8.3. Mars Reconnaissance Orbiter CTX¶

8.3.1. Overview¶

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 14.6.2 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).

../_images/n_terra_meridiani_ctx_combined.png — Fig. 8.2 Example output possible with the CTX imager aboard MRO.¶

The images are for the North Terra Meridiani region.

8.3.2. Download¶

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

https://ode.rsl.wustl.edu/mars/indexproductsearch.aspx

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

8.3.3. Creation of cub files¶

Install ISIS (Section 2.3.1).

Convert the .IMG files to ISIS .cub files, initialize the metadata (SPICE), 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 web = true
ISIS> spiceinit from = P03_002258_1817.cub web = true
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

Optionally run ctxevenodd on the cal.cub files.

The spiceinit command initializes the cub file metadata. The option web = true fetches the needed data on-the-fly. If it does not work, it is necessary to download this data manually, from the mro directory of the ISIS data area.

8.3.4. Running stereo¶

Run parallel_stereo (Section 16.52) and point2dem (Section 16.57):

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    \
  results/out
point2dem -r mars --stereographic --auto-proj-center \
  results/out-PC.tif

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

It is recommended to bundle-adjust the CTX cameras before stereo (Section 12). Then the cam2map4stereo.py script cannot be used, but mapprojection can be done with mapproject (Section 6.1.7).

8.3.5. Further processing¶

It is strongly suggested to use CSM camera models for improved performance (Section 8.11). See Section 8.11.2.1 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 16.39.7.

Shape-from-Shading with CTX images is illustrated in Section 11.13.

8.3.6. 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.