I am confused, as per where to start and what R packages should be used? Also readImage() in PNG package is generating a 3d array i dont know what this third dimension is for.
I am able to get a raster array out of png image, but still looking for connection with R
you can use EBImage library in R.It has lot of functions like resize color mode choosing etc.3rd dimension usually represents the color channel.For identification of images u can use convolutional neural network .
You can follow the below link: