Hi Rich:
[
EDITED]
NOTE: Your *.INF "
Layer" parameter is set to "
Imagery"; thus you have told Resample that your source data is intended to be treated as
aerial imagery,
not Elevation data.
And it is possible that FSX SDK Resample 'may' ignore the
FractionBits parameter in your test scenario.
However, since that parameter has a potential impact when processing multi-channel data sets, one might wonder if use of FractionBits in the [Destination] sub-section of an *.INF file with a Geo-TIFF as a input source file "might" influence how Resample behaves when multi-channel aerial imagery is submitted, causing confusion resulting in a bias towards elevation processing rather than imagery processing in Resample's internal work-flow.
AFAIK, although aerial imagery does contain some raster data values which can be used for 'displacement modifiers' in establishing elevation values for terrain mesh, elevation processing requires specific data set file formats (including a *.INF file with correct parameters pertinent
only to elevation processing) ...to achieve normal processing to output terrain mesh BGLs.
One might wonder whether 'confused' internal processing by Resample might assume that it completed properly without actually having the data source file format it needed for elevation output, and thus fail to output either aerial imagery
or terrain mesh ...without having initially rejected the GeoTIFF file for processing and/or having output an error message ?
The "
Red" you see, as Ian alludes to above, is IIUC, a result of Resample processing input source file(s) to completion without actually getting required unique data for a geographic area, but which somehow wrote a BGL while failing to trigger an error report prior to / during processing by Resample.
Regardless, be aware that FractionBits is
not used in either the [Source] or [Destination] sub-sections of an *.INF file for Resampling
aerial imagery.
Instead, FractionBits is used
only in specific mixed
elevation data set value range scenarios when processing elevation data into terrain mesh.
So, it's best not to risk confusing the FSX SDK Resample utility; try your *.INF file again without the FractionBits parameter.
BTW: If you wish to experiment with using "
SamplingMethod=Majority" versus "
SamplingMethod=Gaussian" to create custom photo-real Land Class using aerial imagery, try using all other 'required parameters' which, IIUC (
I haven't tried this yet !) apparently must accompany that "Majority" parametric value.
Apparently, use of the "
SamplingMethod" parameter for aerial imagery as a form of custom land class (aka "Land Classification") requires additional parameters in the *.INF file to qualify use of that option.
See: "
SamplingMethod", "
MajorityMode", "
SampleWeight.n", "
Layer"
http://msdn.microsoft.com/en-us/library/cc707102.aspx#TheResampleTool
PS: An interesting related article on this latter topic as applied to deriving data from Land Use / Land Classification data sets:
"
Sampling Method and Sample Placement - How Do They Affect the Accuracy of Remotely Sensed Maps?"
http://www.asprs.org/a/publications/pers/2003journal/march/2003_mar_289-297.pdf
...And a explanation of methods used when 'resampling' raster data:
http://grasswiki.osgeo.org/wiki/Interpolation
http://grass.osgeo.org/grass65/manuals/rasterintro.html
[
END_EDIT]
Hope this helps !
GaryGB