Spatial Point Process Models for the Clustering Behavior of Northern Plains Boulders - Robotics Institute Carnegie Mellon University

Spatial Point Process Models for the Clustering Behavior of Northern Plains Boulders

Conference Paper, Proceedings of 38th Lunar and Planetary Science Conference (LPSC '07), March, 2007

Abstract

Spatial point processes describe the arrangement of discrete objects on the plane [1,2]. Here we use them to model clustering patterns for approximately 500,000 boulders in a HiRISE orbital image. Summary statistics indicate varying degrees of aggregation at different image locations. Nevertheless, the overall clustering behavior exhibits considerable structure and local continuity. The space of observed summary statistics has a low intrinsic dimensionality, and the observed clustering is well-described by a small handful of archetypes.

BibTeX

@conference{Thompson-2007-9666,
author = {David R. Thompson and David Wettergreen},
title = {Spatial Point Process Models for the Clustering Behavior of Northern Plains Boulders},
booktitle = {Proceedings of 38th Lunar and Planetary Science Conference (LPSC '07)},
year = {2007},
month = {March},
}