8.3. Mars Reconnaissance Orbiter CTX

CTX is a moderately difficult camera to work with. The processing time can be pretty long when using the Bayes EM subpixel refinement (subpixel-mode 2). Otherwise the disparity between images is relatively small, allowing efficient computation and a reasonable processing time.

In this example we use mapprojected images, which is the most reliable way to align the images for correlation. Mapprojection is discussed in Section 11.4.2 and Section 6.1.7. Note that mapprojection can slow down the triangulation step, but given that parallel_stereo performs the triangulation using multiple processes, that is not a concern.

This example’s recipe is is in the examples/CTX directory shipped with ASP (type ‘make’ there to run it).


Fig. 8.2 Example output possible with the CTX imager aboard MRO.

The images are for the North Terra Meridiani region.

Download the CTX images P02_001981_1823_XI_02N356W.IMG and P03_002258_1817_XI_01N356W.IMG from PDS, at:

The download commands are:

wget https://pds-imaging.jpl.nasa.gov/data/mro/mars_reconnaissance_orbiter/ctx/mrox_0031/data/P02_001981_1823_XI_02N356W.IMG
wget https://pds-imaging.jpl.nasa.gov/data/mro/mars_reconnaissance_orbiter/ctx/mrox_0042/data/P03_002258_1817_XI_01N356W.IMG

Convert the .IMG files to ISIS .cub files, initialize the spice information, and calibrate:

ISIS> mroctx2isis from = P02_001981_1823_XI_02N356W.IMG \
        to = P02_001981_1823.cub
ISIS> mroctx2isis from = P03_002258_1817_XI_01N356W.IMG \
        to = P03_002258_1817.cub
ISIS> spiceinit from = P02_001981_1823.cub
ISIS> spiceinit from = P03_002258_1817.cub
ISIS> ctxcal from = P02_001981_1823.cub to = P02_001981_1823.cal.cub
ISIS> ctxcal from = P03_002258_1817.cub to = P03_002258_1817.cal.cub

(Here one can optionally run ctxevenodd on the cal.cub files, if needed.)

Run parallel_stereo (Section 16.51) and point2dem (Section 16.56):

cam2map4stereo.py P02_001981_1823.cal.cub P03_002258_1817.cal.cub
parallel_stereo                                      \
  --stereo-algorithm asp_mgm --subpixel-mode 9       \
  P02_001981_1823.map.cub P03_002258_1817.map.cub    \
point2dem -r mars --stereographic --auto-proj-center \

Higher quality results can be obtained by using the aspm_mgm algorithm and mapprojection (Section 6).

CTX cameras can benefit from bundle adjustment (Section 12). It is is also strongly suggested to use CSM camera models for improved performance (Section 8.12). See Section for how to create CSM camera models for linescan cameras, including for CTX.

CTX cameras can exhibit jitter. How to correct it is discussed in Section 16.38.6.

8.4. Automated Processing of HiRISE and CTX

While he was at the University of Chicago, David Mayer developed a set of scripts for automating Stereo Pipeline for CTX and HiRISE images. Those scripts and more information can now be found at https://github.com/USGS-Astrogeology/asp_scripts.