This is done by dumping its state into a file, thus avoiding costly
initialization.

When only latitude and longitude are specified, the elevation is
automatically computed from the Digital Elevation Model.

The new name explains more clearly that several sensors can be set up at
once.

We now use a single position for all pixels (but of course still
different line-of-sight vectors), as it doesn't really make sense to
consider different positions as pixels are only a few microns away from
each other.

The former implementation ignored the flat-body flag and in fact
corrected the flat-body ...