Bug #2244: Pixel (Line/Sample) Projection Offset Issue
mimap2isis - Import PDS formatted Kaguya MI MAP file to ISIS3 cube format
Mentor: Trent Hare
As part of the pixel offset issue (import/export)
#6 Updated by Trent Hare over 1 year ago
These are going to be preliminary numbers I'm not sure what the correct values should be!
(1) grab map projected Kaguya L35 data (please place in testing location):
(2) A bug was just fix and not in stable so first set isis. This fix correctly adjusts the incorrect PDS SAMPLE_PROJECTION_OFFSET
log into prog21 to run beta area
$ isis /work/projects/isis/latest/m02358/isis
(3) run conversion
$ mimap2isis from=MIA_3C5_03_02024S140E3586SC.img to=MIA_3C5_03_02024S140E3586SC_mimap.cub setnullrange=yes nullmin=-30000 nullmax=-30000
(4) test ProjectionOffset X,Y using maplab (in typical ISIS latest area)
maplab from=MIA_3C5_03_02024S140E3586SC_mimap.cub map=MIA_3C5_03_02024S140E3586SC_maplab.cub sample=0.5 line=0.5 coordinates=latlon lat=-13.73339844 lon=358.31103516
Now look at X,Y offsets in both ISIS cubes. They should be the same (but are not).
Just using the pure Simple Cylindrical map projections equation they should be
X = 10865191.07999640 (X = min_lon * PI/180.0 * 1737400) where min_lon = 358.3110352
Y = -416442.65341059 (Y = max_lat * PI/180.0 * 1737400) where max_lat=-13.73339844
ISIS sets them to these values (mimap.cub)
UpperLeftCornerX = -51207.65772 for a difference of 7.41497739
UpperLeftCornerY = -416449.95948 for a difference of 7.50025059
ISIS is apparently also wrapping the the min longitude 358.3110352 (min_Lon - 360) to -1.68896484 (which must be another issue in the label). Perhaps ISIS is running (min_lon - 360) for Longitudes above 180? Anyway, using this the map projection equation would be:
X = -51215.07269739 (X = -1.68896484 PI/180.0 * 1737400)
If ISIS is converting the Lons then it should be reflected in the ISIS output label. Currently they are still being reported as ~358 degrees.
X is wonky but seems correct but with a 1/2 pixel offset
Y seems correct but but with a 1/2 pixel offset
So products in PDS standard are suppose to apply a PIXEL_OFFSET by adding 0.5 pixels to get to the corner from the pixel center. For this label the offset should be 0.0.
The PDS updates resulted in the following change to the ISIS cube (tested by the mimap2isis default test):
UpperLeftCornerX = -970339.58436 <meters> UpperLeftCornerY = 1971009.91524 <meters> to UpperLeftCornerX = -970354.39068 <meters> UpperLeftCornerY = 1971024.72156 <meters>
this is consistent with the 1 pixel offset at a scale of
MAP_SCALE = 1.480632e-02 <km/pixel>^M