What can you check with a USAF target?
The target gives a controlled test object for comparing optical configurations. Its result describes the system under the recorded conditions, including focus, illumination, sampling, vibration and image processing. It is particularly useful when a team needs a simple pass/fail detail check or a repeatable comparison before and after an adjustment.
- Lens and camera selection: compare resolved detail at the same working distance, aperture and magnification.
- Microscope setup: check the objective, camera adapter and focus as one optical path.
- Machine-vision acceptance: verify that required detail remains distinguishable in the actual inspection setup.
- Maintenance and repeat checks: compare a recorded baseline after replacing a lens, remounting the camera or changing the lighting.
Understand groups, elements and line pairs
Each group contains six elements, numbered 1 to 6. The group number G and element number E together identify the spatial frequency. Moving forward by one element multiplies frequency by 2^(1/6), approximately 1.1225; the same element in the next group doubles the frequency. One dark bar plus one adjacent light space forms one line-pair period.
- Read the group label near the first element, then the element number next to the bar pair. A numeral beside a bar is not automatically an lp/mm value.
- In the classic layout, element 1 of an even-numbered group sits at the lower right, separated from elements 2 to 6 on the left. Follow the labels rather than assuming every group is one vertical column.
- Use lp/mm or cycles/mm for spatial frequency. The ambiguous term lines/mm should be clarified before comparing a supplier table or a test report.
Convert group and element into feature size
The target-plane frequency is f = 2^(G + (E − 1)/6) lp/mm. The line-pair period is 1/f mm, while the width of one bar or one equal space is 500/f µm. The selected examples below are mathematical pattern values, not a statement that every target includes all listed groups or that these values are manufacturing tolerances.
| Group G | Element E | Frequency (lp/mm) | Bar width (µm) |
|---|---|---|---|
| 0 | 1 | 1.00 | 500.00 |
| 2 | 3 | 5.04 | 99.21 |
| 4 | 1 | 16.00 | 31.25 |
| 6 | 1 | 64.00 | 7.81 |
| 7 | 1 | 128.00 | 3.91 |
| 7 | 6 | 228.07 | 2.19 |
| 8 | 1 | 256.00 | 1.95 |
| 9 | 1 | 512.00 | 0.98 |
Select clear bars or opaque bars
Opaque bars on a clear field and clear bars in an opaque field have the same nominal bar geometry but different illuminated areas. Match the layout to the illumination and control exposure and flare when comparing results. State the observed side and film orientation for the mounted target.
A practical resolution-check procedure
Keep the setup and reading rule consistent. A useful report records the highest correctly resolved element in each direction, together with the test conditions, rather than only attaching a sharpened image.
- Mount the target in the intended object plane and check cleanliness, orientation and stability. Define the working distance and aperture.
- Use controlled illumination and exposure. For system comparisons, fix sharpening, denoising, resampling and other image-processing settings.
- Focus on a resolvable element, then progress from coarse to fine. Count three distinct bars with separating spaces in each orientation.
- Record the last resolved group/element for horizontal bars and vertical bars separately. If one conservative system value is required, use the lower limit under the agreed rule.
- Repeat at selected field positions and retain the original image, magnification, illumination, exposure and interpretation notes with the result.
Distinguish target-plane and image-plane resolution
Let the absolute lateral magnification be M = image size / object size. A target frequency f_object forms an image-plane frequency f_image = f_object / M. For example, group 2, element 3 is about 5.04 lp/mm at the target. At M = 0.055, it corresponds to about 91.63 lp/mm at the sensor or film plane. The two values describe the same imaged feature at different planes.
Compare the centre and edges of the field
A centre-only reading can miss off-axis image degradation. Measure the centre and selected edge or corner positions at the same working plane. State whether focus stays fixed or is optimised separately at each position: fixed focus tests the field at one setting, while refocusing measures local best-focus performance. These answer different questions.
Avoid false resolution and misleading comparisons
Beyond the true resolution limit, bars may merge and then seem to reappear because of aliasing, contrast reversal, defocus or image processing. Do not accept an element merely because stripes are visible: the correct three bars and their spaces must be identifiable. A USAF threshold result is also different from an MTF curve, pixel count or camera dimensional accuracy.