Step in Rasterize function

[Sile]

From my understanding, the step function is starting from the bottom and then working its way up to the top of the point cloud identifying each point it comes in contact with and saving its height value. If a point is located between a step, is the point given a height value of zero?

[daniel]

Hum, not really.

Each point is projected in the nearest grid cell (each cell as a dimension of 'step' x 'step'). Basically you divide the coordinate expressed relatively to a corner of the grid and you divide it by 'step', and then you keep the integer part to deduce the cell position in the grid.

And we 'add' the height of each point to the corresponding cell, depending on the current parameter (we keep either the lowest, the highest or the average height).

[Sile]

Is it possible then for a point in the middle of a tree to be recognized as a base point, therefore dropping it to the ground when creating a DSM ?

[daniel]

As usual, I'm not sure to understand ;D

[Sile]

I am sure you are aware that this is related to creating the DSM from the other question forum. I set a step of 0.0025 to maximize the resolution. When I change the scalar field property to "height range". Some of my points in various heights on each tree stem turn blue, indicating they are ground points. When I set the scalar field to Z coordinates, those blue points drop to the ground, leaving me a lesser resolute tree.

[daniel]

If you set a too small grid step, then there's a chance that only one point that falls inside a cell. Or even if there's several points, there's a good chance they will all be very close to each other (in altitude). In this case, the 'height difference' inside the cells will be very small.