33Visual | Capture Workflow Example

Caboose Scan Comparison: Immersive Tour vs. 3DGS Proof of Concept

This example shows the difference between a Realsee/Galois capture designed for immersive navigation and an FJD Trion P2 capture processed as an early 3D Gaussian Splat proof of concept. Both are useful, but they serve different purposes.

Galois M2 + Realsee

Immersive VR Tour

The Realsee version is strongest as an immersive walkthrough. It gives the viewer a clean way to move through the subject, understand the space, and experience the object from selected scan positions.

Best for navigation, presentation, walkthrough experience, and client-friendly viewing.
Exterior “dollhouse” views may not always look as strong as the inside navigation experience.
Ideal when the goal is to show a place or object through an accessible VR-style interface.
FJD Trion P2 + 3DGS

Gaussian Splat Proof of Concept

The FJD Trion P2 scan was done quickly as a proof of concept to test whether the capture could support a 3D Gaussian Splat workflow. The point cloud was excellent, and the visual result was promising, but more passes and stronger camera overlap would improve the final splat.

Best for testing exterior 3D reconstruction and Gaussian Splat potential.
The point cloud quality was strong; the visual camera data would benefit from more disciplined overlap.
This is not a one-button workflow. Better results require slower passes, repeated angles, and visual coverage.

What this example shows

This caboose is a useful comparison because it makes the workflow difference clear. Realsee/Galois is better for immersive tour presentation. The FJD P2 capture is better suited for point cloud acquisition and testing a 3DGS-style reconstruction workflow. For polished exterior models, the result depends less on a single device and more on capture discipline: overlap, exposure, distance, height variation, and repeated passes.

Realsee / Galois Stronger for immersive navigation and client-facing walkthroughs.
FJD P2 / 3DGS Stronger as a technical capture path for point clouds and Gaussian Splat testing.
Next Improvement More visual passes, more overlap, and slower camera movement would produce a stronger 3DGS result.
33Visual — Business Stories for Humans. Human Stories for Business. | 33visual.com