There are a lot of misconceptions and confusion over thermal imagers that can result in hunters making regretted purchases. In this article, I will concentrate on the specifications that matter and how they should be applied to real-world hunting conditions.
Thermal imagers used for hunting vary greatly in specifications, each of which will be better for certain tasks. The most important specification for identifying a target is image resolution.
Part one: Image resolution.
The two most important considerations for spotting and identifying animals are sensor resolution (number of pixels) and field-of-view. A popular misconception is that smaller pixels form a higher resolution image. This is not the case, image resolution is determined by how many pixels are available to capture an image within the devices field-of-view, regardless of pixel size or pitch.
However, comparing two thermal imagers with the same focal length lens, each with the same number of sensor pixels, the sensor with smaller pixels will have a higher image resolution because its field-of-view is narrower.
To calculate image resolution, divide the horizontal linear field-of-view, usually expressed as metres at 100 metres (m@100m), by the number of horizontal pixels in the sensor array.
The following is a comparison between a 12μm and a 17μm thermal that have the same working resolution:
Example 1: 640×480 / 12μm sensor 35mm F/l lens, field-of-view 21.9m @100m
Step one: multiply field-of-view in metres by 1000 to convert to millimetres.
Step two: divide field-of-view in millimetres by the number of horizontal sensor pixels.
Equation: 21.9m x 1000 = 21900mm ÷ by 640 = 34.2mm.
At 100m each pixel resolves an area of 34.2mm. (smaller this number, higher the image resolution)
Example 2: 640×480 / 17μm sensor 50mm F/l lens, field-of-view 21.8m @100m
Equation: 21.8m x 1000 = 21800mm ÷ by 640 = 34.0mm.
At 100m each pixel resolves an area of 34mm
The above shows that a 12μm sensor with a 35mm lens has the same resolution as a 17μm sensor with a 50mm lens, because the same number of pixels are divided into the same (very similar field-of-view).
The examples that I have discussed are a linear measure of resolution, based on a distance from the observer of 100m. Resolution decreases as distance increases, so a 34mm @100m resolution becomes 68mm @200m, 136mm @400m and so on. The area that a single pixel resolves is known as the Instantaneous Field-Of-View or IFOV.
that is produced by ALL the pixels in a sensor.
Another method to calculate IFOV is angular, such as Minute-of-Angle (MOA). MOA remains the same at all distances but is more difficult to work with when applying to objects of a known height and distance.
Applying to real-world situations
The main consideration for choosing a thermal for target identification is distance. If the animals are relatively close, such as in bush, forestry or farmland, less resolution is required for identification because at closer distances, animals appear larger. A wide field-of-view is more important in these distances to reduce scanning time.
At longer distances, resolution is increased at the expense of field-of-view. This is fine however, because linear field-of-view increases proportionally to distance between target and observer.
How much resolution is required for target identification?
The Johnson Criteria is an industry standard that was formulated by the US Department of Defence for calculating target Detection, Recognition and Identification (DRI) based on:
- Device’s pixel pitch of sensor and lens focal length
- Target size
- Target distance from the observer.
According to the Johnson Criteria, the difference between Detection, Recognition and Identification is quoted as:
Detection – “there is something there”
Recognition – “It is a vehicle”
Identification – “It is a Humvee”
We can use for Recognition “it has four legs” and Identification “it is a deer” as a better model for shooting.
While the Johnson Criteria is still a relevant model, it has a huge flaw in that it is based on a probability of 50% success, which means that an observer has a 50/50 chance of detecting and identifying what they are looking at. While this may be OK for detection, it is certainly is not good enough for identification.
These are the Johnson Criteria minimum number of pixels required for DRI based on 50% probability:
D = 1.5px +/- .5px
R = 6px +/- 1.5px
I = 12px +/- 2px
To convert to a theoretical 100% probability, I have doubled the amount of pixels:
D = 3 pixels
R = 12 pixels
I = 24 pixels
Note that I have used the word theoretical, because this formulae does not compensate for lens and sensor quality, user experience and viewing conditions, all of which affect the outcome. It does, however, provide a rule-of-thumb guide when choosing your thermal and a starting point for getting to know your thermal’s capability.
Applying our modified Johnson Criteria figures
Firstly, it must be noted that these figures are based on a stationary object, if the object is moving it greatly improves identification.
Calculating:
- Take the average height measurement of your target in mm’s,
- Divide this by the resolution in millimetres of your thermal.
Example: 800mm high pig and a thermal imager with a resolution of 34mm @100m
- Divide the height of the pig by the resolution of the thermal to find number of vertical pixels that fit into the height of the pig.
Equation: 800mm ÷ 34mm = 23.5 pixels.
According to our modified Johnson Criteria, 24 vertical pixels are required to Identify a stationary object. Our figure of 23.5 pixels is pretty bang-on for Identification at 100m, Recognition at 200m and Detection at 800m, though these distances can increase if the animal is moving. Also, distances can also increase because animals are longer than higher in shape.
of an image that is produced by a SINGLE pixel in
the sensor. IFOV is a measure of the device’s
image resolution.
I hope this sheds a little light on the topic which I have tried to keep as simple as possible. I will be developing a distance guide for different animals, thermals and distances in the near future, though I find the maths rather challenging.
In the next article I will discuss sensor sensitivity and NETD, which is the second most important consideration when choosing a thermal optic for hunting.
