Summary
The UAS Wildland Fire Collaborative says its pipeline turned drone uploads into pilot-approved wildfire maps in a median of 18 minutes during the 2026 fire season. The service standardizes thermal imagery into map products for incident teams and uses Lenovo Legion computers for development and local processing.
The UAS Wildland Fire Collaborative says its drone-data pipeline turned uploads into pilot-approved wildfire maps in a median of 18 minutes during the 2026 fire season. The figure, which includes the pilot’s review, is reported by Collaborative founder Eli Polsky in an account published by Lenovo on 3 November 2026.
The nonprofit’s service converts thermal drone imagery into standardized products that incident teams can use in their existing mapping tools. Lenovo Legion computers supported the system’s development and local processing; much of the processing for users now takes place in the cloud.
From thermal images to incident maps
Thermal cameras capture heat patterns that can help pilots identify hotspots and possible heat beyond a fire’s mapped perimeter. Turning hundreds of images into useful information requires more than stitching them together: pilots and agencies may use different software, data formats and incident names. Polsky describes one fire whose name appeared in 34 different versions in the data the system encountered.
After a pilot uploads a flight, the pipeline identifies the files, matches them to the incident and produces topographic and aerial map sheets, Google Earth files and GIS-compatible datasets. It also creates a daily infrared activity log documenting the flight and the pilot’s interpretation. Maps mark detected heat perimeters, individual hotspots, areas of intense heat and possible heat signatures for closer inspection. The pilot reviews and approves the products before they are shared through mapping tools used by incident teams and the Collaborative’s public archive.
The East Evans Creek fire in Oregon offers an example of the system in use. Over three weeks, it produced 37 map sets across 15 nights. The busiest map documented 3,327 individual heat sources. At Woodrat Mountain, a survey map marked five possible heat sources; Polsky later found glowing embers at one location while walking the area with the map. Such signatures can point responders towards places to inspect, while interpretation remains with people on the ground.
As of 21 October 2026, the public archive contained 894 files representing 79 nights of maps across 12 wildfires. These figures describe delivered mapping products, rather than measured changes to firefighting decisions.
Where the Legion computers fit
Polsky used a Lenovo Legion 5 as his main development machine to build the processing pipeline, create agency-specific map templates and test changes against wildfire data. Its NVIDIA graphics card supported work in Agisoft Metashape, which stitches drone images into larger thermal maps, and ArcGIS Pro, used for GIS analysis and map development. The source says cloud processing now handles much of the work for Collaborative users, while local computing remains important for development, the founder’s own flights and large datasets such as LiDAR.
For one October flight, 195 thermal photographs were processed into a map about six minutes after upload to the cloud-based stitching system. That is a single example of the stitching stage; the reported 18-minute median covers the broader interval from upload to final approval.
After the graphics card in the original laptop failed, Lenovo provided a Legion Pro 7i with an Intel Core Ultra 9 processor and NVIDIA GeForce RTX 5070 Ti graphics. Polsky says the replacement will support local mapping workflows and software development. The Collaborative also plans to work on LiDAR and multispectral mapping for wildfire-mitigation projects and make its system easier for additional agencies to adopt.
The practical aim is to make overnight aerial observations available to incident teams before morning briefings in a consistent format. The account reports processing times and map deliveries, but says the organization does not yet have reliable data on firefighter hours saved or on how individual operational decisions have changed.