Crowd Spotting

This Call is a part of the Providence Search & Rescue UAV Project.

Searching for lost persons at sea can be a daunting task. Looking from above is often helpful. With the introduction of drones, Search & Rescue organizations will have new ways to stream aerial video from a search pattern. What if we could enlist the public to help looking at the video streams via social media? (The Swedish Sea Rescue Society, for example, has more than 50k Facebook followers).

Here’s how it might work:

• A drone is remotely launched in response to a missing person at sea (See Providence)
• The drone pilot sets up a search pattern, follows the progress and scans the video stream
• A ping for help goes out over social media
• As helpers enter the Crowd Spotting site the video feed is dynamically split up into smaller frames and time shifted to make sure that every frame gets viewed by as many spotters as possible

A Crowd Spotting system might be further enhanced in various ways:

• Combine human spotters with learning machine vision
• Research the optimum attention span of spotters in different conditions and time shift the video feed accordingly
• Simplifying the interface to the point where anyone can instantly help with no prior training

Inspiration/Research

Planethunters.org

Tomnod.com

Earth Watchers

Micromappers

Sentient ViDAR - Visual Detection and Ranging – which might be good enough or an excellent compliment to crowd spotting