How to interpret the Dynamic Range and Signal to Noise Ratio (SNR) in …

Posted by Gretchen  Alper on Wed, May 18, 2016

With machine vision applications, some of the most important specifications beyond resolution and frame speed to determine whether the camera meet the measurement requirements, are full well capacity, Signal to Noise Ratio (SNR), and dynamic range (DR) specifications.  Interpreting these values from specification sheets can be challenging though.  The full well capacity and SNR definitions that are used for the image sensor do not always match those that are used in the specification sheets of the resulting industrial camera. Dynamic range in particular can be confusing as there are different ways to calculate it. 

As an example, let’s consider the OnSemi Vita25k which is utilized in the Adimec S-25A70/CXP and S-25A30/CL camera to dig into these specifications. 

The number of electrons which can be contained in a pixel is referred to as the full well capacity.  The sensor supplier specifies the maximum full well figure, that is 22 kel. In the lower range of the full well the number of electrons increases linear with the amount of light that is incident on the image sensor. In the upper range of the full well the sensor behavior is not linear anymore. As the camera is mostly used for automated measurement and inspection systems, we optimize the camera for the requirements of these applications.  The linear part is the only useful part in these cases. Adimec measures all sensors characteristics and from these measurements we can tell which is the linear part of the full well. This is only until 13 kel.

(For other applications like outdoor for example the linearity is of less importance and other camera and other design decisions may be made.)

The SNR in a specification sheet is determined by taking the maximum signal (Full well) divided by the shot noise of the maximum signal, or SQRT(Full Well).  However, if you would use the full well as defined by the sensor manufacturer you include the non-linear operation regime. Therefore, we measure most specifications ourselves to identify the usable working regimes.

Dynamic range is defined as the ratio of the largest signal the image sensor can handle (linearly) to the readout noise (in the dark) of the camera system. From an intuitive point of view, dynamic range defines the brightest and darkest data in a given image which can reliably and faithfully be reproduced by the camera electronics. [source: Quantum Imaging]

DR and SNR calculations for both the 2 situations (first 22 kel and later 13 kel)

The Sensor manufacturer provides the following specs:

If you would use the full well as defined by the sensor manufacturer you are already far in the non-linear operation regime.

With our measurements it was identified that up to 13000e the sensor was still working linear. Creating the same table as above gives:

To make it a little bit more confusing, Camera manufacturers set their analog to digital converter (ADC) such that they will convert a maximum well size that is somewhat larger as compared to the well size for the linear regime. By doing this it is assured that the full linear regime can be used without having clipping problems.

Actually in our manufacturing process we test all our cameras and measure the dynamic range and the ADC accessible full well. From averaging the result of 45 manufacturing tests on S-25 cameras we can provide the specifications as listed in the following table:

Conclusion

How to report the DR and SNR depends a lot on how you define the full well. How you define the full well again depends on how much deviation from linear is allowed for your application. The highest dynamic range and signal to noise that the camera can achieve is given by the last table. However, in some applications you might be limited already by the dynamic range given in the second table.

The next time you read a DR, SNR or Full Well Capacity, you should thus always check if a definition is provided. If this is not the case, ask the manufacturer for the definition.

http://info.adimec.com/blogposts/how-to-interpret-the-dynamic-range-and-signal-to-noise-ratio-snr-in-image-sensor-and-industrial-camera-specifications