Skip to contents

Prepare training points for remote sensing algorithm

Source: R/preparing.R
get_lu_points.Rd

This function prepares training points for a remote sensing algorithm based on land use and crop classes.

Usage

get_lu_points(
  df,
  year,
  lookup,
  lu_constant = c(),
  nb_pts = 100,
  col_name = "type"
)

Arguments

df

An sf data.frame with land use. The "type" column should be present.

year

A numeric value representing the year of land use.

lookup

A dataframe with a "lc1" column for numeric codes and a "type" column for text.

lu_constant

(optional) A vector of strings with land uses to be kept constant in land use (e.g., water, urban areas, etc.). Default lu_constant = c().

nb_pts

(optional) A numeric value representing the number of points per land use/crop class. Default nb_pts = 100.

col_name

A string with the name of the column to be used representing the type of crops/land use. Default col_name = "type".

Value

An sf data.frame with point input for a remote sensing training algorithm.

References

Mészáros, J., & Szabó, B. (2022). Script to derive and apply crop classification based on Sentinel 1 satellite radar images in Google Earth Engine platform. https://doi.org/10.5281/zenodo.6700122

Examples

if (FALSE) {
  library(sf)
  
  # Loading land use/crop layer
  lu_path <- system.file("extdata", "GIS/lu_layer.shp", package = "SWATprepR")
  lu <- st_read(lu_path, quiet = TRUE)
  
  # Preparing lookup table
  lookup <- data.frame(lc1 = seq(1:length(unique(c(lu$type)))), 
                       type = unique(c(lu$type)))
                       
  # Setting land uses to be kept constant
  lu_constant <- c("fesc", "orch", "frst", "frse", "frsd", "urld", "urhd", 
                   "wetl", "past", "watr", "agrl")
  
  # Getting training points
  pts <- get_lu_points(lu, 2021, lookup, lu_constant)
}