Airfree
Research & Development

Applied research at the frontier of spatial computing

Airfree's research division advances the science and engineering of geospatial intelligence — publishing peer-reviewed work, releasing open datasets, and translating discoveries directly into production infrastructure used by governments and enterprises across the Asia-Pacific.

4
Active funded projects
6+
Papers published 2023–25
8
University partners
4
Open datasets released

Research areas

Six interconnected disciplines that together define Airfree's research agenda and differentiate our platform capabilities.

Spatial AI

Developing foundation models for geospatial reasoning, integrating large language models with spatial graph structures to enable natural-language querying of complex geographic datasets and automated feature extraction at continental scale.

Digital Twins

Advancing real-time synchronisation architectures for city-scale digital twins, including physics-informed simulation layers, IoT sensor fusion pipelines, and photorealistic rendering of dynamic urban environments.

Earth Observation Analytics

Building automated change-detection and land-cover classification pipelines over multi-spectral and SAR imagery, with particular focus on bushfire recovery, coastal erosion, and agricultural monitoring across the Indo-Pacific region.

Sovereign Cloud Architectures

Designing air-gapped and sovereignty-preserving infrastructure patterns for sensitive geospatial workloads in government and critical-infrastructure contexts, covering encryption-at-rest, data residency enforcement, and federated identity.

LiDAR Processing

Creating next-generation point-cloud compression, semantic segmentation, and real-time streaming algorithms capable of handling billion-point datasets on commodity hardware, with applications in infrastructure inspection and autonomous navigation.

Satellite Image Fusion

Researching super-resolution and data-fusion techniques that combine optical, hyperspectral, and radar imagery from heterogeneous constellations into unified, analysis-ready data cubes aligned to open OGC standards.

Active research projects

Current externally funded programs conducted in partnership with government agencies and leading research universities.

ARIA — Automated Remote-sensing Intelligence for Agriculture

A three-year program developing AI-driven crop stress detection and yield estimation tools using fused Sentinel-2 and PlanetScope imagery across 40 million hectares of Australian farmland.

Partner institutions
  • University of Queensland
  • CSIRO Agriculture & Food
Funding body

Australian Research Council (ARC) Linkage Grant LP230100412

Expected outcomes
  • —Open benchmark dataset of 500k labelled field parcels
  • —Production-grade inference API integrated into Airfree GIS Platform
  • —Peer-reviewed methodology for near-real-time harvest forecasting

SovereignMesh — Federated Geospatial Infrastructure for Government

Designing a multi-agency, sovereignty-preserving data-sharing architecture that lets federal and state agencies exchange classified spatial datasets without centralising custody or breaching protective marking controls.

Partner institutions
  • Australian National University
  • Defence Science and Technology Group
Funding body

Dept. of Home Affairs — Protective Security Research Initiative 2024

Expected outcomes
  • —Reference architecture published as open standard
  • —Prototype deployed across three pilot agencies
  • —Policy framework for cross-jurisdictional spatial data governance

TwinCity — Real-Time Urban Digital Twin Synchronisation

Investigating low-latency event-streaming architectures that synchronise physical-world sensor readings into a living 3D city model at sub-second cadence, enabling operational dashboards for emergency services and utilities.

Partner institutions
  • UNSW Sydney — School of Built Environment
  • City of Brisbane
Funding body

Smart Cities and Suburbs Program — Round 5, Dept. of Infrastructure

Expected outcomes
  • —Open-source event-broker connectors for SCADA and IoT protocols
  • —Validation study across 12 km² of inner Brisbane
  • —Framework paper on temporal consistency in live digital twins

DeepLiDAR — Foundation Models for Point-Cloud Understanding

Training large-scale transformer models on diverse aerial and terrestrial LiDAR corpora to achieve zero-shot semantic segmentation of infrastructure assets including powerlines, road markings, and building facades.

Partner institutions
  • ANU School of Computing
  • Geoscience Australia
Funding body

National AI Centre — Applied Research Grant 2025

Expected outcomes
  • —Publicly released pre-trained model weights (Creative Commons)
  • —Benchmark suite covering five asset categories
  • —Integration pathway into Airfree Point-Cloud Processing service

Recent publications

Selected peer-reviewed output from the Airfree research team, 2023–2025.

Full bibliography available on request
01

Federated Spatial Data Exchange: A Sovereignty-Preserving Architecture for Multi-Agency Environments

Nnam U., Chen L., Prasad R. · International Journal of Geographic Information Science (2025)

02

Real-Time Sensor Fusion for City-Scale Digital Twins: Latency Bounds and Consistency Guarantees

Walker M., Nnam U., Ito S., Okonkwo A. · ACM SIGSPATIAL Conference on Advances in Geographic Information Systems (2025)

03

Super-Resolution Fusion of Sentinel-2 and SAR Imagery Using Diffusion-Based Priors

Mbeki T., Chen L., Singh P. · IEEE Transactions on Geoscience and Remote Sensing (2024)

04

LiDAR Semantic Segmentation at Billion-Point Scale: Streaming Architectures and Memory Efficiency

Nnam U., Walker M., Okonkwo A. · ISPRS Journal of Photogrammetry and Remote Sensing (2024)

05

Natural-Language Interfaces for Spatial Query: Benchmarking LLM Grounding in GIS Contexts

Prasad R., Ito S., Mbeki T. · Computers, Environment and Urban Systems (2024)

06

Autonomous Change Detection Over Multi-Temporal Satellite Archives: A Continental-Scale Study

Singh P., Chen L., Nnam U. · Remote Sensing of Environment (2023)

Research partnerships

We collaborate with government funding bodies, research universities, and international standards organisations to ensure our work is rigorous, relevant, and reproducible.

Australian Government

  • Australian Research Council (ARC) — Linkage and Discovery programs
  • National AI Centre — Applied Research Grants
  • Geoscience Australia — Collaborative research agreements
  • Dept. of Infrastructure — Smart Cities & Suburbs Program
  • Dept. of Home Affairs — Protective Security Research Initiative

University Collaborations

  • Australian National University — School of Computing & College of Engineering
  • University of Queensland — Remote Sensing Research Centre
  • UNSW Sydney — School of Built Environment & Spatial Sciences
  • University of Melbourne — Infrastructure Engineering
  • Curtin University — Spatial Sciences

International Bodies

  • Open Geospatial Consortium (OGC) — Standards working groups
  • Group on Earth Observations (GEO) — GEOSS data contribution
  • International Society for Photogrammetry and Remote Sensing (ISPRS)
  • ESA Phi-Lab — Copernicus ecosystem collaboration
Open data initiative

Publicly released datasets

Airfree believes that foundational geospatial data should be a shared resource. We release curated, analysis-ready datasets to accelerate the broader research community.

AusUrban-LiDAR-1B

CC BY 4.02.4 TB

One billion classified point-cloud returns across six Australian capital cities, collected at 20 pts/m², with building, vegetation, ground, and infrastructure labels.

AgriBench-500k

CC BY-NC 4.0180 GB

Labelled field-parcel dataset covering 500,000 agricultural parcels across Queensland and New South Wales, with multi-season NDVI time series and yield ground-truth annotations.

CoastalChange-AUS-20yr

CC BY 4.095 GB

Twenty-year shoreline change analysis for the entire Australian coastline derived from Landsat and Sentinel-2 archives, with erosion and accretion polygons at 10-metre resolution.

SovereignMesh-TestBed

Apache 2.012 GB

Synthetic multi-agency geospatial dataset designed for evaluating federated data-sharing protocols, including simulated protective markings, provenance metadata, and access-control scenarios.

PhD & researcher program

Airfree co-sponsors PhD candidates and hosts research interns through structured programs with our university partners. Candidates work on live research problems with direct access to production-scale data infrastructure, publication support, and a pathway to full-time employment upon graduation.

We currently offer placements in Spatial AI, Digital Twins, and Earth Observation Analytics. Stipends are competitive with ARC scholarships and are supplemented by Airfree top-up funding. Candidates retain full intellectual-property rights over their thesis contributions and are co-authors on any resulting publications.

Duration
3 years (PhD) · 3–6 months (intern)
Stipend top-up
Up to $15,000 AUD/year above ARC base
Next intake
February 2027 (applications open Q4 2026)
Research collaboration

Collaborate with our research team

Whether you're a government agency, university researcher, or international standards body, we welcome proposals that advance the science of geospatial intelligence and deliver real-world impact.