FDFernanda Beatriz Jordan Rojas Dallaqua

Fernanda Beatriz Jordan Rojas Dallaqua

*This is an algorithmically generated author record based on publicly available information.

Fernanda Beatriz Jordan Rojas Dallaqua has a total of 8 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    CBERS 4A/WPM Image Classification for Mapping Land Use and Land Cover in TCRAs

    by Carla Rodrigues Santos, Francisco Salazar, Fernanda Beatriz Jordan Rojas Dallaqua, Bruno Schultz

    Tropical restoration areas present high spectral heterogeneity and spatial fragmentation, making land use and land cover (LULC) classification challenging, particularly when using medium-resolution satellite imagery. Although spectral mixture analysis, object-based image analysis, and machine learning approaches have been widely explored individually, their integrated application using freely available China–Brazil Earth Resources Satellite 4A Wide-field Panchromatic and Multispectral (CBERS-4A/WPM) imagery for monitoring Environmental Restoration Commitment Agreements (TCRAs) remains limited. This study investigated whether the integration of Linear Spectral Mixture Model (LSMM) fractions, Geographic Object-Based Image Analysis (GEOBIA), and machine learning algorithms could improve LULC classification for mapping TCRAs in the state of São Paulo, Brazil. Spectral, textural, and sub-pixel attributes derived from pan-sharpened CBERS-4A/WPM imagery were combined into a unified feature dataset, and Random Forest, XGBoost, Multilayer Perceptron, and Support Vector Machine classifiers were evaluated. Random Forest achieved the best classification performance, with an overall accuracy of approximately 0.66, demonstrating superior predictive capability compared with the other tested algorithms. Feature importance analysis indicated that LSMM-derived vegetation and shadow fractions contributed significantly to class discrimination. The integration of spectral, spatial, and machine learning approaches provides an interpretable and cost-effective framework for LULC mapping in highly fragmented tropical restoration landscapes, supporting environmental monitoring and restoration assessment.

  • Open Access

    Land Use and Land Cover Mapping in Fragmented Areas of São Paulo: Application of SITS and LSMM in TCRAs

    by Carla Rodrigues Santos, Bruno Schultz, Fernanda Beatriz Jordan Rojas Dallaqua, Ana Larissa Ribeiro de Freitas, Júlio Bandeira Guerra, Francisco Salazar

    The state of São Paulo is home to remnants of the Atlantic Forest and Cerrado biomes, both of which face intense anthropogenic pressure and hight fragmentation. In this context, Environmental Recovery Commitment Agreements (TCRAs) serve as essential instruments for restoring degraded areas and monitoring vegetation recovery over time. This study assesses land use and land cover (LULC) classification performance in TCRA sites by integrating Satellite Image Time Series (SITS) with spectral fractions derived from the Linear Spectral Mixture Model (LSMM), utilizing 2025 Sentinel-2A/2B imagery. Data were organized into spatiotemporal cubes within R environment and classified using the Random Forest algorithm. Model performance was assessed using a confusion matrix and accuracy metrics, including User’s Accuracy (UA), Producer’s Accuracy (PA), F1-score, and Intersection over Union (IoU), as well as spatial analysis of agreement and disagreement between predicted maps and reference data. Results demonstrate high classification precision for vegetation classes, specifically pasture (F1 = 0.91) and forest formations (F1 = 0.87). Primary misclassifications occurred between spectrally similar classes, particularly within small fragments and intermediate regeneration stages. Overall, the integration of SITS and LSMM enhanced class separability by incorporating temporal dynamics and mitigating spectral mixing effects, highlighting its potential as an operational tool for environmental restoration monitoring.

  • Open Access

    Cluster Analysis for Different Physiognomies and Spatiotemporal Patterns from Vegetation Indices in São Paulo State

    by Francisco Javier Tipan Salazar, Carla Rodrigues Santos, Fernanda Beatriz Jordan Rojas Dallaqua, Bruno Schultz

    Multi-temporal orbital satellite imagery is an alternative for measuring behavioral patterns or trends in different physiognomies through vegetation indices (VIs) and Spectral Linear Mixture Models (SLMMs). In this study, time series of Landsat 7/8/9 and Sentinel-2 have been used to classify a considerable quantity of areas spread over the São Paulo state from 2021 to 2024. Because the large amount of samples considered in our analysis, self-organizing maps (SOMs) have been applied as a convenient method to group similar satellite image time series samples with respect to a certain vegetation index or green vegetation fraction (VEG). Since every dataset area belongs to different types of physiognomies, each cluster has been labeled according to the plurality technique. Additionally, we obtained the mean spectral behavior of the VIs and VEG in the 2021–2024 seasonal cycle of all samples. The results showed similar variations from the rainy to the dry season for most of the physiognomies. On the other hand, this research indicates that the proposed method for classification the Brazilian areas spread over the São Paulo state is consistently good, obtaining the best performance (quantization error) associated with Normalized Difference Vegetation Index (NDVI) time series samples.

  • Open Access

    ForestEyes Project: Conception, Enhancements, and Challenges

    by Fernanda B. J. R. Dallaqua, Álvaro Luiz Fazenda, Fabio A. Faria

    Rainforests play an important role in the global ecosystem. However, significant regions of them are facing deforestation and degradation due to several reasons. Diverse government and private initiatives were created to monitor and alert for deforestation increases from remote sensing images, using different ways to deal with the notable amount of generated data. Citizen Science projects can also be used to reach the same goal. Citizen Science consists of scientific research involving nonprofessional volunteers for analyzing, collecting data, and using their computational resources to outcome advancements in science and to increase the public's understanding of problems in specific knowledge areas such as astronomy, chemistry, mathematics, and physics. In this sense, this work presents a Citizen Science project called ForestEyes, which uses volunteer's answers through the analysis and classification of remote sensing images to monitor deforestation regions in rainforests. To evaluate the quality of those answers, different campaigns/workflows were launched using remote sensing images from Brazilian Legal Amazon and their results were compared to an official groundtruth from the Amazon Deforestation Monitoring Project PRODES. In this work, the first two workflows that enclose the State of Rond\^onia in the years 2013 and 2016 received more than $35,000$ answers from $383$ volunteers in the $2,050$ created tasks in only two and a half weeks after their launch. For the other four workflows, even enclosing the same area (Rond\^onia) and different setups (e.g., image segmentation method, image resolution, and detection target), they received $51,035$ volunteers' answers gathered from $281$ volunteers in $3,358$ tasks. In the performed experiments...

  • Open Access

    Cidadãos Comuns Ajudando no Combate ao Desmatamento das Florestas Tropicais

    by Fernanda B. J. R. Dallaqua, Álvaro L. Fazenda, Fabio A. Faria

    A área de Ciência Aberta, que vem ganhando cada vez mais espaço, almeja a abertura e o compartilhamento de todo o processo científico, promovendo o acesso tanto às publicações como também aos dados utilizados e incentivando novas formas de produção de conhecimento. Cidadãos comuns podem ser utilizados para a coleta, análise e classificação de dados na solução de problemas técnicos e científicos, o que é conhecido como Ciência Cidadã outra área emergente devido a grande quantidade de dados gerados, que são de boa qualidade, e a criação de uma estreita conexão entre cientistas e o público em geral, que se torna mais consciente e engajado na soluções de desafios da sociedade, como por exemplo o desmatamento das florestas tropicais. O projeto de Ciência Cidadã ForestEyes tem como objetivo utilizar as contribuições dos voluntários como conjunto de treinamento de um sistema inteligente que irá reconhecer desmatamento em novas áreas de florestas tropicais. Esse projeto recebeu mais de 86.000 contribuições de 644 voluntários e o sistema inteligente construído a partir dessas respostas conseguiu mais de 80% de acurácia média, mostrando potencial para se tornar um produto auxiliar no monitoramento de desmatamento. Open Science is an emergent area promoting free and easy access to both publications and data, which also encourages new forms of knowledge production. In this context, ordinary citizens can collect, analyze and classify data necessary to solve technical and scientific problems, which is known as Citizen Science. This also emerging subject can generate a large amount of data with good quality, creating a close connection between scientists and the general public, which increases the engagement and awareness in solving society's challenges, such as the deforestation of tropical forests. ForestEyes is a Citizen Science project that aims to use volunteer contributions to create a training set to an intelligent system that will recognize deforestation in new areas of tropical forests. This project received more than 86,000 contributions from 644 volunteers, and the intelligent system built from these data achieved more than 80% average accuracy, showing ForestEyes' potential to become an auxiliary product in deforestation monitoring.

  • Building Data Sets for Rainforest Deforestation Detection Through a Citizen Science Project

    by Fernanda Beatriz Jordan Rojas Dallaqua, Fabio Augusto Faria, Alvaro Luiz Fazenda

    Originally, the ForestEyes project aims to detect deforestation in tropical forests based on citizen science (CS) and machine learning (ML) approaches, in which the volunteers analyze and label segments of remote sensing images to build new training sets for creating different classification models. In previous work, only three modules related to CS have been proposed. In this letter, two new modules are created: 1) organization and selection and 2) ML. Therefore, these modules turn the ForestEyes project a more robust system in the deforestation detection task, building high-confidence labeled collections, increasing the monitoring coverage, and decreasing volunteer dependence. Performed experiments show that volunteers create better data sets than those based on automatic PRODES-based approaches, selecting the most relevant samples and discarding noisy segments that might disrupt ML techniques. Finally, the results showed the feasibility of allying CS with ML for rainforest deforestation detection task.