R/split_transects_from_pcl.R

In forestr: Ecosystem and Canopy Structural Complexity Metrics from LiDAR

#' Split transects from PCL
#'
#' \code{split_transects_from_pcl} places data values into x-bins (x-coordinates
#' and) and z-bins (z-coordinates)
#'
#' Function to add two additional columns
#' to the pcl dataset, one for the segment
#'  (which should only be from 1-4) and is
#' designated by a -99999999 value in the
#' return_distance column
#' The only required parameters are the data
#' frame of pcl data, with the length of transect and the marker
#' spacing.
#'
#' @param pcl_data data frame of unprocessed PCL data.
#' @param transect.length total transect length. Default value is 40 meters.
#' @param marker.spacing distance between markers in meters within the PCL data. Default value is 10 m.
#' @param data_dir directory where PCL data .csv are stored if value is used.
#' @param output_file_name old code relic that doesn't do much.
#' @param DEBUG check to see order of final output. Default is FALSE.
#'
#'
#' @export
#' @examples
#' # Function that has the algorithm that splits the raw data into defined, equidistant x-bins.
#'  pcl_split <- split_transects_from_pcl(pcl_adjusted,
#'  transect.length = 40, marker.spacing = 10)
#'
split_transects_from_pcl <- function(pcl_data, transect.length, marker.spacing, DEBUG = FALSE,  data_dir, output_file_name) {

  index <- NULL
  # Initialize count for segments (expecting 4 segments per transect)
  # Some returns before beginning of first segment and some after last
  segment_num <- 0

  # Check for how many segment boundaries we have (should be 5)
  #stopifnot(length(pcl_data[pcl_data$return_distance == -99999999, 2]) = 5)
  # stopifnot(length(which(pcl_data$return_distance < -9999)) == 4)


  # Walk through rows and add the segment number in a new column
  for (i in 1:nrow(pcl_data)) {
    pcl_data$seg_num[i] <- segment_num

    if(pcl_data$return_distance[i] <= -9999 & !is.na(pcl_data$return_distance[i])){
      segment_num <- segment_num + 1

    }
    if (segment_num == ((transect.length/marker.spacing) + 1)) {
      break
    }
  }

  # Check to see if it worked
  if (DEBUG) utils::head(pcl_data)

  # Initialize empty data frame to store results
  results <- data.frame()

  # but we're flexible here. Uses cut() with labels = FALSE to return

  #### this needs to be adjusted to account for smaller transects
  # a vector of integer categories for each "chunk" within each segment
  # This should go from 1-10 and be spaced evenly in "index" space
  for (i in 1:(max(pcl_data$seg_num) - 1)) {
    for (i in 1:(max(pcl_data$seg_num))) {
      this_segment <- subset(pcl_data, pcl_data$seg_num == i)
      this_segment$chunk_num <- cut(this_segment$index, marker.spacing, labels = FALSE)
      results <- rbind(results, this_segment)
    }

    # Make sure we didn't make too many chunks in any segment
    #stopifnot(max(results$chunk_num) < 11)
    stopifnot(max(results$seg_num) < ((transect.length/marker.spacing) + 1))
  }
  # Code segment to create zbin and xbin
  results$xbin <- ((results$seg_num * marker.spacing) - marker.spacing)  + results$chunk_num
  results$zbin <- round(results$return_distance)
  results$zbin[results$sky_hit == "TRUE"] <- 0
  # Check final output
  if (DEBUG) utils::head(results)
  if (DEBUG) utils::tail(results)


  results <- dplyr::distinct(results, index, .keep_all = TRUE)
  return(results)
}