How To Test Thermal Imaging Cameras

By Deepshikha Shukla


Share and Enjoy !

Thermal sensitivity is the ability to distinguish tiny differences in temperature. It has a high impact on detection performance.

A thermal camera measures temperature of a surface by interpreting density of an infrared (IR) signal emanating from that surface and reaching the camera itself. Every object or environment includes materials with different levels of emissivity and reflectivity, and understanding these properties is vital for accurate temperature measurements. The ability to measure the surface temperature from a distance not only helps ensure safety of the person operating the camera but also ensures that the item being measured is not influenced by the measuring instrument.

Thermal imaging works on the principle of measuring reflected IR radiation from an object or environment to detect flaws or different objects. IR waves are a part of electromagnetic radiation whose wavelength is between 700nm and 1mm. However, there are three spectral ranges that are considered important for thermal sensor applications: 1µm to 3µm short-wave IR, 3µm to 5µm middle-wave IR, and 8µm to 14µm long-wave IR.


An IR camera used in thermal imaging needs to be tested for sensitivity to the temperature difference between an object and its background. This requires simulation of the object and its background by placing a target in front of the emissive surface of a blackbody to have a simultaneous and accurate measurement of both.

An IR target is a thin sheet with high emissivity coating and stuck patterns. The object is made by the blackbody seen through holes, while the background is the solid part of the sheet. Consequently, temperature of the object is temperature of the blackbody, while temperature of the background is measured by inserting a temperature sensor into the mount of the target.

Different targets for different tests

The most frequent targets for usual tests of an IR camera are: pinhole target for optical axis alignment of optronic system, square or rectangle target for field-of-view (FOV) measurement, grid holes for distortion image measurement, half-moon target for modulation transfer function (MTF) calculation and four-bar target for thermal resolution measurement. Noise tests usually do not require any target.

Fig. 1: Different targets for different tests (Credit:
Fig. 1: Different targets for different tests (Credit:

Tests that can be performed on an IR camera using a blackbody are divided into four categories, as follows:

  • Correction and calibration tests: linearity measurement, non-uniformity correction, thermal calibration of signal, dynamic range, etc
  • Noise measurement: thermal resolution, noise equivalent temperature difference (NETD), temporal noise, fixed pattern noise, spatial noise, etc
  • Spatial resolution and geometrical specifications: MTF, alignment, FOV, magnification, etc
  • Range evaluation: minimum resolvable temperature difference (MRTD), detection, recognition and identification (DRI) ranges, etc

Performance measurement parameters

Several factors influence the performance ranges, including spatial resolution, thermal sensitivity of the IR camera, FOV, shape and camouflage of the target, spectral band of detection and experience of the operator. To compare the performance of a system from different manufacturers, establishing a common test frame is required, in addition to listing all conditions influencing DRI range values.

NATO has developed STANAG 4347 standard that defines the target parameters resulting from size, shape, temperature, material properties and emissivity. It also defines conditions of use resulting from ambient temperature and type of background scene to take into account in MRTD computation, giving values of DRI ranges. This standard has been widely adopted in the thermal imaging industry.

MRTD is a standard performance measurement for thermal imagers. This measurement leads to determination of DRI ranges of the IR camera under test. DRI ranges—expressed in kilometres—can be found in the specification table of IR camera brochures. To select the right sensor that meets application requirements, it is essential to first define DRI ranges perfectly and then assess with regards to globally-adopted industrial standards.

Detection is the ability to distinguish an object from the background; for example, to detect a target several kilometres out of the background.

Recognition is the ability to classify the object class (animal, human, vehicle, boat, etc); for example, the target is recognised as a human walking along the fence.

Identification is the ability to describe the object in detail (man with hat, deer, jeep, etc); for example, two males wearing trousers and jackets are identified, and one is smoking.

John Johnson, a scientist from Army Night Vision Laboratory, has defined thresholds as the minimum number of line pairs to detect, recognise or identify targets captured by scene images. According to Johnson’s criteria, DRI lower limits are typically in the following range:

  • Detection limit: 0.75/1lp
  • Recognition limit: 3/4lp
  • Identification limit: 6/8lp

where lp is the line pair as one white line adjacent to a black line

MRTD measurement is performed using a differential blackbody as the IR reference source, a four-bar target as the reference object, positioned in front of the emissive surface of the reference source and a collimator projecting the object in front of the camera under test. A differential blackbody enables to set a positive or negative temperature difference between target bars and their background. The new generation of electronic controllers at the heart of the blackbody control and monitoring are ideally suited for testing and calibration of IR systems.

NETD can be considered as the thermal resolution that measures the temperature difference between an object and its environment, required to generate a variation of the IR camera signal equal to its temporal noise. The camera’s ability to capture tiny details at great distances is defined as spatial resolution. It is closely related to the number of camera pixels—the more the pixels, the better the spatial resolution will be and the greater the detection range.

Thermal sensitivity is the ability to distinguish tiny differences in temperature. It has a high impact on detection performance. Cooled IR cameras provide better thermal sensitivity than uncooled ones for the same number of pixels.

Fig. 2: Uncooled (left) versus cooled (right) camera thermal sensitivity
Fig. 2: Uncooled (left) versus cooled (right) camera thermal sensitivity

Effects of atmosphere

To sense the temperature remotely, you must account for a variety of surface and environmental conditions. Earth’s atmosphere will both absorb and emit IR radiation based on its ambient temperature, air density, smoke, snow, humidity, rain, fog, dust, and the distance between the camera and surface.

The amount of absorption or emission of IR radiation from the atmosphere has a marked effect on thermal imaging readings, so that must be taken into account. If measurements at longer distances are unavoidable, atmospheric conditions must be characterised and taken into account to calculate atmospheric transmission.

In addition, an increased number of pixels are required where thermal images are out of focus due to target or imager motion to make accurate radiometric measurements. There are various factors that can affect the accuracy of radiometric surface temperature measurements. Surfaces with high reflection and low emissivity can reduce the impact of reflections, exacerbated oblique reflections, impact of sun glints and background temperature reflection. This could also be achieved by combining a rough surface texture with high emissivity to cancel out the impact of high reflectivity and uncertain emissivity.


Share and Enjoy !



Please enter your comment!
Please enter your name here