search

CNES / MAJA

Level-2A processor used for atmospheric correction and cloud-detection. The active repository is the one below, this one is kept to leave access to the older issues.
https://gitlab.orfeo-toolbox.org/maja/maja
Apache License 2.0
139 stars 25 forks source link

Error while loading EEF file. #36

Closed Slavko12345 closed 4 years ago

Slavko12345 commented 4 years ago

Hi!

I am trying to run MAJA (v.3.3.2) without CAMS option on L1C tiles downloaded from peps. At the moment I am getting such error: vns::Data::ERROR: EarthExplorerXMLFileHandler(0x7af7d0): XMLFileHandler::LoadFile: Error while loading the file [.../Start-MAJA/temp_files/T22KHB/22KHB/GIPP_united/in/S2A_TEST_GIP_L2COMM_L_ALLSITES_00001_20190626_21000101.EEF]. Parsed with errors.

Will be grateful for any ideas or suggestions.

olivierhagolle commented 4 years ago

Hi Slavko,

we have had issues with products that did not include the full information requested by MAJA, regarding viewing angles for instance, but it was with early S2 products.

Does it happen for each product you process, or just for this specific one ? Could you please provide us the command line, log printed on the screen, and log file of the product for which the error is observed ? Thanks in advance, Olivier

Slavko12345 commented 4 years ago

I tried different dates for the same tile - it resulted in the same error. By the way, I found that I can use another tile for which 2A data is available, so at the moment I will close the issue. By the way, thank you for the development and continuous support of the project - it is great!

olivierhagolle commented 4 years ago

Hey, don't give up. there was an error in the xml file (a missing bracket '<' in the xml) If you download the GIPP files again, this error should not happen anymore. Olivier

Slavko12345 commented 4 years ago

In fact, 4 L2COMM GIPP files for Sentinel2-TM in the repository also have the same bug (at line 111, absence of '<'), after fixing that it started working! Thank you for pointing that out!

olivierhagolle commented 4 years ago

I did the correction too quickly, I must have forgotten to push the file, sorry. Glad it works ! I will be happy to get feedback Olivier

Slavko12345 commented 4 years ago

At the moment I run the script without CAMS option and compared result with neighboring tile for which 2A data was available (at the same day). It looks similar, but there is some color difference (probably because of CAMS). So in the nearest future I will probably run it with CAMS and give more feedback. Thanks again!

olivierhagolle commented 4 years ago

Hi Slavko, Theia is not running with CAMS option either (as ECMWF changed the CAMS format in July and our adaptation is not fully ready yet). The colour difference you seem to have observed probably lies in the residual inaccuracy of MAJA, given that the tiles are processed independently (and although MAJA should have lower residuals than usually observed with Sen2cor). Best regards, Olivier

Slavko12345 commented 4 years ago

Ok, thanks for the info! I have one more stupid question. Let's say I want to use WASP for L2A->L3A conversion. Is there any way to use it on L2A data from scihub.copernicus (2A data is available there, but it has a different format)? Maybe, there is a way to convert one format to another, or there are some missing parts in copernicus format? Maybe, the only way is to convert 1C data to 2A using MAJA first? Thank you again!

olivierhagolle commented 4 years ago

Hi Slavko, we did not write an interface to Sentinel-2 L2A from ESA. We could imagine it, but WASP is not adapted to Sen2cor (the processor behind ESA's L2A). WASP requires a very good cloud mask, and, well, Sen2cor cloud masks are not good enough, and would result in data gaps, and artefacts Olivier