Asked by Right Grievous
on 16 Jan 2013

Hi,

I am using scripts written by someone else a long time ago (who has long since gone) and I'm having trouble getting them to do something which is probably quite simple.

I have data from some tracking software and I'm trying to plot the x and y coordinates, as you would do normally but I want to show where the object spent the most time in the form of a pcolor map (so red will signify lots of time spent in an area etc).

I have a matrix (4 x ~30000), rows 1 and 2 contain the x and y coordinates and row 4 contains time in ms (at a 20ms sampling rate I believe).

I am sure I will need to bin the data somehow but I'm not sure how to do this...

Ultimately I want to do the same with velocity (a pcolor map with red signifying 'high speed') but I thought I would tackle this problem first.

Thank you very much for any help,

Rod.

*No products are associated with this question.*

Answer by Doug Hull
on 16 Jan 2013

Edited by Doug Hull
on 16 Jan 2013

Accepted answer

This is a 2-d histogram problem.

http://www.mathworks.com/matlabcentral/fileexchange/29709-function-to-make-a-2d-histogram

Here is a link to a MATLAB file Exchange file that should calculate a 2-d histogram for you.

This video might be good for you too. http://blogs.mathworks.com/videos/2010/01/22/advanced-making-a-2d-or-3d-histogram-to-visualize-data-density/

Doug

Right Grievous
on 16 Jan 2013

Hi Doug and thanks for the reply!

I had a look at your second link and tried out the code there, it does exactly what I want, which is great and really smart (I am guessing that it is plotting x and y coordinates into bins, in my case these are coded to a specific timebase which means that this also plots time?).

In any case I was wondering if it is possible to smooth the plot more? I want to produce a pcolor plot as we use this for similar purposes but we apply gaussian smoothing to these so they obviously look a lot more pleasing to the eye... is it possible to incorporate a smoothing filter into your piece of script?

Thanks,

Rod.

Image Analyst
on 16 Jan 2013

With few data points, it obviously will look quantized. The best option is to just collect some more data. A less desirable option is to blur your 2D histogram as if it were an image. For example

blurred2DHist = conv2(original2DHist, ones(5)/5^2, 'same'); imshow(blurred2DHist); colormap(hot); colorbar;

Opportunities for recent engineering grads.

## 0 Comments