Resumo em Inglês:

Abstract: Tomatoes are widely cultivated, both by family farmers and corporate producers. During the tomato growth cycle, several diseases can affect the plant. The identification of these diseases through short-range images is significant, and computer vision techniques are commonly used to identify diseases in plant leaves. In this paper, a hybrid model that combines a convolutional neural network (CNN) and a Random Forest (RF) decision tree is used for foliar spot detection in tomato leaves. High-level features learned and extracted from CNN are used as input for the RF classifier. To evaluate the proposed model’s performance for plant disease identification, a case study of 2480 low-cost digital RGB images collected in actual field conditions, under different intensities of light exposure, were used, including healthy tomato leaves and leaves with visible symptoms of powdery mildew fungus, which attacks the tomato leaf. The results were compared with six conventional machine learning classifiers: Logistic Regression (LR), Linear Discriminant Analysis (LDA), K- Nearest Neighbors (KNN), Naive Bayes (NB), Support Vector Machine (SVM) and Random Forest (RF). The results show that the proposed model outperformed conventional classifiers, reaching an accuracy of 98%. The results highlight the importance of fusing models to improve the detection plant´s diseases.

Resumo em Inglês:

Abstract: Positional accuracy of cartographic products is typically evaluated using positional discrepancies and point-based techniques. However, using linear features has some advantages over the point-based method, such as a greater amount of geometric and positional information and the fact that approximately 80% of the features on a cartographic basis are lines. Despite these advantages, important parameters for evaluating accuracy using lines have not yet been established or determined, such as the spatial distribution pattern, although it is a relevant factor that can affect the results and determine the validity of an evaluation process. This study proposes a method based on the modification of the Nearest Neighbor Method for points, which can be used to evaluate the spatial distribution pattern of linear features. Instead of the traditional Euclidean distance used by the method for points, the method proposes using the Hausdorff Distance as a measure of the spacing between lines. The proposed method, called Nearest Neighbor Method for Linear Features (NNMLF), was applied to simulated and real data. All experiments with simulated data showed that the NNMLF was effective in estimating spatial distribution pattern up to the third order. Its use on real data showed NNMLF is simple to apply.

Resumo em Inglês:

Abstract: The Story Map Of Evandro Case - Development And Creation Of An Interactive Cartographic Narrative Story maps allow us to present new perspectives on stories, providing a broader understanding of the events and places involved. This approach allows users to navigate time and space, connecting emotionally with the narrative. This paper presents an innovative approach to designing and developing interactive story maps, drawing on agile methodologies widely used in software development and adapting them for map projects. The study examines the process of creating an interactive Story Map for the ‘Evandro Case’, a famous criminal case in Brazil, as a central narrative theme. The methodology employed combines traditional cartographic principles with modern storytelling techniques. This approach enhances maps’ informational value and establishes a more profound emotional resonance with users. The paper highlights the importance of user immersion, advocating for future studies to include user testing and statistical analysis to validate usability and effectiveness. Through the case study of the Evandro Case, the paper demonstrates how story maps can transcend conventional map design, offering fresh insights into narratives. It argues that story maps are a versatile tool applicable to various themes beyond criminal cases. The study concludes that the fusion of spatial representation with narrative elements can create compelling and informative visual narratives, making complex stories accessible and engaging to a wide audience.

Resumo em Inglês:

Abstract: Ionospheric scintillation affects GNSS signals that provide many essential services, making the monitoring of scintillation an important issue. This work presents a system for the real-time acquisition, generation and online dissemination of the S4 scintillation index maps covering Brazil using data from two major networks of GNSS monitoring stations, LISN and INCT. The maps are made using an innovative pre-processing and interpolation scheme. The system is already implemented and tested, being composed of a single real-time server with a database and modules that perform reception from the GNSS stations, data processing, and the online dissemination. All these tasks are executed asynchronously in a pipeline manner using the database as a central hub without any loss of data. The challenges that must be overcome to have real-time capability were: (i) to configure GNSS stations to send S4 data to a real-time server able to (ii) receive S4 data from the GNSS stations, (iii) generate sequences of S4 scintillation maps, and (iv) make these maps available in a web server. The implemented system was able to acquire data from all available stations of the monitoring networks, being robust concerning interruption of connections or different processing times of the tasks.

Resumo em Inglês:

Abstract: This study investigates the influence of geodetic network configuration, stochastic model, and the approach local or global on the determination of minimum detectable displacements (MDD) using sensitivity analyses and significance tests. The proposed approach integrates sensitivity characteristics to establish confidence regions based on MDD. In addition, we examine the equality between the critical value of a significance test and the non-centrality parameter derived from a chi-square distribution to compute concentric ellipsoids representing sensitivity and accuracy. The analyses were focused on evaluate how variations in network configuration, stochastic model, and the type of analysis (if global or local) affect the relationship between sensitivity and accuracy. Our results showed the importance of considering these factors, providing valuable insights for robust network design and analysis in practical applications.

Resumo em Inglês:

Abstract: The Brazilian Pantanal biome, known for its rich biodiversity and wetlands, is experiencing frequent and destructive fires. Detecting and monitoring burnt areas is vital for comprehending their present ecological condition, a key indicator for climate change and protective measures. Optical remote sensing methods, traditionally used in fire mapping, have limitations due to atmospheric conditions. Microwave Synthetic Aperture Radar (SAR) is a promising alternative, excelling in challenging environments and demonstrating sensitivity to surface properties. This study aimed to assesses the potential of SAR images for detecting burnt areas in a conservation unit inserted in the Brazilian Pantanal after intense fires in 2020. For this, the Normalized Burn Ratio (NBR) index was calculated from Sentinel-2 images before and after fire, and then the difference between these images (dNBR). Differences in backscatter coefficients of pre- and post-fire SAR/Sentinel-1 images in the two polarizations (dVH and dVV) were also calculated. To detect burnt areas, the three differences were classified using the Random Forest algorithm. The results showed adequate coincidence of burned areas between dVH and dVV compared to dNBR and high accuracy values of the algorithm model, indicating consistency between SAR and optical data in identifying burnt areas.

Resumo em Inglês:

Abstract: The evolution of Global Navigation Satellite System (GNSS) positioning has greatly benefited several areas of knowledge. For Hydrography, an application improved by this science is the measurement of sea level oscillations resulting from tides. However, to satisfactorily retrieve this information, it is necessary to use low-pass filters (LPF) to match high frequency signals resulting from variation of the vertical component of the GNSS positioning to those of low frequency that characterizes tidal waves. Currently, there is a wide variety of LPF, which are selected according to the required purpose. Thus, the objective of this study is to obtain tidal height variations with high accuracy by applying LPF in GNSS positioning vertical coordinates tracked by an onboard GNSS receiver. For this purpose, field research and the processing of obtained data was performed. Then, two data filters were tested: the Simple Moving Average (SMA) Filter and wavelet compression. In both options, the results reached centimetric accuracy when compared to the real tide in the region of study. However, through quantitative and qualitative evaluations, it was verified that the SMA filter was considered more advantageous because, in addition to its high accuracy, it has a simpler application and less expensive in computational terms.