This approach uses duplicate elements node:
- 1.seperate one point on each spline, that will represent raycasting ID
- 2.set spline Z positions to their ID and extrude
- 3.duplicate points by spline domain size and set their Z to duplicate index
- 4.now you have iterations separated by their Z position I you can start raycast
My english is limited (sorry no AI transl) so below are the images explaining the process + two blend files. One purged and second with extra nodes to help visualize and understand the process. I have not tested this in many scanarios yet, but it seems to work ok and the timing in heavy scenarios is 10-30x faster comparing to the fill curve method.
thanks Markus with help on this!
result: 
engine: 
data for final logic: 
file with extra nodes:
file purged:
