pa_yield: Create an interpolated yield object from raw data

In pacu: Precision Agriculture Computational Utilities

Create an interpolated yield object from raw data

Description

Create an interpolated yield object from raw data

Usage

pa_yield(
  input,
  data.columns = NULL,
  data.units = NULL,
  grid = NULL,
  algorithm = c("none", "simple", "ritas"),
  formula = NULL,
  overlap.threshold = 0.5,
  var.label = "yield",
  boundary = NULL,
  clean = FALSE,
  clean.sd = 3,
  clean.edge.distance = 0,
  smooth.method = c("none", "krige", "idw"),
  fun = c("none", "log"),
  lbs.per.bushel = NULL,
  moisture.adj = NULL,
  lag.adj = 0,
  unit.system = c("none", "metric", "standard"),
  remove.crossed.polygons = FALSE,
  steps = FALSE,
  cores = 1L,
  verbose = TRUE,
  ...
)

Arguments

input	an sf object containing the raw yield monitor data
data.columns	When algorithm is ‘simple’, this argument should be a vector of length 2 or 3 (depends on whether the user wants to adjust for time lag) indicating which column contains the yield data , a column containing moisture information, and a column indicating the time between readings. When algorithm is ‘ritas’, an optional named vector with the column names for the variables ‘mass, flow, moisture, interval, angle, swath, distance’. If a an unnamed vector is supplied, the vector is assumed to be in this order. The default is NULL, in which case the function attempts to guess the columns by using a dictionary of possible guesses.
data.units	When algorithm is ‘simple’, should be a vector of length two, indicating the units of the yield column and the moisture column. Common values would be ‘c('bu/ac', '%')’. When algorithm is ‘ritas’, an optional named vector with strings representing units for the variables ‘mass, flow, moisture, interval, angle, swath, distance’. If a an unnamed vector is supplied, the vector is assumed to be in this order. A typical value for this argument would be ‘c(flow = 'lb/s', moisture = '%', interval = 's', angle = 'degreeN', width = 'ft', distance = 'ft')’. Please see valid_udunits for help with specifying units. The default is NULL, in which case the function attempts to guess the units according to the values of the variable.
grid	an sf or pa_trial object containing the prediction grid. If the user is processing yield data coming from a research trial (i.e. follows a trial design), the user can pass the sf object containing the trial design information to this argument. If the argument ‘formula’ contains any predictions, the predictor should be included in the sf object supplied to this argument. polygons for which the predictions generated.
algorithm	algorithm used to generate the yield object.
formula	formula defining the relationship between the dependent and independent variables. If the dependent variable is a linear function of the coordinates, the formula can be ‘z ~ X + Y’. If the dependent variable is modeled only as a function of the mean spatial process, the formula can be ‘z ~ 1’. If no formula is supplied, it defaults to ‘z ~ 1’.
overlap.threshold	a fraction threshold to remove observations when there is overlap between the vehicular polygons. A value of 0 does not remove any observations. A value of 1 removes all observations that overlap even minimally with neighboring observations.
var.label	optional string to name the final product. Defaults to ‘yield’.
boundary	optional sf object representing the field's outer boundary. If it not supplied, the function attempts to generate a boundary from the observed points.
clean	whether to clean the raw data based on distance from the field edge and global standard deviation.
clean.sd	standard deviation above which the cleaning step will remove data. Defaults to 3.
clean.edge.distance	distance (m) from the field edge above which the cleaning step will remove data. Defaults to 0.
smooth.method	the smoothing method to be used. If ‘none’, no smoothing will be conducted. If ‘idw’, inverse distance weighted interpolation will be conducted. If ‘krige’, kriging will be conducted.
fun	a function used to transform the data. Currently, the option are ‘none’ and ‘log’. If none, data operations are carried out in the data scale. If log, the function will usekrigeTg to perform kriging in the log scale. For now, only relevant when ‘method’ is krige. the log scale and back transform predictions to the data scale. When TRUE, ‘fomula’ should be ‘z ~ 1’.
lbs.per.bushel	a numeric value representing the number of pounds in a bushel (e.g., 60 for soybean and 56 for corn). This argument can be ommitted when the input and output units are in the metric system. It is necessary otherwise.
moisture.adj	an optional numeric value to set the moisture value to which the yield map predictions should be adjusted (e.g., 15.5 for corn, and 13.0 for soybean). If NULL, the function will adjust the moisture to the average moisture of the field.
lag.adj	an optional numeric value used to account for the time lag between the crop being cut by the combine and the time at which the combine records a data point.
unit.system	a string representing the unit system to be used in the function output. If ‘standard’, the function output will be in bushel/acre. Alternatively, if ‘metric’, outputs will be in metric tonnes/hectare.
remove.crossed.polygons	logical, whether to remove vehicle polygons that crossed different experimental units of the grid. This is intented to prevent from diluting the treatment effects. When this argument is TRUE, the argument ‘grid’ must be supplied.
steps	EXPERIMENTAL - whether to return the intermediate steps of the yield processing algorithm
cores	the number of cores used in the operation
verbose	whether to print function progress. ‘FALSE or 0’ will suppress details. ‘TRUE or 1’ will print a progress bar. ‘>1’ will print step by step messages.
...	additional arguments to be passed krige and idw

Details

This function will follow the steps in the selected algorithm to produce a yield map from the raw data.

Value

an object of class yield

Author(s)

Caio dos Santos and Fernando Miguez

Examples

## Not run: 
extd.dir <- system.file("extdata", package = "pacu")
raw.yield <- sf::read_sf(file.path(extd.dir, '2012-basswood.shp'),
                         quiet = TRUE)
## the simple algorithm
pa_yield(input = raw.yield,
algorithm = 'simple',
unit.system = 'metric',
lbs.per.bushel = 56) ## 56 lb/bushel of maize

## the ritas algorithm
pa_yield(input = raw.yield,
algorithm = 'ritas',
unit.system = 'metric',
lbs.per.bushel = 56)

## End(Not run)

pacu documentation built on June 8, 2025, 10:44 a.m.