Spectral 'Fingerprinting' of Phytoplankton Populations by Two-Dimensional Fluorescence and Fourier-Transform-Based Pattern Recognition.

Abstract

The selectivity of fluorescence spectroscopy is exploited for the characterization of marine algae. Two-dimensional, digital images of in vivo fluorescence intensity versus excitation and emission wavelengths, called excitation-emission matrices (EEMs), are used as spectral fingerprints for marine phytoplankton populations. Fourier-transform-based pattern recognition is described along with its inherent strengths and weaknesses for the analysis of natural populations. The EEMs of unknown algae are compared to a library of standard EEMs representing 23 algal species and 6 classes with better than 80% accuracy. The EEMs acquired under different physiological conditions are used in determining pattern recognition reliability. The potential for fingerprinting mixed populations and oceanographic regions is also discussed. Originator supplied keywords include: Fluorescence analysis; Marine analysis; Fourier data analysis; Pattern recognition; Chlorophyll fluorescence; Spectral matching.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 08, 1985
Accession Number
ADA157237

Entities

People

  • E. J. Zillioux
  • I. M. Warner
  • P. B. Oldham

Organizations

  • Emory University

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Accuracy
  • California
  • Chemical Synthesis
  • Chemistry
  • Correlation Analysis
  • Cyanobacteria
  • Data Analysis
  • Detection
  • Frequency Domain
  • Identification
  • Measurement
  • Military Research
  • Oceanography
  • Pattern Recognition
  • Phytoplankton
  • Spectroscopy
  • United States

Readers

  • Aquatic Ecology
  • Chemistry (specifically Chemical Fluorescence)
  • Image Processing and Computer Vision.

Technology Areas

  • AI & ML