Conflicts:
	src/site/fml/faq.fml

The content of this document is now fully available in markdown syntax,
both in the statis site that can be locally generated using maven and
online in the forge wiki pages.

This allows using polynomial calibration models provided by the image
quality fitting process.

There are no changes in the library code, only tests cases were added.

TileUpdater must not return a tile fow which the specified
latitude/longitude are in the last row/column, as these row/columns are
considered not covered by this tile but are used only to interpolate
points in the preceding row/column.

This should fix some problems in the vicinity of ±π longitude.

The problems were encountered in direct localization. Some line-of-sight
was almost parallel to a latitude boundary, at 1.0e-14 level. This
triggered an exception in the Duvenhage algorithm.

Thanks to Silvia Ríos Bergantiños and Beatriz Salazar for the
translations.
        

Thanks to Marina Ludwig for the translation.

Thanks to Francesco Coccoluto for the translation.

When a line sensor was regularly curved, the mean plane corresponding to
the line of sights of all pixel was slanted too much and did not contain
the origin anymore.