NETD is one of the most important performance parameters for infrared imaging systems. It is a signal-to-noise figure which represents the temperature difference which would produce a signal equal to the camera’s temporal noise. In human language: NETD expresses the minimal resolvable temperature difference when the camera is used for relative imaging applications. If the NETD = 50mK then the camera can detect temperature differences as small as 50mK. Keep in mind that the NETD has no relation with the accuracy of radiometric calibrations.
Simply said the NETD is calculated by dividing the temporal noise by the response of the camera.
The temporal noise is typically measured by sampling a series of images on a uniform low blackbody temperature and calculating the standard deviation per pixel. With the use of a blackbody, a 1K response map is created by sampling a series of images on a low temperature and by sampling a series of images on a higher temperature. By subtracting the average response of the lower temperature measurement from average of the higher temperature measurement a response map is created that equals the temperature difference of the two measurements. The 1K response map is calculated by dividing the result to equal a response of 1K in analog to digital units (ADU). As a final step to obtain the NETD, the temporal noise is divided by the 1K response map.
Before you start measuring make sure that all forms of image processing are disabled. As an example NUC’s might contain a global gain or AGC algorithms might interfere with the raw data.
In practice this operation can easily be made with following equipment:
- Infrared camera + GUI to capture images in highest bit depth resolution.
- Very accurate blackbody source for uniform capturing
- ImageJ for calculating with images
- Sample 256 images @ 300K BB and calculate the temporal noise using the STD.
- Take 256 images @ 300K & Average them to AVG_300K
- Take 256 images @ 305K & Average them to AVG_305K
- Subtract AVG_300K from AVG_305K to obtain a 5K response map.
- Divide the 5K response map by 5 to obtain the 1K response map
Divide the noise by the 1K response map to obtain the NETD map