Identification of the seaweed fluorescence spectroscopy based on the KPCA and ICA-SVM

Abstract

The problem of water pollution is very serious. The seaweed is an important feature of eutrophication. It is an important aspect of pollution. Three-dimensional fluorescence spectrum can show entire fingerprint information of fluorescent light that in the range of excitation and emission wavelength, but the dimension of three-dimensional fluorescence spectrum is higher, the characteristic spectrum of different kinds pelagic plant are multifarious, it is complex identification. The kernel principal component analysis (KPCA) is used in this paper. It can reduce the dimensions of the spectroscopy. The independent component analysis (ICA) is used to do the matrix decomposition from the perspective of independence to extract the main feature of the spectroscopy data processed by the KPCA. The support vector machine (SVM) is used to assort the main characteristic root books which are abstracted by the ICA. The correct laboratory sorting of seaweed is realized. Experimental result indicate, this method can identify the chief component of admixture seaweed, the high dimensional spectroscopy information of seaweed is proceed effective feature extraction, the sorting speed is increase greatly, the discrimination of sorting is reach 90% percent.