MATLAB Answers

0

Binning 2D set of coordinates

Asked by Akhil Ghanta on 19 Jun 2015
Latest activity Edited by Walter Roberson
on 13 Oct 2016
I'm pretty new to MATLAB, so sorry if this seems easy. I've got a single long trajectory for a particle below. I have about 100,000 X and Y coordinates from which I've assembled the graph. I'm trying to create a density plot of the coordinates to show where the trajectory is most concentrated. I'm trying to create a density plot of the coordinates to show where the trajectory is most concentrated. I've got the second plot using the following code:
bins = 200;
binsizeX = 3/bins;
binsizeY = 1.2/bins;
xbins = -1.5:binsizeX:1.5;
ybins = -0.6:binsizeY:0.6;
[nx,idxx] = histc(X,xbins);
[ny,idxy] = histc(Y,ybins);
out = accumarray([idxx,idxy], 1);
figure(2);
hold on;
h=imagesc(xbins,ybins,out);hold on;
colorbar; hold on;
axis([-1.5 1.5 -0.6 0.6]); hold on;
But it seems that the density is shifted up and to the right for some reason? Any help is appreciated! Thanks!

  0 Comments

Sign in to comment.

Tags

4 Answers

Answer by alessandro on 13 Oct 2016
Edited by Walter Roberson
on 13 Oct 2016

Old post, but will try to answer anyway, in case anybody else end up with same issue.
The easiest way to do what you are willing to do is to use hist3 function, with which binning resolution is user dependent and density matrix can be stored easily.

  0 Comments

Sign in to comment.


Answer by Katalin
on 19 Jun 2015

imagesc flips the image I think. This is from the help: imagesc(x,y,C) displays C as an image and specifies the bounds of the x- and y-axis with vectors x and y. If x(1) > x(2) or y(1) > y(2), the image is flipped left-right or up-down, respectively.

  1 Comment

If that's true why are the axes still going in the proper direction on the plot? Also, for my vectors x(1)<x(2) and y(1)<y(2)

Sign in to comment.


Answer by Image Analyst
on 19 Jun 2015

With images, since they are arrays, line 1 (row 1) is at the top, and the line numbers get bigger as you go down the screen. This is the standard custom with images.
With a plot, since it's like regular Cartesian coordinates, where the "y" value increases as it goes up the screen.
The right/left are not flipped, just the up/down. You could use flipud(out) to flip the image vertically before displaying it if you want the plot and image to look similar.
imagesc(xbins, ybins, flipud(out));

  2 Comments

Well now, the density is just flipped. The slope of the ellipse is now negative. The big thing is that the density isn't centered like the original trajectory is in the original post.
Image Analyst
on 20 Jun 2015
Is the centroid supposed to be at (0,0)? Please attach your data file with the paper clip icon, along with code to read it in.

Sign in to comment.


Answer by Walter Roberson
on 19 Jun 2015

out = accumarray([idxx,idxy], 1);
should be
out = accumarray([idxy,idxx], 1);
Height (y) is expressed as row number so the y index needs to be first.

  2 Comments

This did seem to fix the slope of the density ellipse, but it's still not centered on the origin like I'd expect it to be.
Could you attach the data as a file?

Sign in to comment.