Repository navigation
Missing Instrument Kernel (IK) for Chandrayaan-1 M3 — blocks spiceinit/cam2map #6165
Unanswered
dhakateparul9-lang
asked this question in
Q&A
Replies: 1 comment
|
@dhakateparul9-lang There were some missing files in the kernel database for chandrayaan1. Those files should be uploaded now, if you can try downloading the |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
ISIS version: 8.3.0 (conda-forge, usgs-astrogeology channel)
Instrument: Moon Mineralogy Mapper (M3) on Chandrayaan-1
Goal: Map-project an M3 L2 reflectance cube using the standard ISIS
workflow (chan1m32isis → spiceinit → cam2map), following the method
described in Suárez-Valencia et al. (2024), "MoonIndex"
(https://doi.org/10.1029/2023EA003464).
What works
well-known LBL trick of pointing ^RDN_IMAGE at the reflectance .IMG).
order: SPK, a DEM kernel database placeholder, CK, and FK for
chandrayaan1 — all downloaded via downloadIsisData chandrayaan1.
file, since none exists upstream for that folder (only the raw .TF
files are present, no selection database). Once done manually,
spiceinit proceeded past FK.
What's blocked
spiceinit finally fails at the IK (instrument kernel) step:
ERROR No existing files found with a numerical version matching
[kernels.????.db] in [.../chandrayaan1/kernels/ik].
Checking the ik/ folder (via both chandrayaan1_naifKernels and
chandrayaan1_usgs rclone remotes) shows only IK files for the other
Chandrayaan-1 instruments:
CH1_C1XS_V04.TI
CH1_SARA_V00.TI
CH1_SIR2_V01.TI
CH1_XSM_V13.TI
There is no M3-specific .TI/IK file anywhere in either remote. I also
checked for an iak/ (instrument addendum kernel) folder as an alternate
location — it's empty for M3 as well. I confirmed the FK file
(CH1_V02.TF) does not contain M3's focal length/pixel-pitch keywords
either.
Looking at Chandrayaan1M3Camera.cpp, the camera model explicitly reads
INS_BORESIGHT_SAMPLE, INS_BORESIGHT_LINE, and
focal-plane-map keywords via CameraFocalPlaneMap, so this isn't
optional — a real IK (or IAK) file is required for the classic camera
model to initialize.
What I also tried (modern CSM path)
I attempted the alternate csminit + ALE/CSM route instead of the
classic Camera class:
python3 -c "import ale; ale.loads('M3G20090813T213525_RFL.cub')"
This fails with:
Exception: No Such Driver for Label
ALE's driver auto-detection appears to require a Kernels group in the
label (normally populated by a successful spiceinit), so this path is
blocked by the same underlying gap.
Question
distribution, or is it hosted somewhere I haven't checked (e.g.,
bundled differently, under a different mission name, or distributed
only through PDS3 M3 archive documentation rather than the standard
SPICE kernel tree)?
community-maintained IK, or documented default focal length/pixel
pitch values I could construct one from) for map-projecting M3
global-mode cubes via cam2map?
All reactions