[U5] - ForestFlux

• 

• 

• 

• 

  Using AI to Identify Tree Species

• 

  Using AI to Identify Tree Species

• 

  Setting the ground level

• 

  Setting the tree top

• 

  Setting the tree radius

• 

  Calculating the carbon stock of trees

• 

  Testing

• 

  Testing

• 

  Testing

• 

  Our team

Inspiration

ForestFlux was conceived in response to significant inefficiencies in traditional forest management, particularly in carbon assessment. Statistically, 40% of U.S. forests are owned by family forest owners, yet fewer than 0.1% participate in carbon offset markets. This discrepancy is largely due to the high costs, extensive labor, and time-consuming nature of current forest assessment methods. For instance, traditional assessments often require manual data collection over vast areas, involving physical tree measurements that are both time-intensive and prone to human error. Recognizing these challenges, we envisioned ForestFlux, a system that utilizes Mixed Reality and conversational AI to streamline and democratize the carbon assessment process, making it accessible and practical for all forest owners.

What it does

ForestFlux is a transformative Mixed Reality system designed to simplify forest carbon assessment. By integrating conversational AI, it allows users to instantly identify tree species and accurately calculate carbon storage. Users simply engage in dialogue with the system, which then guides them through a seamless process of data collection and analysis. This MR technology provides visualizations of carbon data on-site, allowing for immediate decision-making and ongoing forest management improvements.

How we built it

We developed ForestFlux over three days at a hackathon, utilizing cutting-edge technologies across several platforms. Our team used Unity to build the MR interface and C# for developing backend AI algorithms, with Meta's image capture AI kit powering the visual analysis. Our interdisciplinary team comprised of experts in software engineering, AI, MR technology, and climate tech collaborated to integrate these technologies into a cohesive system. Through iterative testing and real-time feedback, we refined the user experience and functionality to meet the specific needs of forest carbon assessment.

To measure biomass accurately, ForestFlux utilizes the camera on the MR device to first identify the tree species using AI. Once the species is determined, the system applies a specific biomass constant, which is predetermined based on extensive ecological research. Users then use the MR controller to pinpoint and click on the base and the top end of the tree, enabling the system to calculate the tree's height. Subsequently, the diameter of the tree is measured by the controller at breast height, a common method in forestry to ensure measurement consistency. Combining these measurements, ForestFlux calculates the biomass by applying the appropriate mathematical models. From this biomass value, the stored carbon amount is then estimated, providing immediate and accurate carbon storage data essential for carbon trading and forest management.

Challenges we ran into

Integrating diverse technologies into a single MR system was our greatest challenge. Ensuring the AI accurately recognized various tree species and calculated carbon storage correctly under different forest conditions required extensive algorithm optimization. Additionally, creating an intuitive user interface that could operate effectively in outdoor, varied lighting conditions presented significant design and technical hurdles.

Accomplishments that we're proud of

We are exceptionally proud of how ForestFlux turned conceptual ideas into a working prototype within a limited timeframe. Achieving functionality that allows for tree species identification and carbon calculation in under ten seconds per tree represents a breakthrough in forest management technology. This innovation not only enhances efficiency but also increases the accuracy and reliability of data, potentially transforming how forest resources are managed globally.

What we learned

Our journey with ForestFlux was incredibly enlightening, offering us deep insights into the complexities of forest ecology and the transformative potential of MR and AI in environmental technology. We learned about the critical need for user-friendly interfaces in environmental applications and gained valuable experience in integrating hardware with software for robust outdoor applications.

What's next for ForestFlux

Moving forward, ForestFlux will enter a phase of expansion and refinement. Over the next year, we plan to enhance the AI’s capability to include more diverse tree species and improve the MR interface for even smoother user interactions. Pilot programs are scheduled with forest management bodies in various ecosystems to adapt and scale the technology. We are also initiating partnerships with environmental agencies to ensure ForestFlux aligns with global sustainability goals. Ultimately, we aim to see ForestFlux adopted as a standard tool in global carbon markets and conservation strategies.

Built With

• ai
• aikit
• csharp
• gpt
• unity