Label each record's platform from how fast it was moving

Classifies every record as "aerial", "vessel" or "stationary" using the median track_speed() of the stretch it belongs to. A whole stretch takes one label, because a platform does not change mid-line.

Usage

classify_platform(
  dat,
  by = NULL,
  aerial_min = 40,
  stationary_max = 2,
  max_gap = 300
)

Arguments

dat

A data frame with LATITUDE, LONGITUDE and TIME.

by

Passed to track_speed().

aerial_min

Knots at or above which a stretch is aerial. Default 40, comfortably between a surveying vessel and the slowest survey aircraft.

stationary_max

Knots at or below which a stretch is stationary. Default 2.

max_gap

Passed to track_speed().

Value

A factor with levels "stationary", "vessel", "aerial", one per row, NA where the stretch has too little usable time to judge.

What this is for

A single NARWC extract can hold both an aerial and a shipboard survey with nothing to tell them apart. On the file this was built against, PLATFORM was NA on all 1,394,556 records and the aerial LEGTYPE codes were used throughout — but 738 line occupations flew at 40-131 knots and 299 moved at about 10. That matters because the distance machinery is not platform-agnostic: a declination angle and a strip width are aircraft measurements, and a shipboard survey measures by reticle and bearing.

Prefer a recorded signal where the file has one. On that extract every aerial occupation carried a LEGNO and no vessel occupation did, agreeing with speed on 1037 of 1038 — a line number is a reading, where a speed threshold is an inference.

See also

track_speed()

Examples

path <- system.file("extdata", "narwc-example.csv", package = "narwcr")
dat <- make_leg_id(read_narwc(path, quiet = TRUE), quiet = TRUE)
table(classify_platform(dat), useNA = "ifany")
#> 
#> stationary     vessel     aerial 
#>          0          0        113