R/AFSreconstruction.R

In stla/SurfaceReconstruction: Surface Reconstruction Using the 'CGAL' C++ Library

#' @title Advancing front surface reconstruction
#' @description Reconstruction of a surface from a cloud of 3D points.
#'
#' @param points numeric matrix which stores the points, one point per row
#'
#' @return A triangle mesh, of class \code{mesh3d} (ready for plotting
#'   with \strong{rgl}).
#'
#' @details See \href{https://doc.cgal.org/latest/Advancing_front_surface_reconstruction/index.html#Chapter_Advancing_Front_Surface_Reconstruction}{Advancing Front Surface Reconstruction}.
#'
#' @export
#' @importFrom rgl tmesh3d
#' @importFrom Rvcg vcgClean
#'
#' @examples 
#' library(SurfaceReconstruction)
#' data(bunny, package = "onion")
#' mesh <- AFSreconstruction(bunny)
#' library(rgl)
#' shade3d(mesh, color = "firebrick")
AFSreconstruction <- function(
  points
){
  if(!is.matrix(points) || !is.numeric(points)){
    stop("The `points` argument must be a numeric matrix.", call. = TRUE)
  }
  dimension <- ncol(points)
  if(dimension != 3L){
    stop("The dimension must be 3.", call. = TRUE)
  }
  if(nrow(points) <= dimension){
    stop("Insufficient number of points.", call. = TRUE)
  }
  if(any(is.na(points))){
    stop("Points with missing values are not allowed.", call. = TRUE)
  }
  storage.mode(points) <- "double"
  afs <- AFSreconstruction_cpp(t(points))
  rglmesh <- tmesh3d(
    afs[["vertices"]], afs[["triangles"]], normals = NULL, homogeneous = FALSE
  )#afs[["normals"]])
  vcgClean(rglmesh, sel = c(0, 7), silent = TRUE)
}