Cookies Policy
The website need some cookies and similar means to function. If you permit us, we will use those means to collect data on your visits for aggregated statistics to improve our service. Find out More
Accept Reject
  • Menu
About

About

PhD in Electrical and Computer Engineering from the University of Porto (Faculdade de Engenharia), and MSc in Electrical and Electronic Engineering - Autonomous Systems branch from ISEP Porto Polytechnic School. Since 2016 she has been a researcher at CRAS – INESC TEC, where she works with hyperspectral cameras, mainly in classification algorithms using deep learning techniques and unsupervised approaches. The main topic of her PhD thesis was to detect and classify marine litter using different machine learning and deep learning methods, which culminated in implementing the first zero-shot learning algorithm with hyperspectral imaging data. She also made part of the team that collected the datasets used to implement these algorithms in a real hotspot, using manned and unmanned aircraft. At the moment, she is working as an assistant researcher In the NewSpace Project (PT-PRR), in the development of novel AI methods for data compression onboard satellites. 

Interest
Topics
Details

Details

  • Name

    Sara Costa Freitas
  • Role

    Assistant Researcher
  • Since

    11th January 2016
001
Publications

2022

Hyperspectral Imaging Zero-Shot Learning for Remote Marine Litter Detection and Classification

Authors
Freitas, S; Silva, H; Silva, E;

Publication
REMOTE SENSING

Abstract
This paper addresses the development of a novel zero-shot learning method for remote marine litter hyperspectral imaging data classification. The work consisted of using an airborne acquired marine litter hyperspectral imaging dataset that contains data about different plastic targets and other materials and assessing the viability of detecting and classifying plastic materials without knowing their exact spectral response in an unsupervised manner. The classification of the marine litter samples was divided into known and unknown classes, i.e., classes that were hidden from the dataset during the training phase. The obtained results show a marine litter automated detection for all the classes, including (in the worst case of an unknown class) a precision rate over 56% and an overall accuracy of 98.71%.

2021

Remote Hyperspectral Imaging Acquisition and Characterization for Marine Litter Detection

Authors
Freitas, S; Silva, H; Silva, E;

Publication
REMOTE SENSING

Abstract
This paper addresses the development of a remote hyperspectral imaging system for detection and characterization of marine litter concentrations in an oceanic environment. The work performed in this paper is the following: (i) an in-situ characterization was conducted in an outdoor laboratory environment with the hyperspectral imaging system to obtain the spatial and spectral response of a batch of marine litter samples; (ii) a real dataset hyperspectral image acquisition was performed using manned and unmanned aerial platforms, of artificial targets composed of the material analyzed in the laboratory; (iii) comparison of the results (spatial and spectral response) obtained in laboratory conditions with the remote observation data acquired during the dataset flights; (iv) implementation of two different supervised machine learning methods, namely Random Forest (RF) and Support Vector Machines (SVM), for marine litter artificial target detection based on previous training. Obtained results show a marine litter automated detection capability with a 70-80% precision rate of detection in all three targets, compared to ground-truth pixels, as well as recall rates over 50%.

2021

Hyperspectral Imaging System for Marine Litter Detection

Authors
Freitas, S; Silva, H; Almeida, C; Viegas, D; Amaral, A; Santos, T; Dias, A; Jorge, PAS; Pham, CK; Moutinho, J; Silva, E;

Publication
OCEANS 2021: SAN DIEGO - PORTO

Abstract
This work addresses the use of hyperspectral imaging systems for remote detection of marine litter concentrations in oceanic environments. The work consisted on mounting an off-the-shelf hyperspectral imaging system (400-2500 nm) in two aerial platforms: manned and unmanned, and performing data acquisition to develop AI methods capable of detecting marine litter concentrations at the water surface. We performed the campaigns at Porto Pim Bay, Fail Island, Azores, resorting to artificial targets built using marine litter samples. During this work, we also developed a Convolutional Neural Network (CNN-3D), using spatial and spectral information to evaluate deep learning methods to detect marine litter in an automated manner. Results show over 84% overall accuracy (OA) in the detection and classification of the different types of marine litter samples present in the artificial targets.

2019

Convolutional neural network target detection in hyperspectral imaging for maritime surveillance

Authors
Freitas, S; Silva, H; Almeida, JM; Silva, E;

Publication
INTERNATIONAL JOURNAL OF ADVANCED ROBOTIC SYSTEMS

Abstract
This work addresses a hyperspectral imaging system for maritime surveillance using unmanned aerial vehicles. The objective was to detect the presence of vessels using purely spatial and spectral hyperspectral information. To accomplish this objective, we implemented a novel 3-D convolutional neural network approach and compared against two implementations of other state-of-the-art methods: spectral angle mapper and hyperspectral derivative anomaly detection. The hyperspectral imaging system was developed during the SUNNY project, and the methods were tested using data collected during the project final demonstration, in Sao Jacinto Air Force Base, Aveiro (Portugal). The obtained results show that a 3-D CNN is able to improve the recall value, depending on the class, by an interval between 27% minimum, to a maximum of over 40%, when compared to spectral angle mapper and hyperspectral derivative anomaly detection approaches. Proving that 3-D CNN deep learning techniques that combine spectral and spatial information can be used to improve the detection of targets classification accuracy in hyperspectral imaging unmanned aerial vehicles maritime surveillance applications.

2018

Supervised Classification for Hyperspectral Imaging in UAV Maritime Target Detection

Authors
Freitas, S; Almeida, C; Silva, H; Almeida, J; Silva, E;

Publication
2018 IEEE INTERNATIONAL CONFERENCE ON AUTONOMOUS ROBOT SYSTEMS AND COMPETITIONS (ICARSC)

Abstract
This paper addresses the use of a hyperspectral image system to detect vessels in maritime operational scenarios. The developed hyperspectral imaging classification methods are based on supervised approaches and allow to detect the presence of vessels using real hyperspectral data. We implemented two different methods for comparison purposes: SVM and SAM. The SVM method, which can be considered one of most utilized methods for image classification, was implemented using linear, RBF, sigmoid and polynomial kernels with PCA for dimensionality reduction, and compared with SAM using a two classes definition, namely vessel and water. The obtained results using real data collected from a UAV allow to conclude that the SVM approach is suitable for detecting the vessel presence in the water with a precision and recall rates favorable when compared to SAM.