I'm working on classifications of very high spatial resolution Quickbird (agricultural plot) and I would like to use SVM classification with the software ENVI. I encounter some difficulties and I would like to know if somebody could help me. In the literature (C-W. HSU, C-C. CHANG and C-J. LIN, A practical Guide to Support Vector Classification, 2008), the RBF kernel is first considered. A simple scaling on the data must have been previously conducted. How can we do this scaling in concrete terms ? Then a cross validation to find the best parameter C and Gamma has to be done. C is the penalty parameter which controls the trade-off between allowing training errors and forcing rigid margins. Gamma parameter is used in the kernel function. How can we concretely do this cross-validation ? Two additional parameters have to be determined : Pyramid Levels field to set the number of hierarchical processing levels to apply during the SVM training and classification process and Pyramid Reclassification Threshold to specify the probability threshold that a pixel classified at a lower resolution level must meet to avoid being reclassified at a finer resolution. How can we choose these values ? I have tested several different values but it makes the software crash. Thanking you in advance Audrey