Transcript Dr. Clark R. Chapman - Southwest Research Institute

Technology Capability Development for Identification and
Interpretation of Martian Craters and Climate History
P.I.:
CoI’s:
Clark R. Chapman (15)
William J. Merline (15)
Steve W. Dellenback (10)
Michael P. Rigney (10)
Michael J. Magee (10)
Collaborator: Prof. James Head III
(Dept. Geology, Brown Univ.)
Proposal submitted 11 March 2002 to:
Southwest Research Initiative for Mars
(SwIM)
Main Features of Proposal
 Strongly interdivisional collaboration to
develop latent talents to be competitive in
NASA’s future Mars exploration programs
 Focus on science issues (geological history
of water) central to NASA interests in Mars
 Strong technological component (Artificial
Intelligence and Data Mining techniques)
 Involvement with prominent Collaborator
 Modest cost (<$95K) sufficient to bring us up
to speed and make us competitive
Science Background: Crater
Degradation and Water on Mars
Sequence of degraded to fresh craters
 First evidence for “rainfall” on Mars was from
“river” valley networks (run-off vs. sapping)
 Craters provide baseline initial conditions to
assess subsequent modification of topography
 Variation in crater degradation classes (fresh to
very degraded) as function of crater diameter led to
hypothesis of “obliteration episode” on Mars,
contemporaneous with valley networks
 Martian craters show much greater variety than on
the Moon; voluminous data not yet studied
A wide variety of surface
modication processes...
 Many kinds of
processes, many
different signatures



Wind

Volcanism

No modification at all!






Lacustrine, oceanic
Volcanic
Aeolian (dunes, storms)
Tectonic
Glacial
Rivers and streams
Subterranean flow
Creep
Isostatic adjustment
Superimposed cratering
Etc., etc.
Degrees of Terrain Softening
None
Some
A Lot
Mars Crater Data Bases
 Original analysis for craters >8 km diam. from
Mariner 9 images only (critical sizes: 10 - 50 km )
 Vast additional imaging sets, with much higher
resolution, better coverage:
Viking Orbiter imaging
 MOC imaging (wide and narrow angle camera), MGS
 New THEMIS images (vis and IR) from Mars Odyssey
 Potential future missions

 Data cry out for cataloging, analysis of
morphology classes…but tedious effort has
inhibited progress, demanding a new approach
Background on Capabilities in
Div. 15 and Div. 10
 Analysis of Martian cratering statistics by
Chapman in 1970s…new data need analysis
 Previously developed A.I. crater recognition
algorithms, applied to simple lunar craters, by
Div. 15 (Merline/Chapman/et al.), collaborating
with JPL…enhanced techniques required for
much more complex Mars craters
 Expertise (although in non-planetary science
applications) in feature recognition and
classification technologies by Div. 10…can
now apply technical experience to Mars
Dr. Chapman’s 1970s
Research on Mars Cratering
 Developed
hypothesis of
erosional
“episode”
 Research
finished before availability of Viking
Orb. images
 Credibility...
but need to
“get up to
speed” on
current data,
issues
Annual Revs. Earth Planet Sci. (1977)
Icarus (1974)
Boulder Office/JPL Past Work on
A.I. Identification of Lunar Craters
 Singular Value Decomposition and “template” approaches
 About 80% reliability for simple cratered terrains; but we
desire >90% reliability for more complex Martian terrains
Output of template
approach to analysis of
simple lunar scene.
Identified craters are
color-coded (yellow =
most reliable). Blue
circles (slightly offset to
upper left) are human
identifications and
sizings.
Simple Cratered Surfaces…
and then there is Mars!
Mars
the Moon
asteroid Gaspra
planet Mercury
Division 10 Expertise in A.I.
And Data Mining
 Expertise in “expert systems” A.I. techniques to
approximating human perception/decisionmaking processes
 Expertise in Data Mining, which can be applied
to Mars crater data base to gain insights
 Past applications of circle-enhancing Hough
transforms to identify wheels, tools, fiducial
marks, rivets, etc.
Illustration of Decoupled Circular
Hough Transform method
Goals, Objectives, General
Approach
 Combine relevant, as-yet-unconsolidated skills
and experience in Divs. 10 & 15 to address new
research opportunities concerning Mars
 Leverage capabilities to begin addressing
fundamental questions concerning Martian
geological history, role and location of water
 Two arenas for development:

Further develop crater identification/classification in
Martian context



Evaluate and augment current lunar algorithms for Mars
Parallel use of Hough circular transforms, contrast enhancement
Evaluate how to interpret Mars crater forms in terms of
processes (in collaboration with Prof. James Head)
Tasks and Expected
Accomplishments
 Develop new, improved crater detection/ID
technology to address wealth of current/future
Mars imaging data
Div. 10 develops filters (e.g. contrast enhance) to preprocess Mars images for analysis by Div. 15 algorithms
 Div. 10 develops, in parallel, alternative (Hough transform)
methods to test on Martian images
 Divs. 10 & 15 collaborate on developing morphological
classification criteria that are practical, geologically useful
 Div. 10 studies “next steps” in classifying, data mining

 Develop interpretive methods in 3 study areas
Div. 15 & Prof. Head select areas, develop morph. indices
 Link indices to previous fresh-to-degraded crater studies
 Preliminary interpretations of geological history in 3 areas

An Example: Contrast
Enhancement
 In the case of the lunar work
previously done, craters are deep,
bowl-shaped depressions
 Martian features are degraded,
shallow and may have confusing
surface shadings
 Raw image (upper left) shows
subdued features, which become
much more prominent after contrast
enhancement (lower left)
 Other pre-processing filters may
improve algorithm success
Opportunities; Benefits to
the Institute
 Current NASA Data Analysis/Research programs:
we’ll have tools, credibility to propose
Mars Data Analysis Program (MDAP): prospects to study
Viking, MGS/MOC, and Mars Odyssey/THEMIS images
 New Mars Fundamental Research Program

 Enchanced prospects to be selected for future
Mars missions (specifics undergoing review, but
NASA commitment to Mars is clear)
Mars Express, Mars Exploration Rovers (Participating
Scientist and/or follow-on research prospects)
 Mars Reconnaissance Orbiter, sample return missions
 Specific landing site selection opportunities

 Non-Mars prospects: e.g. Earth remote-sensing
Program Plan and Schedule
1. Crater ID/Classification
2. Interpretive Methodologies
1 May 02
Explore interfaces (Divs. 10&15)
Explore interfaces (Divs. 10&15)
1 Aug 02
Preprocessing images to
enhance existing algorithms
Selection of 3 Mars study locales
1 Nov 02
Parallel (Hough) methods;
development of classification
criteria, data mining
Develop morphological indices,
link with prior studies
Analysis of “next steps” in
classification, data mining
Preliminary geological
interpretation of 3 study areas;
scientific publication
1 Feb 03
30 Apr 03
Personnel; Project Organization
 Overall project lead: Dr. Clark R. Chapman (15)
 Task 1 (crater identification technology)
Dr. William J. Merline (15; assisted by Mr. Brian Enke)
 Dr. Michael P. Rigney (10)
 Dr. Michael J. Magee (10)
 Dr. Steve W. Dellenback (10; lead on “next steps” task)

 Task 2 (scientific interpretation methodology)
Dr. Clark R. Chapman (15)
 Prof. James Head (collaborator, Brown University)
 Dr. William J. Merline (15)

 Other: trips to facilitate collaboration; no new
equipment. Total budget: $92,544
Past IR&D Work by P.I. and CoI’s
 Dr. Chapman: 1 previous IR, $93K, 1999-2001;
“white paper” received prominent, international
discussion and use; facilitated small grant
continuation; other contacts being pursued
 Dr. Merline: 1 previous QL, $28K, 9/98 - 1/99;
contributed to much subsequent NASA/NSF
funding of asteroid satellite searches
 Div. 10 CoI’s: 2 previous QL’s, 2 previous IR’s
totalling ~$356K, range from being currently
underway to leading to patents, aerospace/
government/industry opportunities
Conclusion
 Cost-effective way to develop latent skills,
enhance inter-Divisional collaboration, be in
good position for major roles in NASA’s chief
Solar System program: Mars exploration
 Strong technology component married with
strong science component -- ideal for NASA
 Prepares us to address some of the most
compelling issues in planetary science (role of
water on Mars) using state-of-the-art Artificial
Intelligence techniques
The End:
Back-up slides to follow
Modelling how time-variable erosion
affects crater morphologies
Total
fresh
slight
moderate
heavily
…degraded
(Chapman, 1974)
Obliteration time history
Signature of “episode” in
morphologic statistics
Among intermediate sized craters
(tens of km diameter), smaller ones
(~10 km) are most heavily degraded,
largest ones (>30 km) only modestly
degraded or nearly fresh.
 (a) sequence f,s,m,h
indicates incompleteness due to resolution
 (b) Mars data (Jones)
 (c) Sequence h,m,s,f
indicates obliteration
episode: smaller
craters are most
affected, largest ones
least affected
Absolute age of obliteration
 Early Mariner 9 interpretations
had obliteration tied to the
declining early cratering flux.
 Depending on calibration of
absolute ages, the obliteration
could have happened toward
the end of the decline (a), or
considerably later (b).
 But the important conclusion is
that it was decoupled from the
end of the early bombardment.
Mars
LHB
LHB on Mars?
 One Mars meteorite (and only
one: ALH84001) is very old and
has an Ar-Ar age of ~3.9 Ga:
statistics of ONE (Ash et al.,
1996)
Lunar rock degassing ages
 Meteorite degassing ages are
very “spread out” compared
with lunar LHB and somewhat
spread out compared with
lunar rocks
 Evidence is dissimilar!


Different impact histories
or
Different selection biases
Kring & Cohen 2002
