PHP implementation of the Ramer–Douglas–Peucker algorithm for polyline simplification
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
sample
src
tests
vendor
.gitignore
LICENSE
README.md
after.png
before.png
beforeAndAfter.png
composer.json
phpunit.xml

README.md

RDP-PHP

PHP implementation of the Ramer–Douglas–Peucker algorithm for polyline simplification.

License: Public Domain

Example simplified polyline

Example polyline before and after simplification

Installation

composer require david-r-edgar/rdp

Example usage

use davidredgar\polyline\RDP;

$line = array(
    array(150, 10),
    array(200, 100),
    array(360, 170),
    array(500, 280));

$rdpResult = RDP::RamerDouglasPeucker2d($line, 30);

$rdpResult will contain a resulting array with the reduced number of points. For this example:

$rdpResult == array(
    array(150, 10),
    array(200, 100),
    array(500, 280));

The second parameter to RamerDouglasPeucker2d() is epsilon, the maximum perpendicular distance for any point from the line between two adjacent points. Try replacing it with, say, 10 or 50 and observe the results.

Use for geographic purposes

I originally implemented this in order to simplify a complex route on a map. Because I was doing this in Great Britain with OSGB36 coordinates, this worked.

However, be careful if you want to want to attempt this with other coordinate systems. The algorithm assumes cartesian coordinates on a 2D plane. Attempting to use latitudes and longitudes on the surface of a sphere will result in incorrect results. For approximate polyline simplification, the results may still be acceptable. As you approach the poles, errors will become more and more apparent: put simply, the degrees of longitude will become much closer to one another than the degrees of latitude, and so incorrect points will be chosen to be removed from the polyline.