It is often useful to know if a path is loop, lollipop, out-and-back or one-way. Add a computed function that describes how much redundancy a path has.
Example idea: arithmetic mean(for all points a, the min(euclidian_dist(a,b)/dist_along_path(a,b) for all other points b))
This should be 0 for out-and-back, 1 for a perfect collinear one-way, and somewhere in the middle for loops with lower values for lollipops.
To give equal weight to all real points on the path, this would have to un-optimize paths. You can't just compute based on the path points that are in the file. Must look at every point along X distance intervals.
It is often useful to know if a path is loop, lollipop, out-and-back or one-way. Add a computed function that describes how much redundancy a path has.
Example idea: arithmetic mean(for all points a, the min(euclidian_dist(a,b)/dist_along_path(a,b) for all other points b))
This should be 0 for out-and-back, 1 for a perfect collinear one-way, and somewhere in the middle for loops with lower values for lollipops.
To give equal weight to all real points on the path, this would have to un-optimize paths. You can't just compute based on the path points that are in the file. Must look at every point along X distance intervals.