Name Computer Vision for the Operation of Unmanned Aerial Vehicles in Maritime and WildfIre Scenarios
Funding Reference 031172
Dates 2019-02|2022-01

This project will develop novel methods for target detection and tracking in airborne and seaborne image sequences. The availability of low-cost visual sensors (visible and IR) and the recent developments on convolutional neural-networks and correlation filters are promising cost-effective and energy-efficient solutions for detection and tracking of targets capable of running onboard on air or sea vehicles. The results of the project can have a direct impact on strategic national activities such as ocean and forest monitoring. To demonstrate the applicability of the methods, three case scenarios will be considered: (i) detection of forest fires from airborne images, (ii) detection and tracking of sea vessels from airborne and seaborne images, and (iii) detection, tracking and pose estimation of aircrafts from seaborne images for teleguidance of unmanned aerial vehicles (UAV’s). These demonstrations will inform the responsible institutions for the protection of the environment and coastal operations of a novel technology to support their activities.

Research Groups Computer and Robot Vision Lab (VisLab)
Project Partners Centro de Investigação da Força Aérea (CIAFA), Centro de Investigação Naval da Marinha (CINAV)
ISR/IST Responsible
Alexandre Bernardino
Ricardo Ribeiro
Jorge S. Marques
José Santos-Victor