pick {tripEstimation} | R Documentation |
pick
plots up series of light data agains record ID, allowing the user to click on the
beginnings and ends of twilight in sequence. picksegs
generates a vector
of segment IDs for each record.
pick(id, val, nsee = 10000) picksegs(twind, n)
id |
index vector to identify records |
val |
sequence of data (light levels) to choose segments from |
nsee |
number of points to plot per screen |
twind |
vector of index pairs generated by pick |
n |
Number of segments values required - length of record |
Choosing twilight segments interactively seems far easier than trying to automate it. Mark Hindell makes the point that you get to see the data in detail, which is good.
pick
returns a vector where each value (obtained using locator
is the x coordinate for the begin or end of a twilight.
picksegs
uses these paired indexes to return a vector of segment IDs, with NAs
for non-twilight periods.
Segments are expected to be chosen as non-overlapping.
It seems best to choose more of the light data than less, using the
ekstrom
keyword to solar.model
we can limit the solar
elevation used.
section{Warning }{Segments are expected to be chosen as non-overlapping. }
Michael D. Sumner
## Not run: d <- sin(seq(0, 10, by = 0.01)) id <- 1:length(d) ## choose a series of start-begin pairs pk <- pick(id, d, 1000) ## your start/ends should be marked as blue versus red plot(id, d, col = c("red", "blue")[is.na(picksegs(pk, 1000))+1]) ## End(Not run)