"a table grid and drop-in terrain pieces" Topic

My gaming friends mostly live in other towns and states and so I want to have players remotely able to play on my table, looking at an ACCURATE map (not my hand-drawn and not-to-scale sketch of the table) and sending orders by email etc. .

Just started this project with GROK and we shall see how it develops. I did a grid for my 5' by 7' table with 1' squares labeled. I first tried getting the AI to put a terrain feature into a specific grid location, but it had trouble doing two things at once, location and also the look and size of the terrain piece. It suggested and I am now doing it in steps:

I am taking top down photos of each piece, with rulers laid out to show size. I put a white disk with ID number in the top center of the north edge. (The ID can be removed later.) I tell the AI to work from the photo, and I have learned not to confuse it with a written description. Once I have my "library" or "menu" of terrain pieces I can have the AI drop them onto the grid. So keep the steps separate.

The main limitation at this point is that top down does not show height clearly. I am also taking side views and the AI is accepting those and waiting for more instructions. What I HOPE is that it will be able to generate angled "player's view" images of the table with the terrain pieces accurately displayed, as well as the top down view.

You can try some of these. They use photogrammetry to make 3d models from multiple angle pictures.

RealityCapture (Epic Games): Industry standard for speed and high-detail geometry. Handles thousands of photos and outputs high-resolution textured 3D assets.

Meshroom (AliceVision): A free, open-source photogrammetry software that provides an intuitive node-based workflow for reconstructing 3D models from images.

Luma AI (NeRF & Gaussian Splatting): Allows rendering photo-like interactive 3D captures from multi-angle photos or videos, viewable in a browser or exported as 3D meshes/splats.

Polycam: One of the most accessible apps for iOS and Android. Converts multi-angle photos into 3D models using cloud photogrammetry and Gaussian Splatting.

Tripo3D (Tripo Multiview): Generates clean 3D assets from 2 to 4 multi-angle reference images.

Disclaimer: I don't personally use any of those. They are recommendations from people who professionally make digital twin environments.

Thanks. My GUESS is that Grok knows a lot of those tricks, and is already well informed about my project in other respects. But I shall keep your list just in case. As I am already doing rules and orders of battle and scenarios with GROK, it will be convenient if everything can be under one AI umbrella.