ENVISAT driver patch

[pacesm]

As discussed at the OSGeo CodeSprint 2014 with Even Rouault, EOX prepared a patch for the ENVISAT products' driver adding following fixes and enhancements:

• MERIS tie-points (GCPs)

The current driver fails to extract GCPs from tiepoints ADS records not exactly matching the products' MDS records.

The patch implements a robust TP-ADS/MDS record matching based on the comparison of the records' time-stamps.

• MERIS tie-points DEM corrections

The patch applies the MERIS DEM tie-point corrections stored in the products TP-ADS, if present.

• unwrapping tie-points for antimeridian crossing products

The tie-points in the ENVISAT products are in WGS84 lat/lon coordinates wrapped around in the +/-180 projection bounds. This posses a problem for the product crossing the atimeriadian as there is a discontinuity in the GCPs not well handled the GDAL transformers.

The patch unwraps the ENVISAT product GCPs returned by the driver so that they can be safely passed to the GDAL transformers.

[pacesm]

patch

[Even Rouault]

Antonio, I've applied the following patch in the Envisat driver. I see that you have contributed the autotest for that driver, so I'd appreciate your testing and reviewing.

trunk r27098 "Envisat: implement more reliable way of extracting GCPs from Meris tie-points; add DEM corrections of TP-ADS products when present; workaround dateline discontinuity in GCPs so they can be used with GDAL warping transformers (patch by Martin Paces, #5423)"

[Even Rouault]

trunk 27099 "Fix Windows build"

[antonio]

Antonio, I've applied the following patch in the Envisat driver. I see that you have contributed the autotest for that driver, so I'd appreciate your testing and reviewing.

Sure! At a first look the patch seems to be OK. I will perform deeper testing tonight.

[antonio]

OK tested on some ASAR and MERIS product. All works correctly. Also I added a small test in r27101. IMO you can safely close now.

[Even Rouault]

Thanks for your feedback and the test Antonio.