Journal Article

Models for point processes observed with noise

J Lund and M Rudemo

in Biometrika

Published on behalf of Biometrika Trust

Volume 87, issue 2, pages 235-249
Published in print June 2000 | ISSN: 0006-3444
Published online June 2000 | e-ISSN: 1464-3510 | DOI: https://dx.doi.org/10.1093/biomet/87.2.235
Models for point processes observed with noise

Show Summary Details

Preview

Consider a pair of point processes, X and Y, where X is regarded as a 'true' point process and Y is an imperfect observation of X. For the transformation from X to Y, we consider a number of disturbance mechanisms covering random thinning, displacement, censoring of the displaced points and superposition of extra points. We present the conditional likelihood of Y given X. When both point processes are observed the likelihood may be used for inference about the disturbance mechanisms. The likelihood is a sum, typically with very many terms, and we discuss an approximation with a small number of terms. The results are applied to an example, where X denotes a set of 'true' positions of tree tops, and Y denotes tree-top positions estimated by template matching in a digital image obtained by high-resolution aerial photography. The parameters governing the various disturbance mechanisms are estimated from the conditional likelihood.

Keywords: censoring of point processes; conditional likelihood; image data; incomplete observation; matching of point sets; random displacement; superposition; thinning

Journal Article.  0 words. 

Subjects: Probability and Statistics

Full text: subscription required

How to subscribe Recommend to my Librarian

Users without a subscription are not able to see the full content. Please, subscribe or login to access all content. subscribe or login to access all content.