From points to surfaces
An editable processing recipe distinguishes an isolated object model from the surface of an entire scene.
02 / iOS · Spatial capture
Magic CameraThe world has more than two dimensions.
I am exploring what I can build with a phone’s LiDAR: capturing a room or an object in 3D, editing the model and preparing it for further use.
EXPLORING SPATIAL DATA
A point cloud is a starting point.
I wanted to capture a room or an object with a phone, then edit the model, export it or show it to someone in a browser. Magic Camera grew out of exploring the LiDAR depth sensor and spatial data.
The app combines depth capture, point processing, surface reconstruction, photographic textures, editing and export. The result can move into a 3D tool or a standalone interactive HTML viewer.
- 01
Capture a space
The mode and detail level affect which depth data the app collects.
- 02
Process and edit
The recipe continues with reconstruction and, optionally, a photographic texture. The model can then be edited further.
- 03
Use the result
Export distinguishes a point cloud, a mesh and a textured model; the HTML version lets people explore the result interactively.
BROWSER SIMULATION
Try capturing the third dimension.
Try selected workflows directly in your browser.
This is a separate simulation using fictional data.
MAGIC CAMERA / INTERACTIVE DEMOSIMULATION · NO SIGN-IN
Loading the demo… A text overview is available below.
What you can try in this demo
- Capture modes — On the home screen, choose Spatial Scan for a room demo or Quick 3D for an object. Room Plan and Object Capture first explain the corresponding feature of the actual app, then offer a prepared 3D demo.
- Start a scan — Choose a space and detail level, then start the synthetic point capture.
- Explore in 3D — Switch between the point cloud and surface. Drag the scene or use the rotation buttons.
- Gallery — Finish the scan and reopen the model in the gallery.
TECHNICAL DECISIONS
The full path from capture to sharing.
Editable geometry
Model Studio adds shapes, transformations, colours and Boolean operations on geometry.
Ready for further use
Point clouds and models have their own export formats. An HTML viewer using three.js is included too.
An object model and a surface serve different purposes
Object isolation and closing the base suit a model; retaining the wider scene suits a room surface.
Processing and viewing are separate
Capture and processing run in the native app. The standalone web export is for viewing the result.
The assistant uses Studio tools
On a compatible system, a language model can call deterministic operations. Manual editing does not depend on it.
My contribution
Design and development of my own iOS project combining spatial capture, reconstruction, geometry editing and sharing the result.
Project and demo status
The experimental project includes capture, model editing and export, including an interactive HTML viewer. The web demo lets you explore a prepared 3D scene; it is not a real scan or a measurement of LiDAR accuracy.
EVIDENCE AND LIMITS
The sources are local code, a verified public commit and the owner's screenshots. LiDAR, RoomPlan, Object Capture and the Studio assistant depend on a compatible device and operating system. Processing is local; the app may also use iCloud storage depending on settings and availability.