Smooth 2D data in non-regular distribution pattern for colormap-based representation

Hello everyone

I have a discrete set of data which looks like the one represented by the dots within the white background-based region displayed in this sketch

My goal is to obtain a smoother data set in the aforementioned region to represent it through the image environment. My approach to this problem, usually, would be to use RegularizeData3D, as suggested by @Mathieu NOE in his answer to a previous post, Smooth 2D colormap based on non-evenly distributed data. However, in this case, due to the spatial distribution, this approach does not seem very appropiate. Could someone recommend a way to increase the number of points within the white region, without potential interpolation being affected by the non-existence of values in the out-of-borders-based gray area?

Answer 1

[Image Analyst]

You could use the convex hull and then use linspace to create more points in between the vertices. But like I said in your other post, I don't think you need to smooth your data.

[Richard Wood]

Hi @Image Analyst

Thanks for your comment. If I don't interpolate my data, at least using uimagesc, I don't get a color map that is smooth, as you can see in this figure

I attach in this message the files that I have used, and I just did something like this

u1=figure('visible','off','units','pixels','position',[0 0 1920 1080]);
uimagesc(space_a,space_b,mc');

On the other hand, I am also interested in summing the mc matrix that I put in the previous plot, with the one given in the mc_2.txt file. The problem is that, even if the non-null 2D spatial region for both matrices is the same, these non-zero values do not fall into the same elements of the array. That is why I also sought to interpolate the data.