BHARAT 3D: 3D Land Administration Architecture for India
From a 2D Parcel
to a 3D Property
Our journey towards building a 3D ULPIN Mapping System for representing vertical, underground and elevated property.
We started with a simple question: What happens when land is no longer just a flat piece of land?
A conventional cadastral map is fundamentally two-dimensional. But modern cities are not.
Apartments occupy different floors. Parking can exist below ground. Metro lines can pass beneath properties. Roads can pass below buildings. Tunnels, pipelines and elevated structures can occupy the same geographic location at different heights.
This made us realize that the problem was not simply about drawing buildings in 3D. The real problem was understanding how a property identity could be connected to a volumetric space.
1. Understanding the Existing Foundation
India's ULPIN — Unique Land Parcel Identification Number — provides a digital identity for land parcels based on cadastral and geographic information.
Our first decision was therefore important: we should not try to replace the existing 2D parcel.
Instead, the 2D parcel can remain the administrative foundation while additional 3D spatial objects are associated with it.
2. Why 3D Is Necessary
Consider a single parcel containing a multi-storey apartment. A 2D footprint may show the building's location, but it does not naturally distinguish the spatial extent of individual floors or units.
The same problem becomes even more interesting underground. A surface parcel, basement, metro tunnel and utility pipeline may occupy overlapping areas in X and Y while existing at completely different elevations.
| Object | Spatial requirement |
|---|---|
| Land parcel | 2D boundary |
| Apartment unit | 2D boundary + vertical extent |
| Basement | Subsurface volume |
| Metro / tunnel | 3D corridor or volume |
| Elevated structure | 3D spatial extent |
3. What We Learned From Existing Research
Research from countries such as the Netherlands and Singapore showed us that 3D cadastre is not simply a visualization problem. It involves legal objects, cadastral registration, survey control, standards, data models and institutional workflows.
Research on Delhi also highlighted that traditional 2D representations become increasingly limited as urban space becomes vertically complex.
One of the most important lessons was therefore: technology alone cannot solve a cadastral problem.
A practical system has to respect existing records and introduce 3D capabilities incrementally.
4. Our 3D ULPIN Mapping Model
We explored a model where a conventional parcel remains the parent spatial object and volumetric objects are associated with it.
For example, a floor can be represented as a collection of units:
This gives us a simple conceptual bridge between a 2D cadastral footprint and a three-dimensional spatial object.
5. Proposed Volumetric Identity
We explored a hierarchical identifier called a VPRID (Volumetric Property Reference ID).
Here the existing parcel identity can act as the parent reference, while the additional components identify a building, floor and unit.
The important idea is not the exact string format. The important idea is maintaining a clear relationship between parent parcel → building → floor → volumetric unit.
6. The Data Problem
Building a 3D mapping system immediately exposed another problem: where does the 3D data come from?
A useful system may need to combine several sources:
7. Where AI Fits
We do not see AI as a replacement for surveyors. We see it as an assistant for processing large amounts of spatial data.
A possible pipeline is:
AI can assist with building segmentation, LiDAR classification, height estimation and change detection.
But every automatically generated geometry should remain provisional until validated through an appropriate surveyor / administrative workflow.
8. Geometry Cannot Be Ignored
A 3D model can look correct on screen and still be geometrically wrong.
Therefore, our system needs computational geometry checks such as:
- Closed geometry
- Manifold validation
- Self-intersection detection
- Duplicate geometry detection
- Unit overlap checks
- Elevation consistency
- Topology validation
These checks help determine whether a generated model is technically valid. They do not, by themselves, determine legal title.
9. The System We Are Building
Our current engineering direction is a web-based 3D ULPIN Mapping System.
| Layer | Technology / Role |
|---|---|
| Frontend | React + TypeScript + Cesium |
| Backend | FastAPI |
| Spatial Database | PostgreSQL + PostGIS |
| 3D Streaming | 3D Tiles |
| Processing | Asynchronous spatial / AI workers |
| Surveyor Interface | Upload, inspect, edit and validate |
10. What We Faced
The hardest part was realizing that the problem was much larger than simply creating a 3D map.
- Data availability: high-quality, consistently structured 3D cadastral data is difficult to obtain.
- Legal complexity: a technical 3D object is not automatically a legally recognized property right.
- Data fusion: imagery, LiDAR, CAD, GNSS and cadastral datasets have different formats, coordinate systems and levels of accuracy.
- AI uncertainty: automated reconstruction can produce incorrect boundaries, heights or classifications.
- Scale: processing national-scale spatial datasets requires considerably more infrastructure than a simple web application.
- Validation: a visually impressive model is not enough. Geometry and source lineage have to be traceable.
11. What We Absorbed
Our thinking changed several times during this work.
Initially, we were thinking mainly about creating a 3D map.
Then we realized we were actually dealing with spatial identity, geometry, data lineage, validation and cadastral workflows.
That changed the direction of the project.
12. From Demonstration to Real System
Our current demonstrations are early prototypes. They help us test the concept, interface and workflow, but they should not be confused with a production cadastral platform.
The next stage is therefore not simply adding more visual effects. It is improving the actual pipeline:
- Reliable spatial data ingestion
- Coordinate-system normalization
- 3D object generation
- Automated geometry validation
- Surveyor correction workflow
- Persistent spatial storage
- 3D visualization and streaming
- Audit and data lineage
- Quantitative evaluation
13. Evaluation
We do not want to claim accuracy numbers without experiments. The system therefore needs measurable evaluation.
| Area | Possible metric |
|---|---|
| Building segmentation | IoU, precision, recall |
| Height estimation | MAE, RMSE |
| Geometry | Topology validity / error rate |
| System performance | Query and processing time |
| Human validation | Correction time and failure cases |
14. Why This Matters for India
India's cities are becoming increasingly vertical. Land is no longer used only at the surface.
Buildings, basements, transport infrastructure, utilities and underground structures increasingly share the same geographic footprint at different elevations.
A 3D cadastral approach could eventually provide a better spatial representation of these relationships while remaining connected to existing parcel identities.
The opportunity is therefore not simply to make maps look three-dimensional. It is to create a structured spatial foundation for increasingly three-dimensional cities.
15. Standards We Are Looking At
16. India: The Existing Digital Foundation
India's Digital India Land Records Modernization Programme provides the broader cadastral digitization context in which ULPIN operates.
More recently, official discussions have also explored synergies between ULPIN and DIGIPIN, including pilot projects and standards-based API approaches.
For our work, this reinforces an important principle: the 3D system should be designed as an extension around existing digital land infrastructure rather than as an isolated map.
17. Our Current Position
We are not claiming that we have solved India's 3D cadastral problem.
We are building and testing a technical approach for a 3D ULPIN Mapping System that explores how existing parcel identities could be connected with validated volumetric representations.
The prototypes are helping us understand what works, what breaks, what data is missing and what needs to be validated before such a system could move beyond demonstration.
18. What Comes Next
19. Our Research & Demonstrations
The work is accompanied by early demonstration repositories:
These repositories represent development and experimentation, not a nationally deployed cadastral system.
20. Final Thought
We began by asking how to put property into 3D.
We ended up asking a much bigger question: how should a digital land system understand space when the same location can contain multiple layers of property?
That is what our 3D ULPIN Mapping System is trying to explore.
Not replacing the existing cadastral foundation. Not pretending that technology can decide legal ownership. But creating a technically structured path from 2D parcel identity to validated 3D spatial representation.
Comments
Post a Comment