Ground-Based Monocular Vision System for UAV Tracking : Performance Evaluation in Realistic Simulations

November 28, 2025

Unmanned Aerial Vehicles (UAVs) have become essential across numerous civilian and military applications. Accurate ground-based pose tracking of UAVs is crucial for applications such as defence systems, environmental monitoring, and intelligence gathering. Among existing ground-based pose tracking methods, monocular approaches using the UAV’s 3D CAD model are convenient but challenging due to noise, view ambiguities, non-linear UAV dynamics, et cetera. To improve a popular 3D-model-based pose tracking method, this thesis proposes the use of accurate UAV dynamics in an autonomous landing scenario, i.e., assuming knowledge of UAV commands. We have used the dynamical equations of a small fixed-wing UAV commonly used in intelligence, surveillance, and reconnaissance. To evaluate our approaches, we developed a graphical and physical simulation using Unreal Engine. Our method was tested with seven simulated experiments with different characteristics. We evaluated the pose-tracking method with and without the UAV’s dynamical model.

Authors in the community:

• 

  M

  Miguel Alexandre Marques Gomes Brandão Bandeira

  ist1100035

Supervisors of this institution:

• 

  A

  Alexandre José Malheiro Bernardino

  ist13761

• 

  N

  Nuno Alexandre Antunes Martins Pessanha Santos

  ist157680

electrical-engineering-electronic-engineering-information-engineering - Electrical engineering, electronic engineering, information engineering

eng - English

Instituto Superior Técnico