How to do supervised image classification in saga with the help of R script or Python script

Question

I am using Ubuntu 12.04 and want to do image classification in SAGA. SAGA has numerous classification algorithms which is not present in GRASS and QGIS. I am new to SAGA, but more accustomed to work on R or python. I installed RSAGA package in R, but could not find any command to do image classification. Please let me know how to do image classification in SAGA through an R/python/shell scripts.

Thanks, Karthik K.

Answer 1

It's looks like you can´t use those SAGA modules with RSAGA, but if you are trying to do cluster analysis (unsupervised classification) you can do it directly in R:

library(raster)
files <- list.files()
EVI <- stack(files)
plot (EVI)

EVI.df <- as.data.frame(EVI)  
wss <- (nrow(EVI.df)-1)*sum(apply(EVI.df,2,var))
for (i in 2:12) {wss[i] <- sum(kmeans(EVI.df, centers=i)$withinss)} ## to choose numer of clusters...
plot(1:12, wss, type="b", xlab="Number of Clusters", ylab="Within groups sum of squares")

kmeans.EVI <- kmeans(EVI.df, 12) ### kmeans, with 12 clusters

clusters <- setValues(EVI, kmeans.EVI$cluster)  
plot(clusters)