Build a strip-by-strip prescription map from a field polygon

Generates a set of parallel, machine-width strips that tile a field polygon along a user-supplied A-B line (or the longest edge of the field, if no A-B line is given) and assigns a target N rate to every strip according to the chosen variability method. The output is ready to be passed to export_prescription for SHP / GeoJSON / ISOXML / John Deere / Trimble export.

Usage

build_strip_prescription(
  field,
  machine_width = 24,
  cell_length = NULL,
  angle_deg = NULL,
  ab_line = NULL,
  variability = c("uniform", "calibration", "nni", "classes"),
  vi_raster = NULL,
  nni_raster = NULL,
  n_target = 160,
  min_dose = 40,
  max_dose = 220,
  vi_low = 0.35,
  vi_high = 0.8,
  thr_lo = 0.9,
  thr_hi = 1.1,
  n_classes = 5,
  preserve_mean = TRUE,
  crs_metric = 3857
)

Arguments

field: sf POLYGON (single feature) of the field boundary.
machine_width: Numeric, working width of the spreader / spray boom in metres (default 24).
cell_length: Optional numeric. If supplied and >0, every strip is further subdivided into rectangular cells of this length (metres) along the A-B direction, producing a true 2-D prescription grid. If NULL or 0, each strip is left as a single polygon (the classic "strip map").
angle_deg: Optional numeric. Azimuth of the A-B line in degrees (0 = east, 90 = north). Overrides ab_line when supplied.
ab_line: Optional sf LINESTRING giving the driving direction. If both ab_line and angle_deg are NULL, the long side of the field is used.
variability: One of "uniform", "calibration", "nni", "classes".
vi_raster: terra::SpatRaster or raster::RasterLayer with vegetation index (0-1). Required for "calibration" and "classes".
nni_raster: Same, with NNI values. Required for "nni".
n_target: Target field-level mean dose (kg N ha\(^{-1}\)).
min_dose, max_dose: Numeric caps (kg N ha\(^{-1}\)).
vi_low, vi_high: Two-point calibration anchors (variability = "calibration"). Below vi_low the dose is capped to max_dose; above vi_high to min_dose.
thr_lo, thr_hi: NNI thresholds (variability = "nni").
n_classes: Integer number of equal-spaced dose classes for variability = "classes" (default 5).
preserve_mean: Logical. If TRUE (default) the strip- level dose is rescaled so the area-weighted mean matches n_target. Guarantees mass-balance compliance with the balance-based field ceiling.
crs_metric: EPSG code of a metric CRS used internally (default 3857, Web Mercator). Use a local UTM for higher accuracy on larger fields.

Value

An sf object of strip polygons with columns strip_id, dose (kg N ha\(^{-1}\)), area_ha, mean_vi or mean_nni when available. CRS matches the input.

Details

The builder works entirely in a metric CRS (Web Mercator by default); the final sf object is returned in the original CRS of the input field so that downstream tools see consistent coordinates.

Variability methods

"uniform": Every strip receives the same n_target dose. Useful as a baseline.
"calibration": Samples the mean vi_raster value per strip and maps it to a dose via a two- or three-point calibration curve. The dose decreases with vigour (higher VI = less N).
"nni": Samples the mean nni_raster value per strip and translates it into the three agronomic zones of Lemaire & Gastal: deficient (NNI < thr_lo) gets max_dose, optimal (thr_lo <= NNI <= thr_hi) gets n_target, excessive (NNI > thr_hi) gets min_dose.
"classes": Assigns doses equal to seq(min_dose, max_dose, length.out = n_classes) based on the k-quantile class of the mean vi_raster value per strip. Useful when a categorical map is preferred.

Regardless of the method, the mean dose across the field is rescaled to match n_target (mass-balance constraint) when preserve_mean = TRUE (default).

A-B line handling

When ab_line is NULL, the A-B direction is taken from the longest edge of the polygon's minimum-area bounding rectangle (via sf::st_minimum_rotated_rectangle()). This typically recovers the "long side" convention farmers use in the field. When ab_line is a sf LINESTRING, its first segment defines the driving direction.

Examples

if (FALSE) { # \dontrun{
# Pick one feature from the demo farm
ex <- system.file("extdata/example_farm.geojson", package = "NFert")
farm <- sf::st_read(ex, quiet = TRUE)
field <- farm[1, ]  # 5.2 ha silage-maize plot

# Uniform 180 kg N/ha, 24 m spreader
rx <- build_strip_prescription(field, machine_width = 24,
                                variability = "uniform",
                                n_target = 180)

# VI-based 40-180 band calibration, 36 m machine
rx2 <- build_strip_prescription(field, machine_width = 36,
                                 variability = "calibration",
                                 vi_raster = my_ndvi,
                                 n_target = 160,
                                 min_dose = 40, max_dose = 200)

export_prescription(rx, "strips.shp", format = "shp")
} # }