Reducing processing time required to produce l2gen formatted metadata

Use this Forum to find information on, or ask a question about, NASA Earth Science data.
Post Reply
ryan_e_oshea
Posts: 1
Joined: Thu Jan 06, 2022 2:07 pm America/New_York
Answers: 0

Reducing processing time required to produce l2gen formatted metadata

by ryan_e_oshea » Fri Jan 14, 2022 12:24 pm America/New_York

Hello,

I am attempting to create an 'imitation' l2gen netcdf that can be used with l2bin/l3bin from a level 2 file generated by an alternative atmospheric correction algorithm (e.g., ACOLITE) and algorithm (e.g., PC/Chl from a machine learning algorithm). I successfully did this by running l2gen on a single vertical line of the imagery, creating a new imitation netcdf using the metadata from the l2gen produced netcdf, and appending the water quality product calculated from the alternative water quality algorithm/atmospheric correction method. However, on high resolution imagery (e.g., MSI) this process can take a while to basically just generate l2gen formatted metadata for my application. Is there a way to only generate the l2gen metadata (loaded in SeaDAS as 'Global_attributes' and 'scan_line_attrbutes') for the netcdf, without estimating any product? Or is there a product I can specify that is rapidly generated across the entire image? I attempted to turn off atmospheric correction and select a l2prod that shouldn't take long to calculate (using the below arguments/keywords), but the process still takes a relatively long time.

I currently use '--epixl=2', '"--atmocorr=0"' and '--l2prod="pixnum"' as arguments/keywords to l2gen, to reduce required processing time.

Of course, I could write my own code to grab the metadata, and appropriately format it, but I was hoping there might be a way to speed up generation of the metadata using l2gen.

Thanks for your help,
Ryan

Tags:

OB SeaDAS - xuanyang02
Subject Matter Expert
Subject Matter Expert
Posts: 644
Joined: Tue Feb 09, 2021 5:42 pm America/New_York
Answers: 1
Been thanked: 1 time

Re: Reducing processing time required to produce l2gen formatted metadata

by OB SeaDAS - xuanyang02 » Fri Apr 22, 2022 11:06 am America/New_York

Try "--epixl=100", "--eline=100" as arguments

OB SeaDAS - dshea
Subject Matter Expert
Subject Matter Expert
Posts: 258
Joined: Thu Mar 05, 2009 10:25 am America/New_York
Answers: 0
Been thanked: 2 times

Re: Reducing processing time required to produce l2gen formatted metadata

by OB SeaDAS - dshea » Thu May 05, 2022 11:50 am America/New_York

MSI files take a long time to read into l2gen. Since we do not really support MSI data, we just whipped up a less than optimal reader. If you want the scan line data for each line, all the data will be read and it takes a long time. If you do not need the data for each line and just the global attributes, then using --epixl=2 --eline=10 will not take long at all.

don

Post Reply