Skip to content

Case File2020–2022

UAV Visual Tracking & Localization

NYU Abu Dhabi · CAIR

Multi-paper research thread on detecting, tracking, and relatively localizing UAVs in real time — from airborne PTZ visual lock and cooperative spherical localization through RGB-thermal fusion and deep-learning / Siamese aerial trackers, validated on live flights in the Kinesis arena.

System Record

Application
APPCounter-UAS
Period
2020–2022

My Contribution

System integration and experimental development across the CAIR UAV perception thread — arena instrumentation, RGB-T PTZ pipelines, and field validation that underpins six peer-reviewed papers (2020–2022).

Development & Validation

  • Thread opens with airborne PTZ visual tracking and relative visual localization for cooperative UAS, then layers computationally efficient RGB-thermal detection so thermal cues pull small drones out of clutter while RGB refines boxes at frame rate.
  • Deep-learning evader pursuit and a Siamese adaptive transformer tracker extend the same arena stack to agile targets; relative spherical-visual localization closes the loop for multi-UAV cooperative localization.
  • Flagship RGB-T detection and tracking demo integrated and flight-tested inside NYUAD's netted Kinesis arena with pan-tilt-zoom camera coverage — evidence spans six peer-reviewed outputs plus the live arena video.

Credits & Collaborators

  • Anthony Tzes

    NYU Abu Dhabi · CAIR

    Principal Investigator

  • Athanasios Tsoukalas

    NYU Abu Dhabi · CAIR

    Co-author

  • Nikolaos Evangeliou

    NYU Abu Dhabi · CAIR

    Co-author

  • Dengqing Xing

    NYU Abu Dhabi · CAIR

    Co-author

  • Scott Holter

    NYU Abu Dhabi · CAIR

    Co-author

Evidence

Connected Work