R/image_processing.R

In swinersha/UAVforestR: Tools for Assessing Forest Quality with UAVs

# A couple of wrappers for running the sobel edge detection and blurring as part
# of the pre-processing for ITC segment


#' Perform Sobel convolution
#'
#' @param imagery A raster image
#' @return A sobel convolved raster
#' @export
#' @author Tom Swinfield
#' @details
#'
#' Created 17-02-16
#'

sobel_edge<-function(imagery){
  # Calculating edge strength using a Sobel edge detector:
  kernX<-matrix(c(-1,-2,-1,0,0,0,1,2,1), 3, 3)
  kernY<-matrix(c(-1,-2,-1,0,0,0,1,2,1), 3, 3, byrow=TRUE)
  im_sobel <- raster::focal(imagery, w = matrix(1, 3, 3),
                            fun = function(x) {
                              y<-sqrt(sum(x* kernX)^2 + sum(x* kernY)^2)
                              if(is.na(y))
                                y<-0
                              return(y)
                            })
  return(im_sobel)
}


#' Perform Gaussian blur
#'
#' @param imagery A raster image
#' @return A blurred raster
#' @export
#' @author Tom Swinfield
#' @details
#'
#' Created 17-02-16
#'

blur<-function(imagery){
  imagery <- raster::focal(imagery, w = matrix(1, 3, 3),
                           fun = function(x) {
                             mean(x, na.rm = T)
                           })
  return(imagery)
}