A carregar...
Projeto de investigação
Sem título
Financiador
Autores
Publicações
Monitoring forest loss in Central Portugal with Sentinel-2 data [poster]
Publication . Moraes, Daniel; Barbosa, Bruno; Costa, Hugo; Moreira, Francisco D.; Benevides, Pedro; Caetano, Mário; Campagnolo, Manuel Lameiras; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)
Background • Land cover monitoring offers significant societal benefits, being essential for environmental modeling and resource management • Advances in satellite data availability, computing storage and processing power introduced a new land cover monitoring paradigm: continuous and rapid identification of changes • Advanced techniques such as the Continuous Change Detection and Classification (CCDC) algorithm also play an important role • Lack of studies using CCDC with Sentinel-2 data ‒ Dense time series (5-day revisit time) ‒ Good spatial detail (10 m) Objective • Evaluate the potential of combining CCDC with Sentinel-2 data to detect changes associated with forest loss caused by clearing or fire
A Weakly Supervised and Self-Supervised Learning Approach for Semantic Segmentation of Land Cover in Satellite Images with National Forest Inventory Data
Publication . Moraes, Daniel; Campagnolo, Manuel L.; Caetano, Mário; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS); Molecular Diversity Preservation International (MDPI)
National Forest Inventories (NFIs) provide valuable land cover (LC) information but often lack spatial continuity and an adequate update frequency. Satellite-based remote sensing offers a viable alternative, employing machine learning to extract thematic data. State-of-the-art methods such as convolutional neural networks rely on fully pixel-level annotated images, which are difficult to obtain. Although reference LC datasets have been widely used to derive annotations, NFIs consist of point-based data, providing only sparse annotations. Weakly supervised and self-supervised learning approaches help address this issue by reducing dependence on fully annotated images and leveraging unlabeled data. However, their potential for large-scale LC mapping needs further investigation. This study explored the use of NFI data with deep learning and weakly supervised and self-supervised methods. Using Sentinel-2 images and the Portuguese NFI, which covers other LC types beyond forest, as sparse labels, we performed weakly supervised semantic segmentation with a convolutional neural network to create an updated and spatially continuous national LC map. Additionally, we investigated the potential of self-supervised learning by pretraining a masked autoencoder on 65,000 Sentinel-2 image chips and then fine-tuning the model with NFI-derived sparse labels. The weakly supervised baseline achieved a validation accuracy of 69.60%, surpassing Random Forest (67.90%). The self-supervised model achieved 71.29%, performing on par with the baseline using half the training data. The results demonstrated that integrating both learning approaches enabled successful countrywide LC mapping with limited training data.
Training data in satellite image classification for land cover mapping
Publication . Moraes, Daniel; Campagnolo, Manuel Lameiras; Caetano, Mário; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS); Taylor & Francis
The current land cover (LC) mapping paradigm relies on automatic satellite imagery classification, predominantly through supervised methods, which depend on training data to calibrate classification algorithms. Hence, training data have a critical influence on classification accuracy. Although research on specific aspects of training data in the LC classification context exists, a study that organizes and synthetizes the multiplicity of aspects and findings of these researches is needed. In this article, we review the training data used for LC classification of satellite imagery. A protocol of identification and selection of relevant documents was followed, resulting in 114 peer-reviewed studies included. Main research topics were identified and documents were characterized according to their contribution to each topic, which allowed uncovering subtopics and categories and synthetizing the main findings regarding different aspects of the training dataset. The analysis found four research topics, namely construction of the training dataset, sample quality, sampling design and advanced learning techniques. Subtopics included sample collection method, sample cleaning procedures, sample size, sampling method, class balance and distribution, among others. A summary of the main findings and approaches provided an overview of the research in this area, which may serve as a starting point for new LC mapping initiatives.
Continuous forest loss monitoring in a dynamic landscape of Central Portugal with Sentinel-2 data
Publication . Moraes, Daniel; Barbosa, Bruno; Costa, Hugo; Moreira, Francisco D.; Benevides, Pedro; Caetano, Mário; Campagnolo, Manuel; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS); Elsevier BV
Recent advances in satellite data availability, computing storage and processing power introduced a new land cover monitoring paradigm, settled on a continuous and timely identification of changes. The Continuous Change Detection and Classification (CCDC) algorithm has emerged as a powerful tool for continuous monitoring, being noteworthy for its ability to process high temporal frequency satellite data with components of seasonality, trend and break. Studies using CCDC were mostly limited to Landsat data, which offer lower spatial and temporal resolution in comparison to Sentinel-2 data. Therefore, our study aims to explore the potential of CCDC with Sentinel-2 data. For that purpose, an extensive reference dataset was developed for change detection accuracy assessment, comprising 290 sites of 200 m radius in a disturbance prone region in Central Portugal, ensuring an adequate representation of areas of vegetation loss. We focused on two specific forest species from this region, eucalyptus and maritime pine. Change date was determined through interpretation of orthophotos and satellite time series. We explored determinant aspects to CCDC performance, namely cloud and cloud shadow masking, algorithm parameterization, use of distinct vegetation indices and detection timeliness. Optimal accuracy was achieved with s2cloudless masking, lambda of 200, chi-square of 0.999, minYears of 1 and the Normalized Difference Vegetation Index. We computed the time lag vs omission error curve, showing comparable results (omission error rate close to 20 % was obtained with a time lag from 30 to 40 days) to methods designed to achieve near-real-time detection. Detections were spatially coherent, with patches of vegetation loss detected only with minor errors, mostly located in polygon borders. Disturbances in the first months resulted in poor model fitting, which undermined detection performance in some cases. Overall, results demonstrated how CCDC and Sentinel-2 data can be used to successfully monitor vegetation loss in a timely manner, especially as the satellite’s time series grows.
Monitoring forest loss in Central Portugal with Sentinel-2 data [abstract]
Publication . Moraes, Daniel; Barbosa, Bruno; Costa, Hugo; Moreira, Francisco D.; Benevides, Pedro; Caetano, Mário; Campagnolo, Manuel Lameiras; Information Management Research Center (MagIC) - NOVA Information Management School; NOVA Information Management School (NOVA IMS)
Avanços na disponibilização de dados de satélites, capacidade de processamento e armazenamento tem contribuído para uma monitorização contínua e ágil da ocupação do solo. O algoritmo Continuous Change Detection and Classifiation(CCDC) decompõe séries temporais de imagens em sazonalidade, tendência e quebras, facilitando a identificação de alterações. Este estudo explorou o uso do CCDC com dados Sentinel-2. Com dados de referência de 290 sítios no Centro de Portugal, o estudo focou na deteção de alterações na floresta de eucalípto e pinheiro-bravo. As datas das alterações foram determinadas por intepretação visual de ortofotos e imagens Sentinel-2. Foram avaliados fatores como máscaras de nuvens, parametrização do algoritmo e índices de vegetação, assim como potencial para deteções em tempo quasereal. Configurações optimas obtiveram um F1-score de 82.07%, com erros de omissão e omissão de 15.04% e 20.63%. Resultados demostraram como o CCDC e dados Sentinel-2 podem ser usados para monitorizar perda de vegetação com bom detalhe espacial e de forma expedita. --- Advances in satellite data availability, computing power and storage have enabled continuous and timely land cover monitoring. The Continuous Change Detection and Classification (CCDC) algorithm decomposes satellite image time series into seasonality, trend, and break, aiding change identification. This study explored using CCDC with Sentinel-2 data. Using a reference dataset of 290 sites in Central Portugal, the study focused on detecting changes in eucalyptus and maritime pine forests. Change dates were determined by visual interpretation of orthophotos and Sentinel-2 images. Factors affecting CCDC performance, including cloud masking, algorithm parameters and vegetation índices were assessed. Potential for near real-time detection was also evaluated. Optimal settings achieved an F1-score of 82.07%, with omission and commission errors of 15.04% and 20.63%. Results demonstrated how CCDC and Sentinel-2 can be used to monitor vegetation loss with good spatial detail in a timely manner.
Unidades organizacionais
Descrição
Palavras-chave
Contribuidores
Financiadores
Entidade financiadora
Fundação para a Ciência e a Tecnologia
Programa de financiamento
OE
Número da atribuição
PRT/BD/153517/2021