Automatic detection and removal of water with python. Is it possible?

[Braeden]

I scan a relatively large area frequently, collecting up to a few thousand photos each scan. After large rain events the area is covered in hundreds of puddles, making masking the water in photos not an option.

When the point cloud is exported and a surface created with it the puddles appear as areas of spikes which can be easily identified by eye, but are very time consuming to crop out.

Would there be any attribute of these points that could be used to filter them from the point cloud pre-export? Would points generated from water areas have worse point confidence or .... anything at all differentiating them from solid stationary points?

If so, how might I go about turning them off, filtering, or deleting them with python?

Thanks for any advice you can give