Spatial data visualization level of detail

Question

I have a 3D point cloud data set with different attributes that I visualize as points so far, and I want to have LOD based on distance from the set. I want to be able to have a generalized view from far away with fewer and larger points, and as I zoom in I want a more points correctly spaced out appearing automatically.

Kind of like this video below, behavior wise: http://vimeo.com/61148577

I thought one solution would be to use an adaptive octree, but I'm not sure if that is a good solution. I've been looking into hierarchical clustering with seamless transitions, but I'm not sure which solution I should go with that fits my goal.

Any ideas, tips on where to start? Or some specific method?

Thanks

Answer 1

i would consider looking into GIS based methods firstly although i do not know exactly how this would relate within a game.

https://gis.stackexchange.com/

if i were to write from scratch for simplicity purposes however, i would scale the 2d positional Vector and group the value by those scaling within a given range i.e if you had a 1000 pixel map condensed into a 100 pixel area you could simply sum the values based upon the Vector2D /10 and provide a single pixel value color based upon this. obviously you could extrapolate this for 3d.

Would probably work very well as a shader but my knowledge does not stretch anywhere near that far

(not sure if this goes anyway near to answering)