9

I have code below which saves an image to my pc. I would like to rotate that image by 45,90 and 135 degrees around its center (or bottom left hand corner) and then save as 3 different images. How could I do that?

library(raster)
r1 <- brick(system.file("external/rlogo.grd", package="raster"))
x <- crop(r1, extent(0,50,0,50))
plotRGB(x)
png(width=50, height=50)
par(mai=c(0,0,0,0))
image(x)
dev.off()

---------update1-------------------------

Based upon the accepted answer the working code is as below

library(raster)
r1 <- brick(system.file("external/rlogo.grd", package="raster"))
r1
x <- crop(r1, extent(0,ncol(r1),0,nrow(r1)))
plotRGB(x)

x1 <- 0:ncol(x)
y1 <- 0:nrow(x)
z <- matrix(1, nrow=length(x1), ncol=length(y1))

col.mat <- t(apply(matrix(rgb(getValues(x)/255), nrow=nrow(x), byrow=TRUE), 2, rev))

# Rotate 45 degrees
persp(x1, y1, z, zlim=c(0,2), theta = 20, phi = 90, 
      col = col.mat, scale=FALSE, border=NA, box=FALSE)
png("SaveThisPlot.png")
persp(x1, y1, z, zlim=c(0,2), theta = 20, phi = 90, 
      col = col.mat, scale=FALSE, border=NA, box=FALSE)
dev.off()
user2543622
  • 5,760
  • 25
  • 91
  • 159

2 Answers2

14

For the 90 degree rotations, this is an easy solution:

image(t(flip(x, 1)))
image(t(flip(x, 2)))
plotRGB(t(flip(x, 1)))
plotRGB(t(flip(x, 2)))

For the 45 degree and 135 degree rotation, it will be a bit trickier. There are probably other ways, but I'll use the persp function and give different angles to the theta argument.

It's just a matter of setting up the call to persp properly. x1, y1, and z are just inputs for the persp function (see ?persp for more about the arguments to that function). col.mat is a matrix holding color values. 

x1 <- 0:ncol(x)
y1 <- 0:nrow(x)
z <- matrix(1, nrow=length(x1), ncol=length(y1))
col.mat <- t(apply(matrix(rgb(getValues(x)/255), nrow=nrow(x), byrow=TRUE), 2, rev))
# the transposing and reversing are just to get the colors in the same 
# spots as they're in when viewing plotRGB(x).
#
# getValues(x) is how you get the rgb colors, in the form of a 3-column matrix.
# rgb(getValues(x)/255) converts them into hex code, which is convenient enough here.

If you find that this is the mirror image of what you're looking for, try filling up the color matrix differently. For example:

col.mat <- matrix(rgb(getValues(x)/255), nrow=nrow(x))

As you can tell, filling up the color matrix properly is the key to making this approach work for you. I'll leave it as an exercise to the reader to figure out how to do any other manipulations to the color matrix.

Now, call persp. Here, I set the zlim values so there is a range including 1. Because I made all the z values 1, you need to set a valid range, otherwise, persp will throw an error about invalid limits. It doesn't like a range from 1 to 1.

# Rotate 45 degrees
persp(x1, y1, z, zlim=c(0,2), theta = 45, phi = 90, 
    col = col.mat, scale=FALSE, border=NA, box=FALSE)

enter image description here

Here's a 135 degrees: 

# Rotate 135 degrees
persp(x1, y1, z, zlim=c(0,2), theta = 135, phi = 90, 
    col = col.mat, scale=FALSE, border=NA, box=FALSE)

enter image description here

Saving the plots can be done in the same way you show in your question:

png("SaveThisPlot.png")
persp(x1, y1, z, zlim=c(0,2), theta = 135, phi = 90, 
    col = col.mat, scale=FALSE, border=NA, box=FALSE)
dev.off()
Jota
  • 17,281
  • 7
  • 63
  • 93
  • The original poster later had some issues when width != height, as notedin a later post: http://stackoverflow.com/questions/31814873/r-raster-rotating-images-and-not-changing-their-size. – aaryno Aug 05 '15 at 16:05
1

The way to do it without throwing an error or warning message (after reading the help page and playing with the example:)

extent(x) <- extent(0, 360, -90, 90)
rr <- rotate(x)
 png()
 plotRGB(x)
dev.off()

But you may not like it because it assumed a particular coordinate system tat you probably did not. If you want to pull the data elements out then use the @ operator:

str(x)  The slots are named 'data' and 'values'
image(x)
image(matrix(x@data@values[,1], 50, byrow=TRUE)) # rotate one layer
IRTFM
  • 258,963
  • 21
  • 364
  • 487
  • I dont think it is rotating my image. Rather it seems that it streching my image along x axis :( – user2543622 Jun 17 '15 at 21:52
  • Actually it's rotating it and deforming but after splitting in half (because of the assumption of a -180 to +180 coordinate system). I warned you wouldn't like it. – IRTFM Jun 17 '15 at 22:10
  • `rotate` is for a single special case. To 'rotate' global lon/lat data from 0..360 to -180..180 degrees longitude.. This (the direct use of S4 slot that you do not know what they are for) `image(matrix(x@data@values[,1], 50, byrow=TRUE))` is a terrible suggestion. – Robert Hijmans Jun 18 '15 at 02:59
  • Yes, `rotate` is for a specific case. That _was_ my point. Agreed the "direct use" of the 'values' matrix may be a "terrible suggestion" for a user who never wants to get under the hood, but I counter that you often do want to understand the nature of the data objects you are working with. – IRTFM Jun 18 '15 at 04:18
  • I agree with the 'wanting to understand', but I think that in this case you are leading people the wrong way. The 'raster' package has methods to get the values (getValues, as.matrix, values() to name a few). If an object is derived from a file, there won't be any values in the slot you referred to. – Robert Hijmans Jun 18 '15 at 05:14