ADAPT: Drone Payload for Data Collection and Real-Time AI Processing in the Field

Webinar
September 21, 2021

ADAPT: Drone Payload for Data Collection and Real-Time AI Processing in the Field

Join the National Oceanic and Atmospheric Administration (NOAA) for the next NOAA Innovators Seminar on September 21st at noon Eastern! Dr. Matt Brown of Kitware Inc, will discuss the Autonomous Data Acquisition and Processing Technologies (ADAPT) project.

ADAPT is a multi-purpose data collection and real-time processing payload for small unmanned aerial vehicles. While this payload can support a variety of missions by swapping AI models, its initial application is real-time image segmentation of river ice structures to support timely prediction of floods. ADAPT includes a lightweight, state-of-the-art neural network for in-flight processing of RGB (red, green, and blue) imagery to extract boundaries between ground, liquid water, and ice. These geo-registered and vectorized boundaries are transmitted wirelessly to a ground station for real-time inspection for flood prediction.

Matt Brown, PhD, is a principal engineer on Kitware’s Computer Vision Team located in Carrboro, North Carolina. He has over 12 years of experience developing advanced imaging systems and image-exploitation algorithms. His expertise spans from the fundamental physics of imaging to the applied aspects of designing and integrating hardware and software to solve challenging problems.