Accessing initial point data inside of point metrics functions with lidR

Question

I am using the R package lidR to do some analysis on a large forestry point cloud. I am trying to write my own point_metrics() function for identifying outliers. How can I access the X, Y, and Z coordinates of the point that the metric is being run on inside of the function? For example with the following code:

is.outlier <- function(x, y, z) {
  return(x)
}
point_metrics(las, ~is.outlier(X, Y, Z), k = 5)

returns the x values of the point's k nearest neighbors. I want to find the x value of the point in addition to the k nearest neighbors in order to find the distance away each knn is from the initial point.

Answer 1

Actually what you are receiving in your function is the processed point + its k-1 neighbors because the current point is considered as the 1-neighbour with a distance 0 to the reference point. I do agree that this is not explicit in the documentation. So to access the central point you can simply use x[1]. Other points are ordered by distance.

is.outlier <- function(x, y, z) {
 dx = x[1] - x[-1]
 dy = y[1] - y[-1]
 dz = z[1] - z[-1]
 sqdist = dx*dx+dy*dy+dz*dz
 return(mean(sqdist) > 250))
}

By the way you cannot do:

f <- function(x) { return(x) }
point_metrics(las, ~f(X), k = 5)

because f is expected to return a single or a list of scalar values. Here you are returning a vector. Your example will fail.