Forum

l2flags is a bit field.  Your example breaks down like this:

decimal = 2112
hex = 0x0840
binary = 0000 1000 0100 0000

That means bit 6 and bit 11 are on.  The first bit is on the right and you start numbering the bits at 0

bit 6 = 64 = COASTZ
bit 11 = 2048 = TURBIDW

don

Your XML file looks like you are asking for 9 millimeter wide pixels.  That would be too many pixels, which you do not have enough memory to store.  You are trying to allocate about 5 PetaBytes of memory.

don

Look at the function:

get_minmax_rowcol()
file: smigen.cpp
line: 2324

Looks like line 2331 does truncate.  Unfortunately, I like the way l3mapgen does it better.

don

Barry,I processed the bin file you sent me with smigen and l3mapgen.  I get the same results as you with l3mapgen having an extra line.  It looks like smigen is using a float for the resolution calculations and l3mapgen is using doubles.  The only thing I came up with that could help ...

I just made a machine with Ubuntu 16.04, installed cmake 3.18.0, git, bison, flex, lib32z1-dev

opt/src would not build unless I deleted

opt/bin
opt/lib

then BuildIt worked fine.

don

Sure,

Not sure if it will work, but send it to my email:

don.shea@nasa.gov

don

Can you put a bin file that exhibits this problem somewhere where I can grab it.  Then send me the exact command line for smigen and l3mapgen and I will take a look at this.

don

l2gen will only output 20m resolution as Sean explained earlier in this thread.  There are not enough bands at 10m for l2gen to do any useful ocean calculations.  Sorry, but we do not have any plans to output 10m resolution for Sentinel-2 MSI.

don

How did you get the lib3 source files (ocssw/opt/src)?  When using the SeaDAS 7.5.3 GUI or the install_ocssw.py script you will get proj4 version 4.9.2.  It is true that our latest in-house development code is using proj version 7.0.0, but it is strange that you have that version using ins...

Barry,

My first guess is that since smigen is deprecated it does not know anything about JPSS.  It is probably using the wavelengths for NPP which has 551 instead of 556.  Might I suggest using l3mapgen which is what the production system here uses to make SMIs.

don