I want to find position of a number of minerals in an area using aster data. So I should to find pixels that their z-profile have highest adapt with z-profile of mineral of interest in spectral library (for example Kaolinite).
I tried to compare spectral profiles visually and for this purpose i applied some transforms such as MNF and PCA to distinguish changes in image but i could reach my goal.
I also applied SFF(spectral feature fitting) method and I did not get good result.
main reason that makes the comparability between z-profiles impossible is low spectral resolution of aster data than spectral library profiles.
My question is: can who has a method that I can compare? If yes please tell me step by step and with adequate description.
Thanks in advance
Statically you can prepare the two spectral profile in spread sheet (Wavelength vs % ref) and then apply one way Anova analysis from (Data menu.... Data Analysis...... Anova single factor). You can also use SAS or SPSS which give much more figures and data analysis; and if you want specific narrow wavebands you can use one of the multivariate techniques
On the raster map: try to use the Spectral Angle Mapper after applying MNF technique.
https://www.utd.edu/~herve/Abdi-MultivariateAnalysis-pretty.pdf
- Badges
- Science topic