Dissertation > Excellent graduate degree dissertation topics show

A Technique for Assessment of Phytoplankton Class Abundance Developed by 3-D Fluorescence Spectra and Multiwavelet

Author: ShenJun
Tutor: SuRongGuo
School: Ocean University of China
Course: Analytical Chemistry
Keywords: Phytoplankton Community composition Three-dimensional fluorescence spectroscopy Multiwavelet Recognition
CLC: Q948.8
Type: Master's thesis
Year: 2010
Downloads: 24
Quote: 0
Read: Download Dissertation

Abstract

Frequent red tides in the world and China's coastal waters, have been expanding, and the marine environment issues has become a common concern of the whole society. According to statistics, nearly 60 years of China's mainland coast of the red tide occurred 265 times, which the East China Sea, the highest frequency of occurrence. There is an urgent need to establish a rapid, real-time, effective phytoplankton community composition measurement techniques to achieve the routine and emergency monitoring of red tides. Fluorescent technology is gradually developed in recent years a phytoplankton species determination of their own characteristic spectrum information technology, fluorescence spectroscopy can provide for different types of phytoplankton. In addition, due to the high sensitivity of this method, the instrument is easy to use, and therefore easy to achieve real-time detection. Of phytoplankton community composition characteristics of China's coastal waters red tide occurred during non-bloom period summarizing research on the basis of the final selection of the 43 kinds of lab-grown offshore common dominant species of red tide phytoplankton study selected algae species belonging to 32 genera of the seven gates. Which diatoms door 9 genera and 14 species, the golden algae door 6 genera and 6 species, genera and 4 species of green algae, dinoflagellates door 6 genera and 11 species, Xanthophyta 2 2 species cryptalgal door 2 species, blue-green algae door 4 4 species. The phytoplankton in different light, different temperature, different growth cycle in vivo 3D fluorescence spectra measured in laboratory culture environment, respectively, the use of multi-wavelet analysis, non-negative least squares technique to establish our typical offshore phytoplankton red tide algae species and dominant species of the door is a a hierarchical recognition technology. Concrete achieved the following results: 1. Choice of two types of multi-wavelet function (GHM multiwavelet SA4 multiwavelet) for feature extraction of the living phytoplankton 3D scanning fluorescence spectroscopy, in turn different scale components and wavelet components, application of Bayes discriminant The principle selected best identifying characteristic components, preliminary selected two types of multi-wavelet second layer scale components Ca2 the third layer the scale component Ca3 both joint Ca2-Ca3 as a characteristic fluorescence spectral component. 2 were selected GHM multiwavelets and the SA4 multiwavelet selected phytoplankton characteristic fluorescence spectra by hierarchical cluster analysis were extracted from two multiwavelet three standard spectral library Ca2 standard library, of Ca3 standard spectral libraries, and Ca2- The Ca3 standard library, and phytoplankton fluorescence recognition technology combined with non-negative least square method. Use two types of multi-wavelet three standard spectral libraries to develop a single species of algae phytoplankton samples and analog mixed algal samples in the categories are class level grading identify the final two multiwavelet selected based on the recognition results The best standard library. GHM multiwavelets Ca2-of Ca3 standard spectrum library identification: single species of algae samples in the category level to identify the correct rate averaged 94.4% correct recognition rate of 92.7% on average in the genus level. SA4 multiwavelet Ca2 standard spectral library identification: recognition accuracy on the category level of single species of algae samples averaged 94.6% the genus level recognition accuracy of 92.7% on average. Original fluorescence spectra of single species of algae samples scanned before after a series of data processing in accordance with a certain proportion twenty-two spectral simulation were used to the two kinds multiwavelet standard library door case identification, results show that when the different types of phytoplankton fluorescence spectra, respectively, in accordance with 3:1,4:1,6:1,9:1 proportion of the mixture of two when the the GHM multiwavelet Ca2-Ca3 standard spectral library in the category level, the average correct recognition rate range of 97.6 % -97.9%, and the genus level, the average correct recognition rate range of 66.8% -90.7%; the SA4 multiwavelet Ca2 standard spectral library in the category level, the average correct recognition rate range of 98.3% -99.1%, the level of genus The average correct recognition rate range of 63.1% -91.6%. When the analog mixing ratio of not less than 4:1, category-level recognition accuracy gt; 95% the genus level recognition accuracy gt; 70%; When the analog mixing ratio of not less than 6:1, the level of genus on recognition accuracy rate of up to 80% or more. For laboratory-made mixed algal samples the GHM multiwavelet Ca2-Ca3 standard spectrum library correctly in the category level average recognition rate of 90.1%, 76.4% of the average rate of correct recognition of the genus level. Average recognition of the the the SA4 multiwavelet Ca2 Standard spectral library category level correct rate of 90.9%, 74.2% of the average rate of correct recognition of the genus level. Jiaozhou Bay to investigate water samples and identify of Maidao confining impermeable kind, two types of multi-the Wavelet standard spectrum library in the category level can reach more than 90% correctly identified to genus level to identify the correct rate of not less than 80%. The innovation of this paper is to establish a the offshore phytoplankton dominant species in China and the accident species standard fluorescence spectral library, based on wavelet technology combined with non-negative least squares chemometric methods phytoplankton community composition determination of fluorescent recognition technology. The technology to red tide phytoplankton identifying and measuring the category level identification at the genus level, and in the occurrence of red tides caused by red tide phytoplankton.

Related Dissertations

CLC: > Biological Sciences > Botany > Plant ecology and biogeography > Aquatic plant science
© 2012 www.DissertationTopic.Net Mobile