Dear all,
I have a set of matrix and I need to compute several measures on source image
and to generate other matrices, in a moving window way. My input images have
integer and double precision values, and also NAs.
I need to compute some measures that I will define as functions (like the
sample below). As a tentative way of to do so, I made two for() looping, but I
feel that the spear is so slow. For the sample below the time is not so big,
but in fact I will neet about 1.000 matrices, and for several windows sizes
(like 3=3x3.... up to 17x17).
Any idea?
window.size<-7 # 7x7
window.offset<-(window.size-1)/2
vin<-matrix(runif(10000)*20,nc=100)
vin[sample(1:length(vin),1000)]<-NA
x11(1024,768)
image(vin)
vin.range<-vin; vin.range[]<-NA
vin.range<-vin; vin.range[]<-NA
vin.sde<-vin; vin.sde[]<-NA
vin.min<-vin; vin.min[]<-NA
for (ll in (window.offset+1):(dim(vin)[1]-window.offset))
{
for (cc in (window.offset+1):(dim(vin)[2]-window.offset))
{
windows.tmp<-vin[(ll-window.offset):(ll+window.offset),(cc-window.offset):(cc+window.offset)]
if (length(windows.tmp[!is.na(windows.tmp)])>0) {
vin.range[ll,cc]<-range(windows.tmp[!is.na(windows.tmp)])[2]-range(windows.tmp[!is.na(windows.tmp)])[1]
vin.sde[ll,cc]<-sd(windows.tmp[!is.na(windows.tmp)])
vin.min[ll,cc]<-min(windows.tmp[!is.na(windows.tmp)])
}
} #ll
} #cc
par(mfrow=c(2,2))
image(vin,main=paste("VIN - ws
=",window.size,"\n",var.desc),col=rainbow(20),panel.first = grid(8,8))
image(vin.range,main="VIN.RANGE",col=rainbow(20),panel.first = grid(8,8))
image(vin.sde,main="VIN.SD",col=rainbow(20),panel.first = grid(8,8))
image(vin.min,main="VIN.MIN",col=rainbow(20),panel.first = grid(8,8))
Thanks,
Miltinho
Brazil.
para armazenamento!
[[alternative HTML version deleted]]
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
