CODEWITHSUNDEEP
r
Paolo Pol Cristoforetti
Paolo Pol Cristoforetti

Reputation: 173

Combine multiple images together

I'm trying to merge multiples images (500 pics) together in a for loop. The image size is constant and even not too big (225*410 px). What I need is to obtain a single image composed by the 500 initial pics stuck together side by side.

I've tried with a for loop using some functions of package EBImage. abind() it's like the traditional rbind(). The code I've used is the following:

library(abind)
#path=a list containing the paths of the source images
final_image<-readImage(path[1]) #initialize the final image
for (i in 2:500){
  im <- readImage(path[i]) #open the i-esim image
  final_image <- abind(final_image,im,along=1) #paste the i-esim image with the previous one
}

The code works but, obviously, it's really slow because at each iteration the size of final_image gets bigger.

Does anyone know a faster workaround? Thanks!

Upvotes: 3

Views: 1725

Answers (1)

r2evans
r2evans

Reputation: 160447

In general, iteratively rbinding (goes for other *bind funcs, too) is a really bad idea, as it makes a complete copy with each iteration in the loop (as you noticed). Notice that in ?abind, it takes ...:

...  Any number of vectors, matrices, arrays, or data frames. The
     dimensions of all the arrays must match, except on one dimension
     (specified by along=). If these arguments are named, the name will be
     used for the name of the dimension along which the arrays are joined.
     Vectors are treated as having a dim attribute of length one.

which allows us to use do.call to do the binding all at once on a single list of all images. Try this (untested):

list_of_images <- lapply(path, readImage)
combined <- do.call(abind, c(list_of_images, list(along = 1)))

Upvotes: 3

Related Questions