PubAg

Main content area

The fingerprint of linear dunes

Author:
Scuderi, Louis
Source:
Aeolian research 2019 v.39 pp. 1-12
ISSN:
1875-9637
Subject:
algorithms, automation, computer software, dunes, forensic sciences
Abstract:
Fingerprint ridges contain unique defects that allow identification of individuals through a combination of metrics that capture Y-junctions and terminations. Ripples, linear dunes and even Transverse Aeolian Ridges (TARs) on Mars share many of these features with closed Y-junctions where crests merge, open Y-junctions where crests bifurcate, and crest end-terminations. Unfortunately, manual extraction of defects from linear dunefields is extremely time intensive and is not consistently reproducible. Fingerprint minutiae extraction software, a mature tool set used extensively in forensic work, was modified and applied to a range of linear aeolian features to determine whether dune defects could be easily extracted from imagery and used to quantitatively characterize their arrangement. The algorithm produced consistent and repeatable output from a range of imagery sources and resolutions from aeolian environments on Earth and Mars in time frames that were orders-of-magnitude faster than manual identification and extraction approaches. The automation of the extraction of defect location, direction, type and quality of each feature identified allows improved characterization of the spatial arrangement linear aeolian landforms. This approach could significantly augment our ability to analyze these features resulting in an improvement in understanding the organization and dynamics of these features and how they respond to environmental change. Challenges, in the form of extracting defects from complex vegetated dunefields, still remain in adapting this software for use in an aeolian environment.
Agid:
6359798