01 / FEATURED
Frontier buyers sign first $53M in offtake agreements with Charm Industrial
)
Skip to main content
OUR BLOG
23 Mar 2026 | Brian Jamieson & Garrett Lutz
6 MINUTES READ
)
The transition from paper exercise to real operations is never more clear than when a truckload of logs gets dropped off in your yard. )
When your raw material grows in forests or gets baled in cornfields, and shows up in piles taller than a house, inventory management stops being a spreadsheet exercise and starts being an engineering challenge. Running a diversified supply chain, where your sources are piles of wildfire residue in one location, and bales of corn stover in another, presents a real organizational challenge.
At Charm, our business depends on turning irregular biomass residues into stable carbon that is stored underground. As we scale our carbon removal operations, quantifying those piles early, accurately, and repeatedly isn’t just helpful, it’s foundational.
That’s why we’ve leveraged AI to measure, store, and evolve our biomass inventory using photogrammetry and computer vision - designed for real yards, real terrain, and real-world variability.
Prior to processing (see recent post), our biomass doesn’t arrive in neat containers with barcodes. Log piles arrive in different sizes, from forest thinning and wildfire resilience projects. Corn stover is baled into cylinders or squares then hauled to our site. Both end up stacked on sloped and uneven ground, not on a perfect flat slab. Every pile is different. Geometry varies. Moisture varies. Packing density varies. The ground underneath varies. These all present challenges for accurate measurement of the biomass inventory we receive.
)
Traditional inventory methods often rely on visual estimates, loader counts, or bulk volume approximations layered with assumptions about density and moisture. While truck scales help us know what enters or exits our facility, knowing how those inventories change over time is more of a challenge.
We wanted a system that could:
Accurately estimate dynamic pile volume regardless of ground level
Account for air gaps inside piles (packing factor)
Operate in remote sites with limited to no connectivity
Produce geotagged, repeatable records
Scale across multiple sites
Charm needed a new tool for the job, and we turned to our technology partners on this work: Rebulk. Using a combination of drones, fixed monitoring stations, and cell phone imaging we tested two core ideas:
1. 3D reconstruction of pile geometryUsing drone or mobile imagery, overlapping photos can be processed into a dense 3D model of the pile and surrounding yard via a method called photogrammetry.
The system stitches 2D imagery into a 3D volumetric model of the pile. Using spatial data from the photos, the underlying ground surface beneath the pile can be identified to account for sloped ground. Biomass yards are rarely flat, and small baseline errors can meaningfully distort volume estimates at scale. Modeling the true terrain profile improves consistency and reduces bias and was an important input to achieve reliable results from this pilot program.
)
Exemplary figure illustrating how sloping ground could impact the volumetric measurement if not carefully considered for elongated long piles.
2. Packing factor estimationBulk volume alone doesn’t tell the whole story. Biomass piles contain air gaps between logs, branches, or stalks. Traditional bulk inventory volumetrics are like throwing a blanket over the pile: technically correct regarding the total volume occupied by material, but incorrect without factoring in the air gaps present in the pile structure.
Side-view imagery is analyzed using computer vision to estimate the void fraction within the pile. That produces a packing factor - a measure of how much of the pile’s bulk volume is solid biomass versus empty space.
Bulk volume × packing factor = estimated usable material volume.
)
Exemplary figure only, representing void fraction identification. Rebulk makes it possible to utilize multiple images, so estimates are not limited to a single angle. Instead, the tool captures a probable range of void fraction based on the photogrammetric analysis.
While Rebulk has spent a lot of their time looking at the volumetrics of bulk piles like limestone in quarries, this was a new challenge. Analyzing biomass piles required the tool to be able to convert volume to dry mass. By estimating packing factor (or void space) from the images, a more realistic material volume could be achieved. This allows piles with very different internal structures (including piles of different materials) to be measured in a comparable way. Estimating or measuring moisture content of the material closes the gap between pile volumetrics and dry mass available for processing.
)
Integrating pile volume, packing factor, and moisture considerations means that 2D images of biomass piles can be used to track and predict inventory dynamics, without the need to constantly weigh the material piles. Combined with opportune weight scale tickets, Charm can operate with a reliable end to end inventory for decentralized biomass supplies.
What excites us most isn’t just that this works, it’s that it works in the field.
The workflow is designed to be operator-friendly and usable in remote environments. Charm has built out its distributed network of digital infrastructure allowing this at all sites (see this in detail in this blog post). Scans can be captured using standard mobile devices and synced later, which is critical for sites without reliable connectivity like in mountainous Colorado.
Beyond individual scans, the larger opportunity is building a time-series inventory system:
Tracking how piles grow and deplete
Improving forecasting and logistics planning
Creating structured, traceable data from biomass origin through processing
Over time, this kind of visibility can inform decisions about yard layout, material movement, and operational efficiency - all without adding heavy process overhead to field teams. With geotagged inventories we can estimate transportation requirements of different feedstocks to our facility, and estimate associated MRV embodied emissions impacts for our remote inventory. These types of continuous improvements can drive the changes needed to scale the delivery of high-integrity carbon removals, while reducing costs and expanding market access to carbon removal buyers who seek durable, high integrity carbon removals.
Carbon removal doesn’t start at the wellhead. It starts with feedstock.
If you can’t confidently measure your biomass inventory, you can’t reliably:
Plan pyrolyzer utilization
Coordinate logistics
Forecast material availability
Reduce reconciliation friction downstream
Accurate measurement turns irregular piles into actionable data. Actionable data supports better operations. Better operations support scale.
And scale is the whole point.
Find it interesting? Share!
)
Brian Jamieson
Principal Scientist & TEA Guy
)
Garrett Lutz
Carbon Accounting- MRV
Subscribe to follow our journey to inject bio-oil into deep-geological formations, Charm permanently puts CO2 back underground.
Find it interesting? Share!
At Charm, we’re often asked what stands between us and gigaton-scale carbon removal. There are many academic answers to this: policy frameworks; economic market incentives; technological learning curves. All of these are right, in their own way. But as one of the leading carbon removal companies that is removing carbon today, there’s an often overlooked challenge. What does it mean to scale operations for a new hardware and manufacturing technology? What breaks when you go from running one carbon sucking machine some of the time, to a fleet of machines all the time? In this post, we’re going to take you through the key moments and highlights over the last 12 month journey in scaling durable carbon removals. It’s the gritty, zoomed-in view of technological progress, full of engineering surprises, clog-busting ingenuity, and G.O.A.T. hotflows.
)
Dillon Card
Director of Operations Engineering
)
At Charm, we’re often asked what stands between us and gigaton-scale carbon removal. There are many academic answers to this: policy frameworks; economic market incentives; technological learning curves. All of these are right, in their own way. But as one of the leading carbon removal companies that is removing carbon today, there’s an often overlooked challenge. What does it mean to scale operations for a new hardware and manufacturing technology? What breaks when you go from running one carbon sucking machine some of the time, to a fleet of machines all the time? In this post, we’re going to take you through the key moments and highlights over the last 12 month journey in scaling durable carbon removals. It’s the gritty, zoomed-in view of technological progress, full of engineering surprises, clog-busting ingenuity, and G.O.A.T. hotflows.
Humanity has emitted hundreds of gigatonnes of CO₂. Now you can put it back underground.
)
)
)