R/opencv.R

In opencv: Bindings to 'OpenCV' Computer Vision Library

#' OpenCV Computer Vision
#'
#' Tools to experiment with computer vision algorithms. Use [ocv_read] and
#' [ocv_write] to load/save images on disk, or use [ocv_picture] / [ocv_video]
#' to use your webcam. In RSudio IDE the image objects will automatically be
#' displayed in the viewer pane.
#'
#' @export
#' @rdname opencv
#' @importFrom Rcpp sourceCpp
#' @useDynLib opencv
#' @param path image file such as png or jpeg
#' @examples # Silly example
#' mona <- ocv_read('https://jeroen.github.io/images/monalisa.jpg')
#'
#' # Edge detection
#' ocv_edges(mona)
#' ocv_markers(mona)
#'
#' # Find face
#' faces <- ocv_face(mona)
#'
#' # To show locations of faces
#' facemask <- ocv_facemask(mona)
#' attr(facemask, 'faces')
#'
#' # This is not strictly needed
#' ocv_destroy(mona)
ocv_read <- function(path){
  if(grepl("https?://", path)){
    base <- basename(path)
    tmp <- file.path(tempdir(), base)
    utils::download.file(path, tmp, mode = 'wb')
    path <- tmp
  }
  path <- normalizePath(path, mustWork = TRUE)
  cvmat_read(path)
}

#' @export
#' @rdname opencv
#' @param image an ocv image object created from e.g. [ocv_read()]
ocv_write <- function(image, path){
  path <- normalizePath(path, mustWork = FALSE)
  cvmat_write(image, path)
}

#' @export
#' @rdname opencv
ocv_destroy <- function(image){
  cvmat_destroy(image)
}

#' @export
#' @rdname opencv
ocv_bitmap <- function(image){
  cvmat_bitmap(image)
}

#' @export
#' @rdname opencv
ocv_edges <- function(image){
  cvmat_edges(image)
}

#' @export
#' @rdname opencv
ocv_picture <- function(){
  cvmat_camera()
}

#' @export
`print.opencv-image` <- function(x, ...){
  viewer <- getOption("viewer")
  is_knit_image <- isTRUE(getOption('knitr.in.progress'))
  if(!is_knit_image && is.function(viewer) && !cvmat_dead(x)){
    tmp <- file.path(tempdir(), paste0("preview.png"))
    ocv_write(x, path = tmp)
    viewer(tmp)
  }
  base::print.default(x)
  #TODO: copy from magick
}

#' @export
#' @rdname opencv
#' @param width output width in pixels
#' @param height output height in pixels
ocv_resize <- function(image, width = 0, height = 0){
  cvmat_resize(image, as.integer(width), as.integer(height))
}

#' @export
#' @rdname opencv
ocv_mog2 <- function(image){
  cvmat_mog2(image)
}

#' @export
#' @rdname opencv
ocv_knn <- function(image){
  cvmat_knn(image)
}

#' @export
#' @rdname opencv
ocv_hog <- function(image){
  cvmat_hog(image)
}

#' @export
#' @rdname opencv
#' @param ksize size of blurring matrix
ocv_blur <- function(image, ksize = 5){
  cvmat_blur(image, ksize)
}

#' @export
#' @rdname opencv
#' @param color true or false
ocv_sketch <- function(image, color = TRUE){
  cvmat_sketch(image, color)
}

#' @export
#' @rdname opencv
ocv_stylize <- function(image){
  cvmat_stylize(image)
}

#' @export
#' @rdname opencv
ocv_markers <- function(image){
  cvmat_markers(image)
}

#' @export
#' @rdname opencv
ocv_info <- function(image){
  info <- cvmat_info(image)
  info$type  <- enum_label("type", info$type)
  info$depth <- enum_label("depth", info$depth)
  info
}

#' @export
#' @rdname opencv
#' @param target the output image
#' @param mask only copy pixels from the mask
ocv_copyto <- function(image, target, mask){
  cvmat_copyto(image, target, mask)
}

#' @export
#' @rdname opencv
ocv_display <- function(image){
  cvmat_display(image)
}

#' @export
#' @rdname opencv
#' @param filter an R function that takes and returns an opecv image
#' @param stop_on_result stop if an object is detected
ocv_video <- function(filter, stop_on_result = FALSE){
  if(!is.function(filter))
    stop("Filter must be a function")
  message("Starting video window (could be behind this window)")
  livestream(filter, stop_on_result)
}

#' @export
#' @rdname opencv
ocv_grayscale <- function(image){
  cvmat_grayscale(image)
}

#' @export
#' @rdname opencv
ocv_version <- function(){
  cvversion()
}

#' @importFrom magrittr %>%
#' @export
magrittr::`%>%`