Searching for the
biological objects on Mars
A future rover, or
an astronaut, charged among other tasks with search for life on Mars, must
benefit from the today rover imagery [2].
The searcher,
trained to visually recognize the candidate objects like those that can be
observed and analyzed right now in the rover made pictures, will be much more
efficient than if we were to pretend that life search has yet to be done
totally from scratch.
Consequently, the
detection of life candidate objects and images, development of methods of their
analysis, of their concepts, and incorporation of results into the search for
life program, should be a priority.
Let us consider some
of the issues involved.
If in the process of
object shape recognition, we were able to obtain at will the desired pictures
and stereo pairs, we certainly could find out all that there is to know about
that object’s shape.
However, in case of
Martian imagery we have limited number of pictures of an object. How can we
then be able to see what we see, namely the features that those images actually
show?
One method geared
towards that goal, is conversion of the stereo pair’s information into 3D
physical model, which then can be examined at will.
Computer program
MARSOMETRY has been developed, which takes as input pixel coordinates of
related to the object of interest identifiable points, and produces as output
the orthogonal coordinates of those points, along with a map with those points
shown. The algorithm used in the program, is consistent with PANCAM optical
characteristics [4].
Before we take a
look at actual model of the geometry of possible Martian plant, let us consider
results for PANCAM calibration target [4].
Its geometry has
been recovered from the stereo pair taken by the Spirit on Sol 16,
2P127783627EFF0327P2367L7M1/R1M1.
The precision
obtained, apparently, should be sufficient for the shape recognition
applications arising from MER PANCAM imagery.
Figure 1 & 2 & 3
Let us introduce now
an interesting object photographed by Spirit on Sol 111, images
2P136234337EFF3600P2401L7M1/R1M1. Below are a red-cyan anaglyph made of the
clipped from originals stereo pair [1], and a color rendition [3].
Figure 4 & 5
For the modeling we
use 3 base plane points, 2 points for tracking stalk shadow, and 8 object
surface points. Below are the map produced by MARSOMETRY, and a picture of the
measurement model built. Modeling points located on the object surface,
represented by the blue balls, stalk top and bottom by the red balls, and tip
of stalk shadow by a green ball.
One has to conclude
that general appearance of object, as seen in the anaglyph, is consistent with
model.
Figure 6
Figure 7
It is possible to
create concepts of this object, which would remain consistent with the model
and object appearance. Here is one feasible example.
Figure 8
Although different,
those concepts would have enough of features that are present both in the model
and in the actual item - features that make this object a possible candidate
for a Martian plant [3].
As always, deep gratitude goes to NASA for making this work
and much more possible, by taking the magnificent pictures of Mars surface, and
generously providing full public access to them.
REFERENCES
1. NASA image gallery, http://origin.mars5.jpl.nasa.gov/gallery/all
2. Zeltsman, A. (2005) Visual Detection of life on Mars? Marsbugs,
12(20):2-3
3. Isenberg, H.
(2004-2005) PanCam true color images from Spirit and Opportunity Mars
Exploration Rovers,
http://mars.gh.wh.uni-dortmund.de/mer/spirit/111
4. Bell III J. F. et al. (2003) The Mars Exploration Rover Athena Panoramic
Camera (Pancam) Investigation, J. Geophys. Res., Special issue on the Mars
Exploration Rover mission, 73,83