Limiting orthographic camera to only rendering specific objects without using layers

I'm creating a large-scale 2D space strategy game that uses game world segmentation to get around unity's float restrictions. All segments would exist on top of each other in a unity scene, and interaction between objects in different segments is prevented.

Now, in order to render this into a coherent game world when the player zooms far out, it would be really useful to be able to make every segment have it's own camera which only renders objects from that segment only. I could then selectively stich the required camera feeds together using view port rectangles to form the full picture of the game world. Not entirely sure what the performance on something like this would be, but I'm willing to find out.

Using the culling mask behaviour for the cameras would work, but there is a limit on the amount of layers, which means that there would be a cap on the amount of segments that can be on-screen at once. Some kind of "Render only child objects" property for the camera would be perfect for this.

Just wondering if there is another way besides using layers. Any ideas?