Visualization from 3D model and image

[Zsombor]

I have a 3D model integrated into a point cloud and a site photo taken from the same viewpoint. Can the AI Visualizer generate a high-quality visualization by merging these two image inputs, and what are the optimal settings/prompts for this workflow? Nano Banana (Gemini) gave me a great result for this, but I was wondering if I have to leave Vectorworks for that, or if it can be created within the AI Visualizer.

I am using version 2026.4.

Thank you!

[zoomer]

1 hour ago, Zsombor said:

Nano Banana (Gemini) gave me a great result for this,

 

That's pretty cool.

It really revived plants by using the existing plants. I'm impressed.

 

OK, the amount of gargoyles seems a bit too much but shows Gemini thinks about where the water is coming from 🙂

Edited April 2 by zoomer

[Tom W.]

That's one waterlogged hillside!

[Dave Donley]

The point cloud image is pretty noisy but it was able to do the correct outline of the patio and walls.  I changed the material after.  Could try again, I put image prompt 1 as the photo and 2 as the point cloud/model.  In the prompt you have to mention that there are two images one is the woodland site and you want the model added to it.

Gemini by default will change details, you can add a sentence telling it to preserve the exact geometry of the model and don't invent new outlines etc.

[Zsombor]

Thank you @Dave Donley, I'll play with this setup and try how it works for me.