About the project

Preserving Nepal's heritage, digitally

HamroSampada is a digital heritage archive and visualization platform dedicated to documenting and sharing Nepal's cultural monuments using modern 3D reconstruction and open web technology.

Project background

Nepal holds an extraordinary density of cultural heritage — from the UNESCO-listed monuments of the Kathmandu Valley to sacred temples, stupas and archaeological landscapes across the country. Much of this heritage is built of brick, timber and stone, and is vulnerable to earthquakes, weathering, urban pressure, theft of artefacts, and the slow loss of traditional craft knowledge.

The 2015 Gorkha earthquake was a stark reminder: in seconds, centuries-old structures were damaged or lost. Where detailed records existed, reconstruction was possible. Where they did not, knowledge vanished with the rubble. HamroSampada exists to make sure that detailed, shareable records exist before they are needed.

Why Nepal needs digital heritage preservation

  • • Safeguards a record against earthquakes, fire and decay.
  • • Supports accurate restoration and conservation planning.
  • • Makes heritage accessible to students, researchers and the diaspora worldwide.
  • • Documents craftsmanship and detail that photos alone cannot capture.

What HamroSampada does

3D reconstruction

We capture monuments with photogrammetry and 3D Gaussian Splatting to create photorealistic, measurable 3D records.

Structured archive

Each site is documented with history, architecture, condition, conservation notes, coordinates and references.

Map exploration

Every record is geolocated so heritage can be explored by place across all of Nepal.

Interactive 3D viewing

Anyone can rotate, zoom and pan through captured monuments directly in a web browser.

Holographic projection

A working Raspberry Pi holographic unit shows captured monuments as floating, gesture-controlled holograms.

Native mobile AR

Scan a monument's QR to open it in Google Scene Viewer or Apple AR Quick Look — placed in your real space.

The technology

3D Gaussian Splatting, in plain language

Imagine taking many photographs of a temple from every angle. Older 3D methods tried to build a solid mesh of triangles from those photos — often losing fine detail and looking artificial.

3D Gaussian Splattingtakes a different approach. It represents the scene as millions of tiny, soft, coloured 3D blobs (“splats”), each with its own position, colour, transparency and shape. Together they recreate the monument with remarkable realism — soft edges, subtle colour and depth — and they render fast enough to explore live in a web browser.

1

Capture

Photograph or film the monument from many viewpoints.

2

Reconstruct

Train a Gaussian Splat model that learns the 3D scene.

3

Explore

View and interact with the result right in the browser.

The web portal's purpose: make these captures — and the knowledge around them — freely explorable by anyone, on any device, as a trusted public record. The Chandeswori Temple is our first reference capture.

The archive today

01
37
Monuments archived
02
22
Districts covered
03
2
3D models available
04
8
Heritage categories
The roadmap

Future vision

This MVP already runs the full stack — archive, 3D viewer, holographic unit and native mobile AR. Next, we scale coverage and tooling toward a complete, government-compatible digital heritage system.

A larger web archive

Scale from reference monuments to a comprehensive national record, contributed to by institutions and communities.

Admin & data-entry system

Authenticated dashboards, review workflows and a database so heritage officers can manage records directly.

Holographic touring kit

Take the working holographic unit on the road — schools, festivals and regional museums across Nepal.

AR model library

A growing set of .glb and .usdz captures so every monument can be placed in your space in native AR.

Explore the archive

Browse documented monuments, see them on the map, and experience our reference 3D capture.

Project credits

Project team

  • Sanjay Pahari
  • Utkrist Mani Neupane
  • Bipul Lamsal

Mentor

Prof. Dr. Rabindra Bista

Department

Computer Science and Engineering