Mathematics – Logic
Scientific paper
May 2004
adsabs.harvard.edu/cgi-bin/nph-data_query?bibcode=2004agusm.p33d..07m&link_type=abstract
American Geophysical Union, Spring Meeting 2004, abstract #P33D-07
Mathematics
Logic
5464 Remote Sensing, 5494 Instruments And Techniques, 6225 Mars
Scientific paper
Currently Mars missions can collect more data than can be returned. Autonomous systems for data collection, processing and return will aid future Mars rovers in prioritizing and returning geologically important information. We have created a neural net detector that is able to successfully recognize carbonates from Visible/NIR (350-2500 nm) spectra of rocks collected from Mars analog environments (Gilmore et al. JGR 105, 29,223). In order to characterize and improve the detector's sensitivity, we are evaluating the performance of the detector under more realistic Martian environments: 1) carbonate minerals covered with palagonitic dust, and 2) carbonate minerals intimately mixed with basalt and palagonite. Lessons learned will be applied to the generation of additional detectors for minerals of interest (e.g., hydrothermal minerals). Aliquots of Martian Soil Simulant JSC Mars-1 palagonite were sieved to <45 microns and air fall deposited onto a calcite crystal and an adjacent glass slide. Spectra in the Vis/NIR were taken of the calcite after every layer up to a thickness of 270 microns (35 layers) with an Analytical Spectral Devices Field Spec Pro spectrometer operating from 350-2500 nm. The carbonate detector, operating over the range 2000-2400 nm, has an empirically established detection threshold based on training with thousands of synthetic linear combinations of laboratory mineral spectra designed to simulate expected Martian rocks. The detection threshold was reached in this experiment when the dust layer thickness reached 102 -116 microns (78-85% aerial coverage). This corresponds to a real change in the depth of the 2300 nm carbonate band (continuum depth, D= 1- [reflectance at trough center/reflectance at continuum] = ~0.1), and is similar to the detection threshold of the human experimenter. Very thin (~10-20 microns) coatings of palagonite dust had a large effect on the spectral response of the substrate, exemplifying the nonlinear mixing of the two components. We will report on experiments mixing known quantities of carbonate with basalt and palagonite at various grain sizes. We plan to assess the detector's ability to discern and quantify varying amounts of carbonate within the mixtures, and model the spectra as nonlinear mixtures to ascertain if any improvements in the accuracy of the neural net can be achieved. Finally, we will report on the generation and performance of a detector capable of identifying phyllosilicate minerals individually and as a mineral class.
Bornstein Benjamin
Castano Rebecca
Gilmore Martha S.
Greenwood J.
Merrill Matthew D.
No associations
LandOfFree
Generation and Testing of Autonomous Mineral Detectors for Mars Rovers does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Generation and Testing of Autonomous Mineral Detectors for Mars Rovers, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Generation and Testing of Autonomous Mineral Detectors for Mars Rovers will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-1177795