| 摘要 |
PROBLEM TO BE SOLVED: To provide a method for removing noise components capable of removing noise components, while minimizing missing signal components of multipoint spectral data. SOLUTION: The noise component removal method for removing noise components from multipoint spectral data, acquired by measuring (S10) a section to be measured of a sample surface is provided with a concentration variable computation process (S12) for converting characteristic information, indicating characteristics of the section to be measured, except the spectral information as concentration variables of a partial least-squares method into numerical values; a PLS analysis process (S14) for performing multivariate analysis on the multipoint spectral data by the partial least-squares method through the use of spectral information as independent variables of the partial least square method to determine the constituent components of the multipoint spectral data of the section to be measured, in a descending order of characteristic value; and a spectrum reconstruction method (S16) for removing constituent components, having characteristic values lower than a prescribed value from among the constituent components and to reconstruct the multipoint spectral data. COPYRIGHT: (C)2008,JPO&INPIT
|
