Can SmartData host Gaussian Splats now?
SmartData supports Gaussian Splat hosted-view package shapes. Processing from raw imagery is the part being productised with the photogrammetry engine.
Host and prepare Gaussian Splat views for immersive 3D review
SmartData supports Gaussian Splat package shapes today and is building the processing path that turns drone imagery into cleaner splat views for high-fidelity asset context.
Status
Preview
Best-fit audience
Asset teams needing immersive visual context
A future-facing landing page for teams wanting more photorealistic 3D context than conventional mesh or point cloud outputs can provide.
A photorealistic 3D view layer for complex assets.
Gaussian Splatting works as premium visual context, not a survey-grade measurement replacement.
Pair splats with 3D Tiles, point clouds, orthos, inspection imagery, and reports.
What this unlocks
Connect source data, SmartData processing, hosted views, and review outputs so operational teams can move from capture to decisions with less file handling.
BYO Gaussian Splat hosted-view packages
SmartData package root: gaussiansplat-view
model.json driven viewer contract
Pipeline research through COLMAP, Nerfstudio/Splatfacto, and splat filtering
QA reports for visual fidelity and package evaluation
Tie-in to the photogrammetry engine roadmap
Workflow story
1
Package a supported splat output with model.json and referenced assets.
2
Use BYO hosting for viewer-ready splat packages today.
3
For processing-preview work, solve cameras and train the splat outside SmartData's web request path.
4
Filter and QA the splat before packaging it for SmartData.
5
Publish the view for stakeholder review, comparison, and handoff.
Commercial use cases
- A photorealistic 3D view layer for complex assets.
- Gaussian Splatting works as premium visual context, not a survey-grade measurement replacement.
- Pair splats with 3D Tiles, point clouds, orthos, inspection imagery, and reports.
Delivery notes
- The repo contract already defines gaussiansplat-view package roots.
- ODM does not directly produce splats; the pipeline uses camera solve and Splatfacto-style training.
- Survey-grade measurement claims should come from the underlying georeferenced workflow.
Asset teams needing immersive visual context
SmartData gives this team a hosted workflow for turning complex capture outputs into reviewable evidence.
Drone processors experimenting with 3D Gaussian Splatting
SmartData gives this team a hosted workflow for turning complex capture outputs into reviewable evidence.
Infrastructure teams reviewing towers, substations, bridges, and complex sites
SmartData gives this team a hosted workflow for turning complex capture outputs into reviewable evidence.
Common questions
Straight answers about inputs, availability, outputs, and where human review still matters.
Are splats survey-grade?
They are best positioned as immersive visual context. Survey-grade claims should come from the underlying measurement workflow, not the splat alone.
What will reviewers see?
Reviewers should see an immersive splat view alongside related hosted dataset context, such as 3D Tiles, point cloud outputs, or QA/reporting handoff states where available.