Presentacin de PowerPoint

1
The Cyborg Astrobiologist Teaching Computers to
Find Uncommon or Novel Areas of Geological
Scenery in Real-time Network Conference of
the Alexander von Humboldt Foundation, Muenster,
Germany, November 2008. Authors Patrick C.
McGuire1,5,6,, Enrique Díaz Martínez1,3, Jens
Ormö1, Javier Gómez Elvíra1, Virginia
Souza-Egipsy1, Helge Ritter2, Markus Oesker2,
Robert Haschke2, Jörg Ontrup2, Florian Schmidt2,
Alexandra Bartolo4, Richard Bose5, Lorenz
Wendt6 Institutions 1Centro de
Astrobiología (CAB, INTA/CSIC), Instituto
Nacional Técnica Aeroespacial, Torrejón de Ardoz,
Madrid, Spain Robotics Laboratory, Planetary
Geology Laboratory, Transdisciplinary
Laboratory 2Universität Bielefeld, Germany,
Neuroinformatics Group, Computer Science
Department, Technische Fakultät 3(currently at)
Dirreción de Geología y Geofísica, Instituto
Geológico y Minero de España, Tres Cantos,
Madrid, Spain 4Engineering Faculty, University of
Malta, Malta 5McDonnell Center for the
Space Sciences, Washington University, St. Louis,
Missouri, USA 6Institute for Geosciences, Freie
Universität Berlin, Department of Planetary
Science and Remote Sensing, Berlin,
Germany (corresponding author)
Email mcguire_at_epsci.wustl.edu
The Cyborg AstrobiologistGeological Field
Missions to Rivas Vaciamadrid, Guadalajara and
Malta
McGuire, Ormö et al., Intl Journal of
Astrobiology (2004)
McGuire, Diaz-Martinez et al., Intl Journal of
Astrobiology (2005)
CONCEPT
Question Can an autonomous Robotic
Astrobiologist someday be taught to understand a
geological scene like the one pictured below?
Cyborg Astrobiologist -- Basic Idea
Geologists intuition high-level planning
Mission Riba1 Tripod Position 2,Inside the
Cyborg Astrobiologists Brain
McGuire, Diaz-Martinez et al., Intl Journal of
Astrobiology (2005)
Real-time Selection of Interest Points
Wearable Computer
Real-time Image-segmentation
VISION ALGORITHMS
Human Mobility
Video Camera, Firewire communication
(Cyborg Astrobiologist)tested at Rivas
Vaciamadrid (Sept 6, 2005)
Uncommonness
Novelty DetectionCyborg Astrobiologist gets a
memory (of Familiar patterns)
Uncommon Map
Orig (Cam or Micro)
Full-color Image Segmentation
Novel Segments
Cyborg Astrobiologist, with microscope, studying
lichens
1(Cam)
FIELD TESTS
Microscope in Field!
18(Micro)
LichenWhite Sporing Red RockBrown
Tripod!!
Not Yet
19(Micro)
LichenWhite Sporing Red RockBrown
Head-mounted Display failed in
field ? Pictured here tablet display
Not Yet
23(Micro)
Lichen Algae Fungus (in Symbiosis)
Lichen Black, Orange Gypsum Crystal
SmoothGray
Summary of Concept Developing computer-vision
technologies to automate some of a Human
Geologists low-level thought processes. We are
testing these computer vision technologies in the
field in real-time with low-cost platforms, with
the intention of later putting these algorithms
on robotic platforms for exploring the Moon and
Mars. Currently working with collaborators in
Madrid, Malta, St. Louis, Bielefeld and Berlin,
in order to a) Do more comprehensive field
tests at sites of geological and astrobiological
interest using both the uncommon-map and
novelty-detection systems. b) Make the
phone-cam system work with a field laptop enabled
with Bluetooth, for automatic transfer of the
images from the phone to the laptop and
(marked- up) back to the phone. This should speed
up the testing. Future 0) Improved
image-segmentation (to include texture)
1) Improved Novelty Detection (better
statistical measures) 2) Improved
cameras (hyperspectral cameras) 3)
Improve technological readiness for planetary
robotics missions.
The Cyborg AstrobiologistPorted from a wearable
computer to the Astrobiology Phone-cam
Before
HARDWARE IMPROVEMENTS
Now
Bartolo, McGuire et al., International Journal of
Astrobiology (2007)