This video segment focuses on 3 methods of measuring the apparent noise for a given sensor under test.  Noise is defined as the variation in the signal from a sensor that is not caused by the objects visible in the sensors direct field of view.  The apparent temperature difference we are able to detect with the sensor is directly related to its noise: the smaller the noise value, the smaller temperature difference that can be detected.  

This video will discuss 3 different methods for determining this value: Spatial NETD (Noise Equivalent Temperature Difference), Temporal NETD and 3D Noise.  

Spatial NETD takes a single image using a blackbody source. We measure the measurement variations over a specified region of interest. The standard deviation of those measurements is divided by the SiTF to produce the NETD (see previous YouTube video for more information on SiTF).
 
The second method is Temporal NETD which evaluates each element within a ROI of the sensor.  A series of image frames is collected, and then for each element, its corresponding pixel values are collected from all of the frames.  The standard deviation of those measurements is calculated for each pixel and then divided by the SiTF to get the NETD for that pixel. The set of NETD values for all the pixels within the ROI is treated as a population set and displayed in a histogram.

The third method is 3D Noise. A set of frames is collected in the same manner as the Temporal NETD.  However, now we use the entire data set, and we apply a set of directional calculations. This produces standard deviation values that are isolated along the three primary axes as well as the values for cross products of these axes. Dividing these values by the SiTF, we end up with 7 noise values that provide a lot more information about the source of the noise.

We hope you enjoy this video.  As always, feel free to provide feedback or comments on items you would like to see in future releases from Santa Barbara Infrared.