How to handle large vector format file with OTB samples selection/extraction app

OTB Usage
1

Hello Everyone,

I try to manage pixels extraction with large number of features (>1000) and large number of pixels (i…e the file is bigger than 4Go).

I have issue with sqlite and shp format, because they can’t handle either the number of features or the size of the data. I tried also CSV but it seems that geometry are not exported correctly:

RuntimeError: Exception thrown in otbApplication Application_ExecuteAndWriteOutput: /home/mp/pernotg/scratch/otb-7.0/build_rh7-3.7.2/depends/ITK/src/ITK/Modules/Core/Common/src/itkMultiThreader.cxx:399:
itk::ERROR: MultiThreader(0x904b680): Exception occurred during SingleMethodExecute
/home/mp/pernotg/scratch/otb-7.0/otb/Modules/Adapters/GdalAdapters/src/otbOGRFeatureWrapper.cxx:209:
itk::ERROR: Cannot set the geometry:

Is there built-in capability to deal with such large vector format in OTB, without breaking into small chunck my time series ?

Thank for the help,

Mathieu

2

I reply to my self…
I found a solution that solve my problem: there is no problem with the sample selection app, since feature are not written. The problem comes with sample extraction app, and in that case, using the optional “out” parameter with an appropriate format works well. In my case I use CSV.
But for the sample selection app, a vector format that properly handle geometry should be used (e;g. shapefile or sqlite).
Mathieu

2 Likes