Comparison of some methods of analyzing the three-dimensional spectral image using simulation

Abstract

In this article, a new method has been proposed in dealing with the spectral image and analyzing it under the principle components method of Three-Dimensional linearDiscriminant-Analysis (3DPCLDA) and the principle components method – QuadraticDiscriminant-Analysis (3DPCQDA) and compared it with the traditional methods, which is the principle components method - linear discriminant analysis. PCLDA) Ordinary and principal components method - Quadratic Discriminant Analysis (PCQDA), Partial Least Squares Method - PLSDA and SVM Method. Using simulation experiments, it was concluded that the three-dimensional methods are better than the traditional methods, as they have achieved the most standards of accuracy and testing