Rasterize

[sebastien023]

Hi,

first of all, thank you for this great piece of software! It's really usefull and powerfull.

Here is my problem:
I've got a cloud representing a piece of landscape with slopes, carriers... The dimensions are approximalty 900 m X 900 m, and the difference betweem the minumum and the maximum height about 300 m.
I use CANUPO to filter the vegetation with a very good result.
Then, I want to rasterize the cloud (direction Z, step of grid 0.5) in order to get a raster and extract the contours in a GIS software.
My problem is that the filtred cloud without the vegetation has plenty of holes (the place of the vegetation before) and I'm not able to get a raster with the "average" value of height instead of a hole. If I choose average value in the menu "fill empty cells", CC fill with the average value of all the cloud and not only with a local value.
Is there any method to fill the holes in order to get a average Z value and then get a coutour line fitting with the reality.

Thanks for the answer!

[daniel]

Indeed, there's no way right now to 'interpolate' the values inside holes.

In fact it wouldn't be that hard (the qSRA plugins do exactly the same thing - sadly it's not applicable to your case ;). I'll add this to the TODO list.

[sebastien023]

Thanks for your answer.

I've you got any tips to fill in the holes (manually?).
I tried to fit a plane with the border of each hole but it's a very long job and the result is not good.
Is there a way to work on the vegetation filtred cloud and get a minimum value of each "ball" representing the trees?

Have a nice day.

[daniel]

The easiest way would be to mesh the rasterized cloud (with holes) with the Delaunay 2D1/2 tool.

Then sample points on the mesh.

And eventually you can either manually segment the resulting cloud to get the points inside each hole, or better you can compute the distances between this cloud and the original one (as reference) and segment the points with high distances (with Edit > Scalar Fields > Filter by value).

For a nice result you should merge the original cloud (non-rasterized) and those points, and then re-launch the rasterize tool on the resulting cloud.

Hope you understood everything ;)