NAN values after data convertion

[marjanmarbouti]

Dear all,
I have two features from one image (intensity and coherence). I made a stack from both and then used ‘data convertion in SNAP’:
Data convertion….convert datatype….linear (slope and intercept)

But the problem is that it makes some ‘NAN’ values in my features (in intensity feature is more than coherence ).

Figure1. Part of ‘intensity’ feature before conversion.

Part of ‘intensity’ feature after conversion (you can see ‘NAN’ values there):

I tried several method but could not get rid off them . Any help?

[falahfakhri]

Might be the reason is, you selected int, and you could select float instead,

But you said,

@marjanmarbouti

I tried several method but could not get rid off them

So did you give a try?

[marjanmarbouti]

Thanks for answer. no, I only used unit8 but with all types of ‘scaling’.
Now I tried it with int16 and linear (slope and intercept); it seems to work
but really do not know why with unit8, it did not work.

[falahfakhri]

@marjanmarbouti

but really do not know why with unit8, it did not work.

Seemingly it depends on your datatype, as you posted here before,

Source of the post

[ABraun]

maybe you just have to remove 0 as nodata value in the band properties after conversion?