VectorDataExtractROIApplication does a strange reprojection
Mantis Issue 966, reported by gpasero, assigned to gpasero, created: 2014-08-27
The output of the test apTvUtVectorDataExtractROIApplication clearly shows the problem. Inputs :
- vector data (roads over Midi-Pyréennées), projRef : WGS84 lon/lat
- support image QB_Toulouse_ortho , projRef : UTM 31N
Output :
- vector data (roads over Capitole area) , projRef : UTM 31N, the extent covers less than 1 square meter...
At the end of the application, a reprojection filter is used, but I don't think this reprojection step is needed. In addition, the same projRef is used in input and output of the reprojection filter.
1409156577 - gpaseroMy proposal : erase lines 155 to 162 in Application/Utils/otbVectorDataExtractROIApplication.cxx.
There is a dedicated application to do reprojection of vector data (Projection/otbVectorDataReprojection.cxx)
1409321533 - gpaseroTODO : rename into VectorDataExtractROI WARNING : same class name as the base filter Code/Common/otbVectorDataExtractROI.h
1409325840 - gpaseroI would rather rename it into "ExtractROIVectorData" This will avoid the names duplicated in the doxygen, and this is closer to the "ExtractROI" application, dedicated to images.
1409329458 - gpaseroApplication renamed to "VectorDataExtractROI" (less change for users) : http://hg.orfeo-toolbox.org/OTB/rev/b4b5bbb9e67c
Bug fixed : http://hg.orfeo-toolbox.org/OTB/rev/2fa32829c108 And a bonus fix of #942 (closed) : http://hg.orfeo-toolbox.org/OTB/rev/53b6813ee642
1429262562 - julienClosing resolved bugs.