Principal polynomial analysis for remote sensing data processing

V. Laparra*, D. Tuia, S. Jiménez, G. Camps-Valls, J. Malo

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference paperAcademicpeer-review

7 Citations (Scopus)

Abstract

PCA is widely used in this context but its linear features are optimal in error reconstruction terms only in the case that data displays a particular symmetry which is not always guaranteed in remote sensing. The proposed PPA is a non-linear generalization of PCA that relaxes the required symmetry of the data. We analytically proved that PPA always improves PCA performance in reconstruction error after dimensionality reduction as well as in energy compaction. Results on high resolution multispectral images show the suitability of using PPA for remote sensing data analysis.

Original languageEnglish
Title of host publication2011 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2011 - Proceedings
Pages4180-4183
Number of pages4
DOIs
Publication statusPublished - 16 Nov 2011
Externally publishedYes
Event2011 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2011 - Vancouver, BC, Canada
Duration: 24 Jul 201129 Jul 2011

Publication series

NameInternational Geoscience and Remote Sensing Symposium (IGARSS)

Conference/symposium

Conference/symposium2011 IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2011
Country/TerritoryCanada
CityVancouver, BC
Period24/07/1129/07/11

Fingerprint

Dive into the research topics of 'Principal polynomial analysis for remote sensing data processing'. Together they form a unique fingerprint.

Cite this