term readability.

In short: film can still deliver the highest spatial resolution and needs no digital infrastructure; CR is the flexible bridge from film to digital; DDA stands out where speed, repeatability and field throughput matter.

6. Field throughput: where a DDA makes the difference

The real return on a DDA shows up in volume, not in a single shot. Where hundreds of welds must be tested — pipe spool fabrication shops, series weld seams, boiler panel walls, erection sites — the minutes saved per shot turn into days. With motorized manipulators positioning the panel and the weld automatically, one operator can work a continuous flow. Real-time imaging also creates a feedback loop with the welding side: when a repeating discontinuity pattern appears, the problem is solved in production by correcting the welding parameter — testing shifts from sorting out bad parts afterwards to improving the process.

The field costs of a DDA are equally real. The panel is an expensive instrument and sensitive to impact; it is also vulnerable to cumulative-dose radiation damage — there is a limit to the total dose a panel can take over its life. Dead pixel maps and calibration (offset/gain correction) require regular maintenance. At very high energies and in tight access, film can still have the advantage. Choose the method according to the geometry and the volume of the job.

7. Quality assurance: calibration and verification

A DDA image is not used raw. Because every pixel of the panel behaves slightly differently, run offset (dark current) and gain (brightness response) calibration first; then correct known faulty pixels with bad pixel correction, interpolating from neighboring values. These corrections drift over time. Standards require periodic recalibration and regular verification of system performance through duplex wire (SRb) and SNR measurement. On an uncalibrated panel, fixed pattern noise can read as a genuine indication — or, just as easily, mask one.

8. Image processing: helpful, but limited

One attraction of going digital is post-processing: window/level adjustment, contrast stretching, edge sharpening and noise reduction. These make evaluation easier — but the golden rule does not move: image processing cannot rescue an inadequate exposure. If SNR and SRb are not sufficient at the source, no filter will bring back detail that was never captured, and over-sharpening can create artificial edges and lead to misinterpretation. Keep every processing step traceable and reproducible so the examination stays auditable, and always retain the raw data.

9. Radiation safety and regulation

Radiography in the field is a matter of authorization and protection before it is a matter of equipment. Industrial radiography operates under a national radiation regulator's license in essentially every jurisdiction, and no "fast inspection" is legitimate outside that framework. The minimum duties include:

  • License: the organization performing the work must hold a valid license from the national nuclear/radiation regulator.
  • Radiation Protection Officer (RPO): a competent RPO, qualified as the regulations require, must be assigned to the activity.
  • Controlled and supervised areas: the exposure area is delineated, marked and cleared; access is controlled and dose rate is measured.
  • Dose limits and ALARA: statutory dose limits for workers and the public apply, and exposure is kept as low as reasonably achievable. Dosimetry monitoring is mandatory.

The speed of a DDA shortens none of these steps. If anything, the appeal of a real-time workflow — live exposure optimization — carries a risk of increasing irradiation time when it is applied badly. That makes safe triggering, remote control and area management even more important.

Related standards

  • EN ISO 17636-2 — Radiographic testing of welds with digital detectors (CR and DDA); SNR_N and basic spatial resolution requirements, class A/B techniques. Normalizes SNR to a reference SRb of 88.6 µm.
  • EN ISO 19232-5 — Image quality indicators: determination of basic spatial resolution with duplex wire IQIs.
  • EN ISO 19232-1 — Determination of image quality value with wire-type IQIs; EN ISO 19232-2 — determination of image quality value with step/hole-type IQIs.
  • ASME BPVC Section V, Article 2 — Radiographic examination; mandatory appendices for digital detector arrays (DDA) and computed radiography (CR), including SNR, SRb and IQI requirements. Section V gives the method only; it does not give acceptance criteria.
  • ASTM E2597 — Manufacturing characterization of DDAs (manufacturer-side characterization, compared through basic technical measurements).
  • ASTM E2737 — Evaluation of DDA system performance and monitoring of long-term stability (baseline plus periodic tests); applied at system acceptance and at regular intervals.
  • ASTM E2698 — Standard practice for radiographic examination using DDAs; in practice applied together with E2737, which supplies the baseline evaluation and stability procedure.
  • ISO 9712 — Certification and level grading of NDT personnel, including RT. Note: this certificate demonstrates NDT competence; the authorization to work with a radiation source in the field comes separately, from the radiation protection training and certification required by the national nuclear regulator. Do not confuse the two.

One separation has to be crystal clear: SNR_N and SRb are image quality thresholds (technical adequacy) — they decide whether the radiograph meets the standard. They are not acceptance/rejection criteria for the discontinuities in the weld (porosity, slag, cracks, lack of penetration). Acceptance or rejection of the weld is judged against a separate document, the acceptance criteria of the construction code — for example ASME Section VIII Div. 1 UW-51 (full RT) / UW-52 (spot RT) under ASME, or ISO 5817 quality levels under EN/ISO. Permission to use a DDA on a given job, and the applicable thresholds, likewise come from the referencing construction code (for example ASME Section VIII, B31.3); Section V Article 2 alone is not sufficient, and the procedure must also be qualified. Standard selection follows the client specification and the product code: ASME-scope work uses the Section V Article 2 method, while European/ISO-scope weld work is based on EN ISO 17636-2.

From the field

The most insidious trap with a DDA is mistaking speed for quality. An image that appears on screen within seconds is not, for that reason, an image that meets the standard. The mistake I see most often in the field is cutting integration time short, leaving SNR below target, and then accepting the shot because it "looks clean" — low SNR will bury a fine discontinuity in noise without any trouble at all. The second is working without ever measuring SRb with a duplex wire indicator; if the pixel pitch is too coarse, or the geometric magnification is set up wrong, the system simply will not see a sharp-edged discontinuity. The third is skipping panel calibration (offset/gain, dead pixels), which leaves you reading fixed pattern noise as a genuine indication. And one more sits above all of these: confusing image quality thresholds (SNR_N, SRb) with the acceptance criteria for the weld. They belong to two different worlds — one tells you whether the radiograph is adequate, the other tells you what discontinuity the construction code will tolerate. The equipment is fast; the correctness of the examination still rests on SNR, SRb, the right acceptance code and a disciplined procedure. Work every shot knowing the two numbers the standard asks for — normalized SNR and basic spatial resolution. Without them you have an image, not an examination.

Türkiye-specific note

In Türkiye (Turkey), industrial radiography is licensed by the Nuclear Regulatory Authority (NDK), Türkiye's nuclear regulator. The legal basis is Nuclear Regulation Law No. 7381 (Türkiye, 2022; Official Gazette 8 March 2022, No. 31772) together with the national regulation on radiation protection and licensing in industrial radiography. Under that framework the licensee must hold a valid NDK license, assign a Radiation Protection Officer, establish controlled and supervised areas, and comply with statutory dose limits and dosimetry monitoring. NDT certification under ISO 9712 does not replace this authorization.

Disclaimer

This article is for educational purposes and does not replace the official standard. Always work from the current edition of the applicable code or standard, and from the radiation safety legislation in force in your country.

Take it to the field: To keep the NDT methods, standard references and field steps in this article in your pocket — completely offline and free — take a look at the Doawise NDT Guide app.


At Doawise we carry out non-destructive testing services, digital radiography included, to international standards and to the radiation safety regulations in force. From weld and casting examination to fast field radiography with flat panel detectors, we produce results that are stamped with their SNR and basic spatial resolution thresholds, tied to the correct acceptance code, and fully recordable and auditable.