^ForestFlux

^{^{XR-Based Field Measurement of Tree Biomass and Carbon Storage}}

Duration
2024

Researcher
Hanju Seo

Context
Meta x PwC x Innovate UK XR Hackathon

Role
- Principal Investigator
- System Architect (HCTI framework)
- Researcher (allometric equations, carbon accounting)

Collaborator
- Zumeng Liu
- Bilal Ahmad
- Dzhamal Alanov
- Congee

Most tools used to assess a forest treat the tree as an object to be measured after the fact: felled, sawn, or surveyed from a distance. ForestFlux starts from a different premise. If every standing tree can be treated as a live data node, its biological state should be queryable in the field, in real time, without specialist equipment or destructive contact. The project developed a prototype system on the Meta Quest 3 headset during a weekend hackathon organised by Meta, PwC, and Innovate UK, building toward what the research frames as Human-Computer-Tree Interaction: a framework for integrating sensing hardware, computer vision, and decision interfaces into a single field event.

The workflow is compact by design. The headset camera captures a full-colour image of the target stem; Meta's on-device AI vision model identifies the species from that image without requiring a network connection. The user then marks the base and crown tip of the tree using the handheld controller to derive height, and completes a circular sweep around the trunk to record diameter at breast height. From these three inputs, the application automatically calculates above-ground biomass and carbon storage using species-specific allometric equations drawn from forestry literature. The complete record is generated in under ten seconds and uploaded to a cloud dashboard.

Measurement accuracy averaged 8 to 12 percent error compared with conventional tape-and-clinometer values. The speed, however, was the more significant result: a complete tree record in under ten seconds, operable without training. Short demonstration clips shared online attracted unsolicited responses from arborists who described the approach as viable for preliminary surveys and particularly useful for public engagement programmes where manual tools discourage non-specialist volunteers.

ForestFlux prototype demonstration - species identification, height and DBH measurement, and carbon storage estimation using Meta Quest 3.