Skip to main content

Cubic spline interpolation

I spent a while today casting around for a simple, fast and just-good-enough cubic interpolation algorithm to use in the GeoTools library where it is to serve as a 'filter function' that can be used when rendering maps.

The ACM digitial library has a mind-boggling collection of references on interpolation methods, most of which were fascinating but also over the top for this application.

In the end I put together the following code based on a Wikipedia article about cubic Hermite splines. Here it is...

* Cubic hermite spline interpolation. This is adapted from the description of the
* algorithm at:
* Tangent caculations are done with simple finite differencing in the interests
* of speed.
* <p>
* The input arrays xi and yi contain the coordinates of four interpolation
* points defining three segments with the middle segment containing the point
* for which we seek an interpolated value.
* @param x x ordinate of the point for which we seek an interpolated value
* and which lies between xi[1] and xi[2]
* @param xi interpolation point x ordinates
* @param yi interpolation point y ordinates
* @return interpolated value for x
public double cubic(double x, double[] xi, double[] yi) {
double span01 = xi[1] - xi[0];
double span12 = xi[2] - xi[1];
double span23 = xi[3] - xi[2];

double t = (x - xi[1]) / span12;
double t2 = t*t;
double t3 = t2 * t;

double m1 = 0.5 * ((yi[2] - yi[1]) / span12 + (yi[1] - yi[0]) / span01);
double m2 = 0.5 * ((yi[3] - yi[2]) / span23 + (yi[2] - yi[1]) / span12);

double y = (2*t3 - 3*t2 + 1) * yi[1] +
(t3 - 2*t2 + t) * span12 * m1 +
(-2*t3 + 3*t2) * yi[2] +
(t3 - t2) * span12 * m2;

return y;

I make no claims about the elegance or efficiency of this code other than suspecting that it rates badly in both areas, but it does a perfectly acceptable job when I throw some test data at it...


Popular posts from this blog

Fitting an ellipse to point data

Some time ago I wrote an R function to fit an ellipse to point data, using an algorithm developed by Radim Halíř and Jan Flusser1 in Matlab, and posted it to the r-help list. The implementation was a bit hacky, returning odd results for some data. A couple of days ago, an email arrived from John Minter asking for a pointer to the original code. I replied with a link and mentioned that I'd be interested to know if John made any improvements to the code. About ten minutes later, John emailed again with a much improved version ! Not only is it more reliable, but also more efficient. So with many thanks to John, here is the improved code: fit.ellipse <- function (x, y = NULL) { # from: # # # Least squares fitting of an ellipse to point data # using the algorithm described in: # Radim Halir & Jan Flusser. 1998. # Numerically stable direct least squares fitting of ellipses. …

Build an application plus a separate library uber-jar using Maven

I've been working on a small Java application with a colleague to simulate animal movements and look at the efficiency of different survey methods. It uses the GeoTools library to support map projections and shapefile output. GeoTools is great but comes at a cost in terms of size: the jar for our little application alone is less than 50kb but bundling it with GeoTools and its dependencies blows that out to 20Mb.

The application code has been changing on a daily basis as we explore ideas, add features and fix bugs. Working with my colleague at a distance, over a fairly feeble internet connection, I wanted to package the static libraries and the volatile application into separate jars so that he only needed to download the former once (another option would have been for my colleague to set up a local Maven repository but for various reasons this was impractical).

A slight complication with bundling GeoTools modules into a single jar (aka uber-jar) is that individual modules make ext…

Circle packing with R

To visualize the results of a simulation model of woodland trees within R, I needed an algorithm that could arrange a large number of circles within a rectangle such that no two circles overlapped by more than a specified amount. A colleague had approached this problem earlier by sorting the circles in order of descending size, then randomly dropping each one into the rectangle repeatedly until it landed in a position with acceptable overlap.

I suspected a faster and more robust algorithm could be constructed using some kind of "jiggling the circles" approach. Luckily for me, I discovered that Sean McCullough had written a really nice example of circles packing into a cluster using the Processing language. Sean's program is based on an iterative pair-repulsion algorithm in which overlapping circles move away from each other. Based on this, and modifying the algorithm a little, I came up with an R function to produce constrained random layouts of a given set of circles. He…