Skip to content

ADAPT Multi-Mission Payload

An open source platform for deploying state of the art deep-neural-network computer vision in real time on small unmanned aircraft systems (sUAS).

Payload CAD Rendering sUAS

  • Optimized drone-based collection of imagery and geospatial metadata with live feedback to maintain quality control.

  • Integration with the open source do-it-yourself AI toolkit VIAME to annotate data and train mission-specific image-processing models.

  • Upload your models for aerial deployment with real-time, georegistered analytics wirelessly transmitted to a ground station computer and beyond for rapid dissemination.

  • Commodity hardware components, CAD models, and open-source software allows organizations to cheaply and easily build their own payloads

Supports a variety of unique missions

Ongoing work on the ADAPT project is funded by NOAA to support key missions.

Explore

Try the simulator with docker

Events

Papers / Presentations

  • (2022) National Innovation Center Seminar: Slides

  • (2022) Ocean Sciences Meeting: Slides

  • (2022) 1st International Workshop on Practical Deep Learning in the Wild at AAAI Conference on Artificial Intelligence: Paper

  • (2021) NOAA Innovators Series: Slides, Recording

  • (2021) The 3rd NOAA Workshop on Leveraging AI in Environmental Sciences: Slides, Recording

sUAS sUAS