Dissertation > Excellent graduate degree dissertation topics show

The Application of Multivariate Analysis and Excitation-Emission Matrix Fluorescence Spectroscopy of Dissolved Organic Matter in the Identification of Red Tide Algae

Author: LvGuiCai
Tutor: WangJiangTao
School: Ocean University of China
Course: Marine Chemistry
Keywords: Red tide algae Three-dimensional fluorescence spectroscopy Parallel Factor Analysis Principal Component Analysis Wavelet Analysis
CLC: X55
Type: Master's thesis
Year: 2010
Downloads: 122
Quote: 0
Read: Download Dissertation

Abstract

Harmful Algal Blooms wreak havoc on our coastal marine ecosystem, resources, the environment, causing serious damage and significant economic losses. To prevent and reduce the harmful red tide disaster, we must first be able to accurately identifying and measuring in real time of the accident occurrence of red tide algae species fluorescence technique has the advantages of simple instrumentation, high sensitivity, and easy to implement on-site real-time detection and to receive widespread attention. Dimensional fluorescence spectra of phytoplankton phytoplankton species identification studies have achieved certain results, and use of algae in the growth of phytoplankton in the process of three-dimensional fluorescence spectra of the filtrate red tide algae species identification has not been reported in this article from 3D fluorescence spectra of the red tide algae filtrate to proceed, our coastal selected 10 species of microalgae cultured under laboratory conditions, respectively, parallel factor analysis, principal component analysis and wavelet analysis of three-dimensional process of red tide algae growth algal filtrate The fluorescence spectra for feature extraction, identification and measurement of the red tide algae using multiple linear regression. (1) the application of parallel factor analysis of three-dimensional fluorescence spectra analysis of red tide algae filtrate, the spectral characteristics of the red tide algae, there are differences between the different red tide algae characteristic spectral clustering results show the same red tide algae characteristic spectral similarity higher . Each component of red tide algae maximum fluorescence peak intensity and growth stages shows a certain relationship: in the exponential growth phase, the algal filtrate class protein and humic fluorescent substances continue to accumulate, fast the stable phase filtrate class tryptophan and humic fluorescent substance increase, mainly from the aging and death of cells broken release, as well as the role of bacterial degradation. This is consistent with previous findings and the laboratory, parallel factor analysis can effectively extract the fluorescence characteristics of red tide algae, and to examine the relationship between the components of the red tide algae fluorescence peak intensity and growth stage. (2) The principal component analysis in the identification of red tide algae. The principal component analysis feature extraction, three-dimensional fluorescence spectral data of red tide algae filtrate as the identifying characteristic spectrum based on the results of the Bayes discriminant select the first principal component load spectrum, the use of cluster analysis to identify the standard spectrum, using multiple linear regression ( to resolve non-negative least square method) to identify the red tide algae. The Prorocentrum genus of P. donghaiense and marine Prorocentrum correct recognition rate is low, the other three kinds of dinoflagellates recognition rate ≧ 95%, correct identification rate for diatoms Skeletonema costatum and weak Chaetoceros lower three kinds of diatoms recognition rate ≧ 83%, correct identification of the genus Chaetoceros genus level was 96%. (3) The wavelet analysis in the identification of red tide algae. Using three the wavelet function coif2 db-db-7 on algae dimensional fluorescence spectra of the filtrate the third layer of scale components were selected, based on the results of the Bayes discriminant feature extraction as an identifying characteristic spectrum. The cluster analysis method to identify the standard spectrum of each wavelet function. The three standard spectrum multiple linear regression of the red tide algae identification, correct identification rate of the three kinds of standard spectral P. donghaiense Skeletonema costatum and weak Chaetoceros lower other algae species identification rate ≧ 80%. Genus level 3 standard spectrum of the genus Chaetoceros three red tide algae identification rate of 96%. (4) 4 standard spectrum is very low recognition rate of Skeletonema costatum and weak Chaetoceros other relatively high, db-3 standard spectrum identification results. In order to further improve the accuracy of recognition rate and even the identification of field data, but also further samples of red tide algae cultured under the conditions of different light combinations of different temperatures, feature extraction, and to study the spectral characteristics of the mixed algae extract and split.

Related Dissertations

CLC: > Environmental science, safety science > Environmental pollution and its prevention > Marine pollution and its prevention
© 2012 www.DissertationTopic.Net Mobile