Logo do repositório

Percorrer por autor "Filipe, Vítor"

A mostrar 1 - 2 de 2
  • Integrating Computer Vision Object Recognition with Location Based Services for the Blind
    Publication . Fernandes, Hugo; Costa, Paulo; Paredes, Hugo; Filipe, Vítor; Barroso, João
    The task of moving from one place to another is a difficult challenge that involves obstacle avoidance, staying on street walks, finding doors, knowing the current location and keeping on track through the desired path. Nowadays, navigation systems are widely used to find the correct path, or the quickest, between two places. While assistive technology has contributed to the improvement of the quality of life of people with disabilities, people with visual impairment still face enormous limitations in terms of their mobility. In recent years, several approaches have been made to create systems that allow seamless tracking and navigation both in indoor and outdoor environments. However there is still an enormous lack of availability of information that can be used to assist the navigation of users with visual impairments as well as a lack of sufficient precision in terms of the estimation of the user's location. Blavigator is a navigation system designed to help users with visual impairments. In a known location, the use of object recognition algorithms can provide contextual feedback to the user and even serve as a validator to the positioning module and geographic information system of a navigation system for the visually impaired. This paper proposes a method where the use of computer vision algorithms validate the outputs of the positioning system of the Blavigator prototype.
    2014-07-01Comunicação em conferência Acesso restrito Ver mais
  • UAV Landing Using Computer Vision Techniques for Human Detection
    Publication . Safadinho, David; Ramos, João; Ribeiro, Roberto; Filipe, Vítor; Barroso, João; Pereira, António
    The capability of drones to perform autonomous missions has led retail companies to use them for deliveries, saving time and human resources. In these services, the delivery depends on the Global Positioning System (GPS) to define an approximate landing point. However, the landscape can interfere with the satellite signal (e.g., tall buildings), reducing the accuracy of this approach. Changes in the environment can also invalidate the security of a previously defined landing site (e.g., irregular terrain, swimming pool). Therefore, the main goal of this work is to improve the process of goods delivery using drones, focusing on the detection of the potential receiver. We developed a solution that has been improved along its iterative assessment composed of five test scenarios. The built prototype complements the GPS through Computer Vision (CV) algorithms, based on Convolutional Neural Networks (CNN), running in a Raspberry Pi 3 with a Pi NoIR Camera (i.e., No InfraRed-without infrared filter). The experiments were performed with the models Single Shot Detector (SSD) MobileNet-V2, and SSDLite-MobileNet-V2. The best results were obtained in the afternoon, with the SSDLite architecture, for distances and heights between 2.5-10 m, with recalls from 59%-76%. The results confirm that a low computing power and cost-effective system can perform aerial human detection, estimating the landing position without an additional visual marker.
    2020Artigo científico Acesso aberto Ver mais