Cookies
O website necessita de alguns cookies e outros recursos semelhantes para funcionar. Caso o permita, o INESC TEC irá utilizar cookies para recolher dados sobre as suas visitas, contribuindo, assim, para estatísticas agregadas que permitem melhorar o nosso serviço. Ver mais
Aceitar Rejeitar
  • Menu
Sobre

Sobre

Professor Auxiliar com Agregação da Universidade de Trás-os-Montes e Alto Douro (UTAD) e doutorado em Ciências da Engenharia Geográfica, pela Universidade do Porto e pela Universidade de Delft (Holanda), tendo apresenta a tese “Potential of integrating PSInSAR Methodologies in the Detection of Surface Deformation”. Atualmente, é Investigador (membro integrado) do Centre for Robotics in Industry and Intelligent Systems (CRISS), do INESC TEC/Polo UTAD, e investigador (colaborador) do CITAB (Centre for the Research and Technology of Agro-Environmental and Biological Sciences). Nos últimos anos tem-se dedicado, sobretudo, à utilização de Veículos Aéreos Não Tripulados (UAV) para aplicações agroflorestais. Utiliza imagens aéreas de elevada resolução, obtidas por diferentes sensores (RGB, NIR, Multiespectrais, Hiperespectrais e Térmicos) para, usando técnicas de processamento de imagem e desenvolvimento de algoritmos, extrair informações e parâmetros relevantes, sobretudo, na vinha, soutos e olivais. Estas técnicas são, no entanto, extensíveis à deteção e monitorização de grande parte das espécies arbóreas, que integram as nossas florestas, e de vegetação rasteira. É autor de várias publicações em revistas internacionais da especialidade do Remote Sensing. Participa em vários projetos de investigação, destacando-se o PARRA (Plataforma integrAda de monitoRização e avaliação da doença da flavescência douRada na vinha), em que é líder por parte da UTAD (SI I&DT, aviso Nº 08/SI/2015, Projeto em Co-Promoção, parceiros do projeto: TEKEVER ASDS - empresa líder, UTAD, Instituto Politécnico de Viana do Castelo, INIAV, Agrociência. Montante total atribuído 1.602.245,58€) e é membro do projeto Plataforma de Inovação da Vinha e do Vinho, linha Remote sensing and detection of grapevine diseases (Projeto I&DT pelo Norte2020, com um financiamento global de 4.500.000,00 €).

Tópicos
de interesse
Detalhes

Detalhes

  • Nome

    Joaquim João Sousa
  • Cargo

    Investigador Sénior
  • Desde

    01 janeiro 2014
005
Publicações

2024

Comparative Evaluation of Remote Sensing Platforms for Almond Yield Prediction

Autores
Guimaraes, N; Fraga, H; Sousa, JJ; Pádua, L; Bento, A; Couto, P;

Publicação
AGRIENGINEERING

Abstract
Almonds are becoming a central element in the gastronomic and food industry worldwide. Over the last few years, almond production has increased globally. Portugal has become the third most important producer in Europe, where this increasing trend is particularly evident. However, the susceptibility of almond trees to changing climatic conditions presents substantial risks, encompassing yield reduction and quality deterioration. Hence, yield forecasts become crucial for mitigating potential losses and aiding decisionmakers within the agri-food sector. Recent technological advancements and new data analysis techniques have led to the development of more suitable methods to model crop yields. Herein, an innovative approach to predict almond yields in the Tras-os-Montes region of Portugal was developed, by using machine learning regression models (i.e., the random forest regressor, XGBRegressor, gradient boosting regressor, bagging regressor, and AdaBoost regressor), coupled with remote sensing data obtained from different satellite platforms. Satellite data from both proprietary and free platforms at different spatial resolutions were used as features in the study (i.e., the GSMP: 11.13 km, Terra: 1 km, Landsat 8: 30 m, Sentinel-2: 10 m, and PlanetScope: 3 m). The best possible combination of features was analyzed and hyperparameter tuning was applied to enhance the prediction accuracy. Our results suggest that high-resolution data (PlanetScope) combined with irrigation information, vegetation indices, and climate data significantly improves almond yield prediction. The XGBRegressor model performed best when using PlanetScope data, reaching a coefficient of determination (R2) of 0.80. However, alternative options using freely available data with lower spatial resolution, such as GSMaP and Terra MODIS LST, also showed satisfactory performance (R2 = 0.68). This study highlights the potential of integrating machine learning models and remote sensing data for accurate crop yield prediction, providing valuable insights for informed decision support in the almond sector, contributing to the resilience and sustainability of this crop in the face of evolving climate dynamics.

2024

Synthetic Aperture Radar in Vineyard Monitoring: Examples, Demonstrations, and Future Perspectives

Autores
Bakon, M; Teixeira, AC; Padua, L; Morais, R; Papco, J; Kubica, L; Rovnak, M; Perissin, D; Sousa, JJ;

Publicação
REMOTE SENSING

Abstract
Synthetic aperture radar (SAR) technology has emerged as a pivotal tool in viticulture, offering unique capabilities for various applications. This study provides a comprehensive overview of the current state-of-the-art applications of SAR in viticulture, highlighting its significance in addressing key challenges and enhancing viticultural practices. The historical evolution and motivations behind SAR technology are also provided, along with a demonstration of its applications within viticulture, showcasing its effectiveness in various aspects of vineyard management, including delineating vineyard boundaries, assessing grapevine health, and optimizing irrigation strategies. Furthermore, future perspectives and trends in SAR applications in viticulture are discussed, including advancements in SAR technology, integration with other remote sensing techniques, and the potential for enhanced data analytics and decision support systems. Through this article, a comprehensive understanding of the role of SAR in viticulture is provided, along with inspiration for future research endeavors in this rapidly evolving field, contributing to the sustainable development and optimization of vineyard management practices.

2024

Classification of Grapevine Varieties Using UAV Hyperspectral Imaging

Autores
López, A; Ogayar, CJ; Feito, FR; Sousa, JJ;

Publicação
REMOTE SENSING

Abstract
Classifying grapevine varieties is crucial in precision viticulture, as it allows for accurate estimation of vineyard row growth for different varieties and ensures authenticity in the wine industry. This task can be performed with time-consuming destructive methods, including data collection and analysis in the laboratory. In contrast, unmanned aerial vehicles (UAVs) offer a markedly more efficient and less restrictive method for gathering hyperspectral data, even though they may yield data with higher levels of noise. Therefore, the first task is the processing of these data to correct and downsample large amounts of data. In addition, the hyperspectral signatures of grape varieties are very similar. In this study, we propose the use of a convolutional neural network (CNN) to classify seventeen different varieties of red and white grape cultivars. Instead of classifying individual samples, our approach involves processing samples alongside their surrounding neighborhood for enhanced accuracy. The extraction of spatial and spectral features is addressed with (1) a spatial attention layer and (2) inception blocks. The pipeline goes from data preparation to dataset elaboration, finishing with the training phase. The fitted model is evaluated in terms of response time, accuracy and data separability and is compared with other state-of-the-art CNNs for classifying hyperspectral data. Our network was proven to be much more lightweight by using a limited number of input bands (40) and a reduced number of trainable weights (560 k parameters). Hence, it reduced training time (1 h on average) over the collected hyperspectral dataset. In contrast, other state-of-the-art research requires large networks with several million parameters that require hours to be trained. Despite this, the evaluated metrics showed much better results for our network (approximately 99% overall accuracy), in comparison with previous works barely achieving 81% OA over UAV imagery. This notable OA was similarly observed over satellite data. These results demonstrate the efficiency and robustness of our proposed method across different hyperspectral data sources.

2024

Phase Unwrapping using ML methods

Autores
Couto, D; Davies, S; Sousa, J; Cunha, A;

Publicação
Procedia Computer Science

Abstract
Interferometric Synthetic Aperture Radar (InSAR) revolutionizes surface study by measuring precise ground surface changes. Phase unwrapping, a key challenge in InSAR, involves removing ambiguity in measured phase. Deep learning algorithms like Generative Adversarial Networks (GANs) offer a potential solution for simplifying the unwrapping process. This work evaluates GANs for InSAR phase unwrapping, replacing SNAPHU with GANs. GANs achieve significantly faster processing times (2.38 interferograms per minute compared to SNAPHU's 0.78 interferograms per minute) with minimal quality degradation. A comparison of SBAS results shows that approximately 84% of GANs points are within 3 millimeters of SNAPHU. These results represent a significant advancement in phase unwrapping methods. While this experiment does not declare a definitive winner, it demonstrates that GANs are a viable alternative in certain scenarios and may replace SNAPHU as the preferred unwrapping method. © 2024 The Author(s). Published by Elsevier B.V.

2024

Automatic classification of abandonment in Douro's vineyard parcels

Autores
Teixeira, I; Sousa, J; Cunha, A;

Publicação
Procedia Computer Science

Abstract
Port wine plays a crucial role in the Douro region in Portugal, providing significant economic support and international recognition. The efficient and sustainable management of the wine sector is of utmost importance, which includes the verification of abandoned vineyard plots in the region, covering an area of approximately 250,000 hectares. The manual analysis of aerial images for this purpose is a laborious and resource-intensive task. However, several artificial intelligence (AI) methods are available to assist in this process. This paper presents the development of AI models, specifically deep learning models, for the automatic detection of abandoned vineyards using aerial images. A private image database was expanded, containing a larger collection of images with both abandoned and non-abandoned vineyards. Multiple AI algorithms, including Convolutional Neural Networks (CNNs) and Vision Transformers (ViTs), were explored for classification. The results, particularly with the ViTs approach, achieved high accuracy and demonstrated the effectiveness of automatic detection, with the ViT models achieving an accuracy of 99.37% and an F1-score of 98.92%. The proposed AI models provide valuable tools for monitoring and decision-making related to vineyard abandonment. © 2024 The Author(s). Published by Elsevier B.V.

Teses
supervisionadas

2023

Automatic detection of abandoned vineyards using aerial imagery

Autor
Igor Filipe Ribeiro Teixeira

Instituição
UTAD

2023

Phase unwrapping using ml methods

Autor
Diogo Gabriel da Silva Couto

Instituição
UTAD

2021

Hyperspectral data analysis for agriculture applications

Autor
Jonás Hruska

Instituição
UTAD

2021

Irrigation management in olive groves with support of geomatics

Autor
Pedro Miguel Mota Marques

Instituição
UTAD

2021

Aerial high-resolution imagery to assess almond orchard conditions

Autor
Nathalie dos Santos Guimarães

Instituição
UTAD