How to Drone Map a Disaster Zone in 60 minutes or Less

This past November, I coordinated a short experiment on the ground in Paradise, California in the final hours of wrapping up damage assessment from the Camp Fire.

The goal of this experiment was to test more rapid methods for visualizing damage to critical infrastructure after a disaster. Members of the UAS teams from Alameda and Contra Costa County Sheriff’s Office carried out the fieldwork with some help from volunteers (thanks Andrew Maximow of Firmatek on the ground and John Cherbini for blazing fast internet). Here is a brief summary of the results and the full report can be download for free at:

The Question –  How swiftly can we map all major streets in a disaster zone using georeferenced drone video?

The Methods – Three teams were assigned to all major streets in the town of Paradise (three  main thoroughfares, four main crossroads). Each team consisted of a drone pilot to capture video (Mavic or Phantom) and a driver. Drivers were responsible for relocating pilots further down each road segment and to assist in safety operations. Video and drone flight logs were then uploaded to a cloud software called Survae ( for georeferencing and visualization.

Macintosh HD:Users:gregorycrutsinger:Desktop:RapidResponseDroneData.jpg

The Results – All major streets were surveyed in approximately 60 minutes after deployment. In total, 21.7 linear miles (35 km) of video was captured resulting in 17 GB of data. Videos were visualized in the Survae software as a split-screen view, synchronizing the imagery with the flight map. Property boundaries from Butte County records were also overlaid on the base map to improve damage assessment. The cumulative effort of this study was approximately 2.5 hours, with final products delivered to coordinating agencies by the following day and made public.

Discussion: In the aftermath of the Camp Fire, we found that georeferenced drone video and a coordinated capture effort to be effective tools for visualizing the main transportation infrastructure of the town of Paradise. The survey was carried out at a fraction of the time, by fewer teams and with a significantly lower data  compared to traditional drone mapping efforts that had just concluded. Moreover, manual flight and video capture are simple, require no additional training, and easily repeated.

Drone video is not new and has been used in many other disasters and emergencies. However, without spatial context for when and where video was collected, its use in disaster coordination efforts remains limited. While georeferenced video will not replace the need for more traditional aerial mapping, its positioning within the first few hours of an emergency is ideal for providing situational awareness, damage assessment, and precise targeting of more labor intensive drone mapping efforts in the following days.

Finally, the use of micro-experiments during such events can lead to useful advances in response methodologies and real-world vetting of solutions.

The full free report includes suggested improvements to the methods and future directions:

Thanks goes to Survae for donating software support for this work.