Local Incidence and Emission computed wrong for certain types of images
- Local angles are calculated correctly for the data mentioned.
- If this calculation isn't right for this instrument, it isn't correct for other pushframes.
- (If calculation isn't right for this instrument, and is correct for other pushframes, the camera model is messed up)
This certainly fixes the problem in phocube. It would just be a temporary fix. The problem is the local slope is computed in qview, campt and several other programs using sub-pixel coordinates. So for example, in campt or when zoomed in qview at line 14.2 adding 0.5 would cross the framelet boundary and the local slope would be computed incorrectly. I looked at the code and wonder to myself why we just didn't compute the slope strictly from the DTM?
For now if you are stuck I would make the quick change to adding .499 and just be aware that about campt and qview. I will get a ticket into our system to see if we can get this bug repaired correctly.
In terms of impacts on other cameras it would obviously impact MARCI which is identical to the WAC. I also suspect it would impact the CRISM camera model; it collects non-contiguous data because of an onboard gimbal. I can only see an impact to LROC NAC if the spacecraft attitude is being adjusted to paint ground that has already been covered. I've seen this only in images where folks wanted to try to collect super resolution images (I think MGS MOC tried this).
On Tue, Mar 5, 2013 at 9:29 PM, Adam Licht email@example.com wrote:
We've recently come across an issue when running phocube on WACs. Specifically when using the 'localemission'
and 'localincidence' options. The last line of each frame does not seem to be continuous with the rest of the
frame (I've attached a screenshot of this).
I've looked into this a bit and seem to of isolated the problem. When the camera create the local normal
vector, it takes a lattice of points around the pixel (+/- 0.5 in each direction). When this is done on the last
line in a frame, the 'bottom point' ends up in the next frame, causing the error. When changing these values to
0.49 the output for the WAC looks as it should. This would be a quick fix, however, I do not know how this would
effect other cameras.
There is a cub that can be used to reproduce this at '/home/ser/alicht/tmp/4janderson
An example command line to generate the error is:
'phocube from=/home/ser/alicht/tmp/4janderson/M146803466C.vis.even.cal.cub to=out.cub phase=false
emission=false incidence=false localemission=true localincidence=true latitude=false longitude=false'
Do you have any insight regarding if this effects any other cameras? Or perhaps a better fix?
Lunar Reconnaissance Orbiter Camera Science Operations Center
Arizona State University
School of Earth and Space Exploration
Tempe AZ 85287-3603
Supervisory Computer Scientist
Astrogeology Science Center
2255 N Gemini Drive
Steps to reproduce:
Run phocube on a wac image and output the local incidence or emission images