How The ENL (Expected number of looks) calculated in SNAP?

I used IW GRD VV intensity product for ship detection, when I check the statistics of data in analysis tab i see mean, std deviation median intensities coefficient of variation and ENL. This ENL shows very small value?
My first Question is How This ENL is computed in following data,

Only ROI-mask pixels considered: No
Number of pixels total: 36696373
Number of considered pixels: 36696373
Ratio of considered pixels: 100
Minimum: 64
Maximum: 9.79E+08
Mean: 28265.23069
Standard deviation: 453270.3386
Coefficient of variation: 16.03632192
Median: 490225.745
ENL 0.0039

Q2. Is there any relation of This ENL with Speckle noise? When we calculate ENL by normal procedure that is square (std deviation /mean) is not matching with the given ENL


In SAR data like you’re using, ENL helps understand speckle noise. It’s calculated based on the ratio of the square of the mean to the variance of the pixel intensities in the image. In a simplified form, it can be expressed as (Mean²/Variance).

Your low ENL of 0.0039 suggests high speckle noise, which means your image might not be as smooth as expected. A higher ENL value generally signifies less speckle noise and a more ‘smooth’ image.

I think that the mismatch in ENL calculation could be due to how your software processes data, especially with ROI considerations.


Thank you @MysticMosaic, Yes, you mentioned correctly. I am trying this using SNAP toolbox. Provided data is for sentinel image data. taking statistical analysis of the image data. I got this result, hence I get confused about the ENL calculation in SNAP toolbox


This thread provides an excellent summary of how the ENL can be computed from a portion of a SAR image

The ENL can be estimated from the mean and standard deviation of an homogeneous portion of the image. This portion shall effectively contain signal (that is: it should not be selected on dark areas corresponding to thermal noise).