unity – Optimizing 3D tile-based geometry

For my game, I have a tool that lets me quickly create map geometry. You can see a (albeit early version) demo of the tool here. The main problem with the tool is that even a semi-large map can EASILY amount to around thousands to even 10s of thousands of tiles. The easiest way to fix this, I found, was to use a script I found online to merge all of the tiles into one mesh at the start of runtime (this is because I want to still be able to edit my tilemap in the editor)

This works and has a noticable improvement to the performance (especially if post-processing is enabled) however I’m left with a mesh that has a bunch of triangles that just aren’t needed. You can see an image of the mesh here. The problem with this is that if I want to do any kind of point lights with shadows (baked lighting isn’t possible, really, due to the interactivity of my game since you can turn lights on/off and I have a weird day/night system that is split into multiple periods of the day and only changes after a mission/event) would mean I’m rendering that mesh again from each point lights POV (and I need a LOT of point lights). I can’t really just not use shadows since it causes very noticable light bleeding at walls.

I’ve tried to find some sort of algorithm to simplify the mesh and I came across a script that seems to do exactly what I want, but I couldn’t get it to do anything except cause my mesh to become completely black with seemingly no change in mesh topology. Here’s the script by the way

I’ve also found an algorithm called Greedy Meshing which seems to also do what I want, but seems to be meant for voxel based worlds like minecraft and I’m not sure how to implement it in my case since my tiles aren’t voxels.

I’ve been trying to find other algorithms but the various game dev servers on discord can’t seem to help me. Does anyone know how I can solve this problem? I’m not really smart enough to come up with my own algorithm/system to simplify the mesh.