Johnson's criteria
An old rule of thumb from imaging research: about 1.5 pixels to notice something, 6 to tell what kind of thing it is, and 12 to identify it.
What it means
In the late 1950s, John Johnson, a US Army night vision researcher, tested how much detail people needed to find and name objects through image intensifiers. His results became a set of thresholds, still used to calculate range figures for thermal and night vision devices.
Modern spec sheets usually restate them in pixels across the target’s smallest dimension:
- Detection: about 1.5 pixels. Something is there.
- Recognition: about 6 pixels. It is an animal, and roughly what kind (a four-legged animal, not a person or a cow).
- Identification: about 12 pixels. It is a coyote, not a fox or a dog.
Johnson’s original figures were written in line pairs, a dark and a light bar together; two pixels make about one line pair. Each threshold is the point where an observer gets it right about half the time.
What it changes in the field
The criteria explain why box numbers and hunter experience disagree. Makers usually quote detection, the 1.5-pixel threshold. Hunters need identification, at about 12 pixels, which is roughly eight times as many pixels across the animal and so about an eighth of the distance for the same target size. Recognition is about twice the identification distance.
Its limits
The thresholds were built for military targets and trained observers, with good contrast. A coyote at dusk in tall grass, on a humid night, or seen only from the front is harder. A hunter who knows the local animals may call one sooner. Treat the thresholds as a fair way to compare devices, not a promise for any one night.
How we use it on this site
Our identification distance uses the 12-pixel threshold across a coyote’s 22-inch shoulder height. The pixels on target figures let you see where a model sits against all three thresholds at 100, 200 and 300 yards.