HR: 1340h
AN: S43C-1030    [Abstracts]
TI: Towards the development of Hyperspectral Images of trench walls. Robotrench: Automatic Data acquisition
AU: * Ragona, D E
EM: dragona@ucsd.edu
AF: IGPP, Scripps Institution of Oceanography, University of California, 9500 Gilman Dr, La Jolla, CA 92093 United States
AU: Minster, B
EM: jbminster@ucsd.edu
AF: IGPP, Scripps Institution of Oceanography, University of California, 9500 Gilman Dr, La Jolla, CA 92093 United States
AU: Rockwell, T K
EM: trockwel@cox.net
AF: Geological Sciences, San Diego State University, 5500 Campanile Dr., San Diego, CA 92182 United States
AU: Fialko, Y
EM: fialko@radar.ucsd.edu
AF: IGPP, Scripps Institution of Oceanography, University of California, 9500 Gilman Dr, La Jolla, CA 92093 United States
AU: Fialko, Y
EM: fialko@radar.ucsd.edu
AF: Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109 United States
AU: Bloom, R G
EM: Ronald.G.Blom@jpl.nasa.gov
AF: Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109 United States
AU: Hemlinger, M
EM: mch@jord.jpl.nasa.gov
AF: Jet Propulsion Laboratory, 4800 Oak Grove Drive, Pasadena, CA 91109 United States
AB: Previous studies on imaging spectrometry of paleoseismological excavations (Ragona, et. al, 2003, 2004) showed that low resolution Hyperspectral Imagery of a trench wall, processed with a supervised classification algorithm, provided more stratigraphic information than a high-resolution digital photography of the same exposure. Although the low-resolution images depicted the most important variations, a higher resolution hyperspectral image is necessary to assist in the recognition and documentation of paleoseismic events. Because of the fact that our spectroradiometer can only acquire one pixel at the time, creating a 25 psi image of a 1 x 1 m area of a trench wall will require 40000 individual measurements. To ease this extensive task we designed and built a device that can automatically position the spectroradiometer probe along the x-z plane of a trench wall. This device, informally named Robotrench, has two 7 feet long axes of motion (horizontal and vertical) commanded by a stepper motor controller board and a laptop computer. A platform provides the set up for the spectroradiometer probe and for the calibrated illumination system. A small circuit provided the interface between the Robotrench motion and the spectroradiomenter data collection. At its best, Robotrench ?spectroradiometer symbiotic pair can automatically record 1500-2000 pixels/hour, making the image acquisition process slow but feasible. At the time this abstract submission only a small calibration experiment was completed. This experiment was designed to calibrate the X-Z axes and to test the instrument performance. We measured a 20 x 10 cm brick wall at a 25 psi resolution. Three reference marks were set up on the trench wall as control points for the image registration process. The experiment was conducted at night under artificial light (stabilized 2 x 50 W halogen lamps). The data obtained was processed with the Spectral Angle Mapper algorithm. The image recovered from the data showed an excellent classification of the three types of materials observed on the wall (brick, mortar and green felt) as well as an excellent geometrical restitution. More exhaustive experiment will be conducted on Carrizo Wash paleoseismic site during the fall of 2004.The use of Robotrench will provide an excellent platform to acquire large in situ hyperspectral images that can be used as a proof-of concept to design future devices.
DE: 7221 Paleoseismology
SC: Seismology [S]
MN: 2004 AGU Fall Meeting