Post-Wildfire Damage Assessment With Drones: What the Data Actually Looks Like on the Ground in Oreg
The Bootleg Fire burned 413,000 acres across Klamath and Lake Counties in the summer of 2021. The Holiday Farm Fire before that — 173,000 acres up the McKenzie River corridor, thirty miles east of Eugene. When those fires went cold and the containment lines held, the work didn't stop. It shifted. Emergency managers, insurance adjusters, timber companies, county assessors, and private landowners all needed the same thing at roughly the same time: accurate, systematic documentation of what was left and what was gone.
That documentation problem is where drone operators either step up with a real methodology or show up with a consumer quad and a prayer. This is what post-wildfire assessment actually looks like when it's done right.
Why Wildfire Assessment Is Harder Than It Looks From the Air
A burned hillside looks uniform from altitude. That's the deception. From 200 feet up with a wide camera, you see gray ash, standing dead timber, and blackened soil — a monochrome landscape that seems straightforward to document. The reality at ground level is dramatically more complex, and the drone work has to account for that complexity.
The Terrain Problem
Oregon's fire country — the southern Cascades, the Siskiyous, the eastern slope drainages — is not flat. It's steep, broken terrain with ridge fingers, drainage cuts, and road networks that were often compromised by the fire itself. A drone operator who doesn't know this terrain is going to produce coverage gaps on north-facing slopes, miss structure foundations in steep draws, and undercount damage in areas where smoke haze and ash disturb color-based analysis.
The DJI Matrice 30T handles this environment because it was built for it. Forty-one minutes of flight endurance per battery. Wind resistance rated to 15 m/s — relevant when you're flying exposed ridgelines above 4,000 feet with afternoon thermal activity. The 48MP zoom camera and 12MP wide camera together give you the layered resolution you need: wide passes for area coverage, zoom passes for structure-level documentation. The laser rangefinder handles accurate elevation data in terrain where GPS altitude alone isn't sufficient for precise mapping.
The Hazard Window
Post-fire access is not indefinite. Insurance companies and government agencies want assessments fast — often within days of containment, before weather events redistribute ash into waterways, before salvage logging operations begin, and before secondary burn risk returns with the next dry spell. That window matters operationally. You need to be ready to deploy on short notice, fly extended operations with battery redundancy, and deliver usable data the same day.
This is why 30 batteries across a fleet matters in practice. An assessment that covers 600 acres in a single operational day requires planned battery rotation, not ad-hoc charging. Each M30T battery run gives you roughly 35 to 41 minutes of productive flight time depending on altitude and payload. Covering that acreage with the overlap percentage needed for photogrammetric processing — typically 75 to 80 percent front and side overlap — means you're cycling batteries continuously through a full operational day.
What the Thermal Camera Adds That RGB Mapping Can't Provide
RGB photogrammetry — your standard mapping pass with the wide or zoom camera — tells you where structures are standing, where they're destroyed, where vegetation burn severity changed from light surface burn to full crown fire. It produces orthomosaics and 3D point clouds that document the landscape as it exists after the fire.
Thermal imaging adds a layer that RGB can't: it shows you what's still hot.
Hotspot Detection
Containment doesn't mean extinguishment. Oregon fires regularly produce subsurface combustion in root systems, duff layers, and large fuel accumulations that can persist for days or weeks after the fire line is declared controlled. These subsurface hotspots are invisible to RGB cameras and to ground crews walking the area — a cold-looking ash surface can sit on top of a root mass that is actively burning at 400 degrees Fahrenheit three feet underground.
The M30T's 640x512 radiometric thermal sensor identifies these anomalies from operational altitude. A systematic thermal grid pass over a contained area — the same methodical coverage pattern used in search and rescue operations — produces a map of residual heat sources that ground crews can use to prioritize suppression resources. For a 200-acre assessment block, that pass takes less than two hours with proper battery rotation. The alternative is sending crews in on foot to probe and test, which is slower, more dangerous, and produces no spatial record.
Structure Integrity Under Ash Load
When a structure burns and the roof collapses, the thermal signature of the remaining foundation and debris pile tells you whether combustion is complete or ongoing. Insurance documentation that includes thermal data alongside RGB imagery is more defensible, more complete, and more useful to adjusters and engineers who need to determine total loss versus partial loss before any cleanup begins.
The Data Workflow: From Raw Flight to Deliverable Report
A drone assessment that produces a folder of JPEG images is not a deliverable. It's a starting point. What clients — county emergency managers, timber company foresters, private landowners filing insurance claims — actually need is processed, georeferenced, actionable data.
Orthomosaic and Point Cloud Processing
Raw imagery from a mapping pass gets processed through photogrammetry software — Pix4D, DroneDeploy, or Agisoft Metashape — to produce an orthomosaic (a geometrically corrected aerial map) and a dense point cloud (a 3D model of the terrain). For a post-fire assessment, the orthomosaic becomes the base map on which everything else is annotated: structure damage polygons, vegetation burn severity zones, road condition notes, water source locations.
Point cloud data lets engineers calculate debris volumes — relevant when a structure has collapsed and the owner needs to know how much material removal a remediation contractor is looking at before cleanup bids go out.
Burn Severity Classification
Using the orthomosaic, burn severity can be classified into standard categories: unburned, low, moderate, high, and very high. This classification follows established protocols — similar to what the USDA Forest Service uses in Burned Area Emergency Response (BAER) assessments — and gives land managers a spatial picture of where erosion risk is concentrated, where revegetation efforts are most urgent, and where timber salvage is viable versus where it isn't.
For a private timber operation that lost a section of production forest, knowing the spatial distribution of burn severity isn't an academic exercise. It directly informs salvage harvest planning, replanting budgets, and the timeline for returning the land to productive use.
Report Format That Actually Gets Used
A georeferenced PDF report with annotated screenshots, coordinate-referenced damage locations, and a burn severity map overlay is the format that county assessors, insurance carriers, and agency personnel can actually use. Raw drone footage without annotation is not a deliverable — it's homework you're assigning to the client.
Every assessment I produce includes: flight logs with GPS track and altitude data, timestamped imagery with coordinate metadata, an annotated orthomosaic with damage classification, and a written summary of findings. That package is defensible in an insurance dispute, usable by an engineer, and archivable for future comparison if a subsequent fire affects the same area.
The EyesOn Layer: Real-Time Situational Awareness During Assessment Operations
Post-fire assessments don't always happen in a vacuum. Sometimes they happen while suppression crews are still working the edges of a containment zone. Sometimes a county emergency manager twenty miles away needs to see what the drone is seeing in real time — not a summary report hours later, but eyes on the ground as the flight is happening.
EyesOn exists for exactly that scenario. Sub-200ms WebRTC latency on a self-hosted server means the video feed from the M30T — wide camera, zoom camera, or thermal — reaches the incident commander's screen in near real-time. No cloud routing. No third-party infrastructure sitting between the drone feed and the viewer. The companion Android app captures the full DJI controller screen including OSD data: altitude, distance, battery state, gimbal position. The viewer watching the stream gets operator context, not just video.
For an agency that needs to make resource allocation decisions while the assessment flight is in progress — send the ground crew to the draw on the north slope because the thermal pass just showed active hotspots — that latency difference is not a technical footnote. It's the difference between acting on current information and acting on information that's already stale.
EyesOn Personal runs $149 setup and $39 per month. That buys a self-hosted server, unlimited drones, unlimited viewers, and a software stack that keeps running whether or not the subscription is current. No per-drone fees. No per-viewer-minute billing. No data leaving your infrastructure. For an agency doing post-fire assessment work across a season that might involve dozens of flights and multiple stakeholders watching feeds simultaneously, the economics are straightforward.
What to Ask Before You Hire a Drone Operator for Post-Fire Work
Not every drone operator who can fly a mapping grid in a flat agricultural field can execute a post-fire assessment in broken Oregon terrain. The questions worth asking before you commit:
**What aircraft are you flying, and what's your battery capacity for an all-day operation?** A consumer drone with two batteries and a 27-minute flight time is not adequate for a 400-acre assessment.
**Do you carry thermal imaging, or just RGB?** Hotspot detection without thermal is not hotspot detection.
**What deliverables will I receive, and in what format?** If the answer is "a link to your video files," that's not sufficient for insurance documentation or agency reporting.
**Are you Part 107 certified and carrying commercial liability insurance?** In burned areas, you are often operating near active hazard zones. This is not the application for unlicensed recreational flight.
**Can you deliver a georeferenced orthomosaic and a written assessment, not just raw imagery?** The difference between raw imagery and a processed deliverable is the difference between a contractor who is done when the flight ends and one who is done when the client has something they can use.
Post-wildfire Oregon is not a forgiving environment for improvised operations. The terrain is demanding, the hazard window is short, and the stakeholders — landowners, insurers, agencies, timber companies — need data they can act on. That's what this work actually requires.
← Back to all posts