A laptop‑driven solution for a burning problem
When a wildfire rages through remote terrain, the most valuable intelligence often comes from a drone hovering above the flames, its thermal camera spotting heat that ground crews can’t see. The bottleneck, however, has traditionally been getting that data from the pilot’s memory card into a format that incident commanders can use before the next briefing.
Eli Polsky, a veteran wild‑land firefighter and founder of the nonprofit UAS Wildland Fire Collaborative, has built a cloud‑enabled pipeline that turns raw thermal images into standardised GIS maps in a matter of minutes. The engine behind the system is a pair of Lenovo Legion laptops – the Legion 5 that Polsky used for years to write the software, and the newly supplied Legion Pro 7i that brings an Intel Core Ultra 9 CPU and an NVIDIA GeForce RTX 5070 Ti GPU to the table.
"Technology has the greatest value when it helps people solve real problems," – Mark Hubbard, Lenovo senior product manager, NA Gaming.
From raw frames to actionable maps
A typical night flight can generate hundreds of thermal photographs. In an October 2026 sortie over West Fork Evans Creek, 195 images were uploaded to the Collaborative’s cloud‑stitching service and, thanks to the Legion‑powered workflow, a complete thermal map was produced in about six minutes. The pipeline then:
- Identifies the correct incident from a database of fire names (the system can reconcile dozens of naming variations, a problem Polsky encountered when a single fire appeared under 34 different titles).
- Stitches the images into a georeferenced orthomosaic using Agisoft Metashape, a process that leans heavily on the GPU.
- Runs GIS analysis in ArcGIS Pro to outline the heat perimeter, flag individual hot spots, and flag ambiguous signatures for on‑ground verification.
- Exports a suite of files – topographic sheets, Google‑Earth KMZ, and GIS‑compatible layers – that slot straight into the tools used by fire agencies.
- Logs the flight and the pilot’s interpretation in a daily infrared activity report.
The final step is a quick pilot review; once approved, the map is instantly available through the agencies’ existing incident‑mapping platforms and a public archive.
Speed that matters on the fireline
During the 2026 fire season, the Collaborative recorded a median turnaround of 18 minutes from data upload to approved map. That means a drone flown at 3 a.m. can have its findings in the hands of firefighters before the 8 a.m. briefing. In the East Evans Creek fire (Jackson County, Oregon), the system produced 37 map sets over 15 nights, with the busiest night documenting 3,327 individual heat sources – information that helped crews adjust containment lines and plan mop‑up operations.
A smaller but illustrative case occurred on August 19 at Woodrat Mountain. The thermal survey flagged only five possible hot spots; a pilot later found glowing embers at two of those locations, confirming that even faint signatures can be critical when the fire is nearly out.
Why the Legion hardware matters
Polsky’s original development machine, a Lenovo Legion 5, proved essential for two reasons:
- GPU horsepower – The NVIDIA graphics card accelerated both Metashape stitching and ArcGIS Pro rendering, turning what could be an hour‑long task into a few‑minute operation.
- Reliability – When the original GPU failed, opening a large ArcGIS project jumped from a few seconds to two and a half minutes, highlighting how dependent the workflow is on a capable graphics processor.
Lenovo’s response was to provide a Legion Pro 7i equipped with the latest Intel Core Ultra 9 and RTX 5070 Ti. The upgrade restores local processing speed for Polsky’s own flights, speeds up software testing, and gives the Collaborative headroom to add new tools such as high‑resolution thermal viewing and LiDAR point‑cloud analysis.
From a single laptop to a shared resource
As of 21 October 2026 the public archive hosted 894 files covering 79 nights of mapping across 12 wildfires. Partners now include the Oregon Department of Forestry (ODF), its South Cascade District, Hummingbird Drones, and Polsky’s own Rogue Reconnaissance. ODF’s input has shaped the visual language of the maps – drone‑derived heat appears in pink, while data from crewed aircraft shows in red – making it instantly clear where each piece of information originates.
The Collaborative is careful not to over‑state impact; while it can count maps delivered and fires supported, it does not yet have hard numbers on firefighter‑hour savings. Nonetheless, the growing adoption suggests the system is filling a genuine operational gap.
Looking ahead: more pilots, more data, same speed
Polsky’s longer‑term vision is simple: any qualified wildfire‑drone pilot should be able to finish a night flight, upload the data, review an automatically generated map, and have it ready for the morning briefing – all without learning a new file format. To achieve that, the Collaborative plans to:
- Expand the library of agency‑specific map templates, each pre‑tested in desktop GIS before being baked into the automated pipeline.
- Add LiDAR and multispectral capabilities for pre‑fire vegetation health assessments, which generate massive point‑cloud datasets that again rely on high‑end GPU processing.
- Refine the cloud‑stitching service to handle larger image batches while keeping latency under ten minutes.
Lenovo’s continued backing – both in hardware and in‑kind support – appears to be the catalyst that will keep the system scaling. For a tech‑focused audience that follows Android and PC hardware trends, the story illustrates how gaming‑grade laptops are finding life‑saving applications beyond gaming, and why a powerful GPU can be as essential in a fire‑zone command centre as it is in a virtual‑reality headset.
The information in this article is based on a Lenovo newsroom release dated 3 November 2026 and statements from the UAS Wildland Fire Collaborative.