Data Processing And Recognition Algorithm Optimization Based On Three-dimensional Fluorescence Spectrum

For decades, how to use the scientific method to reduce losses from fishtoxic red tides, has been an unsolved problem in academic circles. With thedevelopment of the algae’s three dimensional fluorescence spectroscopy,pigment spectrometry plays an important role in algae recognition work. Forsome red tide algae toxicity and strong, middle or weak recognition, we cannot only provide an instructive basis for decomposing the generatingmechanism of the toxin, but also provide an important theoretical support forred tides work of disaster prevention and reduction.This article mainly focuses on the three dimensional fluorescencespectrum experiment data processing and algorithm optimization, and we usesome methods to deal with algae fluorescence spectral data, for examplematrix maximum normalized, Rayleigh scattering and Coif2wavelet function,etc. At the same time, by using the method of Fisher discriminant function, wecan structure some discriminants of algal toxicity and strong, middle or weak.Finally, On the basis of the existing method, by combining the wholearrangement of different environmental elements and Rayleigh scatteringreuse, we can optimize the original algorithm, and then the recognition ratewill be improved.At last, We set up a vivo fluorescence analysis prediction technology of the red tidealgae’s non-toxic and toxicity of strength at offshore waters. At the same time we can alsoprovide some effective theoretical suggestions for the development of a rapid and efficientfish toxic algae fluorescence spectrum detector.