Leveraging AI, Computer Vision, and Consumer Drones for Carbon Stock Estimation
restoration sites monitored
tons of carbon stock measured
trees observed and geo-referenced
lbs. of biomass estimated
“This technology makes the impact of the management practices we work to implement at Huntingdon Farm real. Being able to document and see carbon stock on a tree by tree basis with such ease will forever change how our operation can be an asset to the surrounding environment and Emma’s Creek.”
Working in partnership with Huntingdon Farm and Coolant, The Commons orchestrated comprehensive drone scans leveraging AI, Computer Vision, and a consumer drone to assess biomass, carbon stock, and tree count totals for selected management practices at Huntingdon Farm.
Our team outlined all Huntingdon Farm restoration sites to track implementation progress and define spatial boundaries.
In partnership with Coolant, our team developed and carried out flight plans from FieldDoc restoration data for automated drone footage collection.
Our partners at Coolant, ran drone gathered imagery through their AI and Computer Vision processing pipeline generating environmental metrics packaged in a stunning web mapping application.
Throughout the lifecycle of the project, our team grew proficiency amongst Huntingdon Farm staff to execute future flight plans with limited intervention.
Huntingdon Farm is a small grass-fed, certified Angus beef operation located in Huntingdon County, PA. Purchased by John Dawes in 1970 with his brother Donald Dawes and Professor Herman R. Purdy from Penn State University, the farm evolved into a national and international source of Angus genetics. At Huntingdon Farm, the environment is a primary consideration alongside premier Angus genetics. Founder John Dawes Sr. and acting Farm Manager Bill Shoemaker have established the operation as a model for sustainability, producing grass-fed Angus beef that is light on the landscape and harmonized with the natural world. As a result, Huntingdon Farm prioritizes maximum conservation over maximum production.
The farm demonstrates no fewer than 30 best management practices (BMPs), implemented with professional expertise and cost-share programs to align with the farm’s core values. Huntingdon Farm primarily comprises grazing lands with alternating crops of Red Clover, Crimson Clover, Annual Rye Grass, Sudan Grass, and soil rejuvenation plots of turnips and alfalfa. These crops are directly fed to the cows, reducing reliance on costly and carbon-intensive feed supplementation. Additionally, forested land within the farm’s boundaries is managed for habitat preservation and supplemental income. A core focus of the operation is to quantify and assess the impact of reforestation and stream restoration efforts on environmental health, specifically Emma’s Creek—a small tributary to the Juniata River that flows into the Chesapeake Bay. To this end, Huntingdon Farm has partnered with The Commons and Coolant to deploy cutting-edge technology, enabling higher transparency and actionable data for evaluating land and water restoration efforts while maintaining a robust business operation.
Despite billions of dollars invested in mitigating industrialization’s adverse effects on the natural world, governments and private foundations lack the tools to trace the efficacy, location, and persistence of their restoration investments. Without clear traceability, improving future restoration strategies and building public trust remain challenging, ultimately compromising ecosystem health and the sustainability of environmental restoration efforts.
The Chesapeake Bay Watershed Agreement highlights this issue. Federal investments of over $14.1 billion toward water quality improvements have only produced aggregated reports, leaving stakeholders unable to pinpoint project locations or assess their performance and impact. This gap contributes to jurisdictions consistently missing water quality targets, eroding public confidence in restoration funding. Outdated verification tools, high costs, and institutional inertia sustain an unacceptable "black box" approach to tracking restoration efforts. These limitations are further compounded by technical barriers and legacy approaches, preventing sectors from achieving a shared understanding of the implementation status and outcomes of functional restoration projects.
Fortunately, platforms like FieldDoc are driving positive changes by enabling users to document restoration and land protection activities with precision. FieldDoc provides clear records of restoration sites and supports active tracking and documentation of associated work and funding sources. This innovation represents a significant step forward in transparency and accountability. However, practitioners still require cost-effective tools for ongoing monitoring as restoration sites develop. This need is addressed by integrating AI, computer vision, and low-cost commercial drones.
AI, computer vision, and drones are transforming how restoration activities are measured, reported, and verified (MRV). They also provide precise tracking of critical indicators such as carbon stock and other environmental benefits. Traditional methodologies, such as sampling and manual measurements of forest stands, are time-intensive and prone to error. For example, foresters often assess variables like diameter at breast height (DBH), canopy coverage, and biomass on representative samples to estimate carbon stock. These methods introduce biases and assumptions that can result in over- or under-estimations.
Drone-based Light Detection and Ranging (LiDAR) offers unparalleled accuracy and repeatability, eliminating assumptions in data collection. However, the high costs—exceeding $40,000 for drone platforms—and the need for skilled GIS analysts make LiDAR prohibitive for many practitioners.
Coolant, leveraging commercial drones equipped with automated flight software and high-resolution cameras, provides an accessible alternative. These platforms collect high-fidelity images of restoration sites, which Coolant transforms into hyperrealistic 3D models comparable to LiDAR data products. Using pre-programmed drone scans, Coolant measures biomass, carbon stock, and canopy height for every tree in a project, systematically eliminating human error and assumptions.
This technology’s affordability marks a pivotal shift in site assessment. For instance, a DJI Mavic Mini 3 costs just over $1,000, and no specialized software modifications are needed. Coolant prepares automated flight plans using tools like DroneDeploy, ensuring proper altitudes and angles. Practitioners only need to charge the drone, execute the flight plan, and upload the imagery to Coolant for processing. This simplicity democratizes access to advanced MRV tools, saving hundreds of thousands of dollars and empowering local practitioners and landowners to generate credible carbon stock and biomass estimates.
The reusability of flight plans allows for consistent rescanning of sites as trees mature, providing ongoing insights into restoration progress. Coolant’s innovations is currently in a pilot phase, with validation efforts underway through leading carbon and ecosystem service credit auditors. Once validated, this approach will streamline the issuance of carbon and ecosystem service credits, reducing costs by enabling direct collaboration between implementers and corporate partners.
The integration of Coolant and FieldDoc creates a powerful synergy for restoration tracking. For example, nonprofits and landowners collaborating with corporate partners like beef and dairy cooperatives can autonomously conduct site scans and produce accurate carbon stock and biomass estimates. These data are essential for corporate partners developing carbon insetting programs to support greener supply chains. FieldDoc complements Coolant’s capabilities by providing transparent metrics and tracking tools. Together, they allow users to assemble comprehensive portfolios of restoration work with full visibility into implementation status and environmental benefits. Stakeholders can easily share this information with corporate partners, ensuring a clear chain of custody for management practices that actively improve the natural world. By combining systems like Coolant and FieldDoc, all aspects of restoration are documented and measurable, leading to greater confidence and efficiency in generating ecosystem service benefits. This transformative approach enables practitioners to achieve meaningful, scalable impacts while promoting accountability and transparency in environmental restoration.
Huntingdon Farm staff demonstrated a clear alignment and understanding of the key environmental data required for the project. The planning process began by identifying critical data elements to be automated through preplanned drone flights. Canopy height, diameter at breast height (DBH), biomass volume, and carbon stock were determined as the most essential metrics. The scope and scale of the project were carefully aligned to ensure compatibility with Coolant’s existing processing capabilities. At the time of project development, Coolant’s pipeline could accommodate these core data elements, with plans for future enhancements to include algorithms for additional ecosystem services such as species diversity and biodiversity assessments.
Once the core restoration sites were selected, the team collaborated with Huntingdon Farm’s local county conservation district office to obtain spatial data for specific practice footprints. These footprints were uploaded into a FieldDoc portfolio as discrete activities, with detailed attributes such as implementation status and funding sources. After practice details were finalized in FieldDoc, the restoration site geographies were transferred to Coolant for integration into DroneLink. Customized flight plans were then created and loaded for use with DJI’s Mavic Mini 2 drone.
With all critical components in place, the team proceeded to the field to execute scans at each restoration site. This fieldwork successfully collected the necessary data for processing through Coolant’s advanced pipeline, laying the groundwork for detailed environmental analysis and restoration tracking.
The core output of this project consisted of a repeatable process to enable Huntingdon Farm staff to easily execute drone scans for continual MRV and quantification of environmental indicators for Huntingdon Farm restoration activities. Our team also issued a fully featured web application for visualization and use of all data observed by the consumer drone. Project outputs are as follows:
FieldDoc Portfolio Deployment: Our team setup and configured necessary metrics in the FieldDoc platform to track critical metrics on restoration activities across Huntingdon Farm. The system enabled our team to document geospatially the respective restoration site footprint and critical implementation details that transparently demonstrate the management practice’s state of implementation.
Reusable Flight Plans and Training: With restoration site footprints delineated in FieldDoc, our team worked with Coolant to integrate these geospatial data into adequate flight plans using Drone Link. This ensured the DJI Mavic Mini 2 had all of the necessary information to successfully gather data on a given mission including flyover height and inflight camera gimbal adjustments to capture necessary angles of the restoration site. These flight plans are saved as well as packaged with an SOP allowing Huntingdon Farm staff to conduct further automated flights that collect necessary data for processing. Drone footage were uploaded to Coolant for processing.
Web Mapping Application Development: Leveraging Coolant’s AI and Computer Vision pipeline, drone footage was reconstructed a series of algorithms that identified and georeferenced all trees at the restoration site, calculated biomass, as well as carbon stock. Additional metric rollups were delivered such as tree count per site, total biomass per site, and carbon stock per site. The web application supports data access to source images used in the 3d reconstruction as well as tabular data export for each tree’s height, width, area, biomass, and carbon stock. Lastly the web mapping application supports 3D views as well as carbon views that color trees by their respective carbon stock.
The integration of innovative tools such as Coolant and FieldDoc has yielded transformative results for Huntingdon Farm and its collaborators. By adopting advanced monitoring and tracking technologies, Huntingdon Farm has successfully:
By combining Coolant and FieldDoc, Huntingdon Farm has not only demonstrated the effectiveness of innovative restoration tools but also set a new standard for sustainable agricultural practices. These outcomes underscore the potential for technology-driven solutions to revolutionize environmental restoration, fostering a future where transparency, efficiency, and measurable impact are the norms.