Real-time object detection can greatly help in automatic ship recognition. Most often though, image-based ship classifiers do not take into account all the class-specific geometric information available in the perspective projection performed by coastal surveillance cameras. This paper introduces therefore a novel Bayesian filter to jointly track the vessel kinematic states and classify it by fusing the labeled bounding boxes detected on incoming frames originating from multiple fully calibrated cameras. Simulation results show that the proposed recursive filter is able to improve the overall classification accuracy compared to a single-frame ship classifier. Finally, we were able to track and consistently classify a real-world ship using coastal surveillance cameras surrounding the Guanabara Bay at Rio de Janeiro.

The Automatic Identification System (AIS) is crucial for maritime navigation and monitoring, but it is vulnerable to data manipulation caused by equipment malfunctions, deliberate tampering, or external attacks. This study focuses on detecting vessels that change their identifiers to evade monitoring systems, with a proposed solution with two steps. The first step is to predict vessel trajectories and the second step is to classify de predicted trajectories. We employ the TrAISformer neural network architecture, adapting it to predict future and past trajectories using multiple AIS datasets with varying sampling intervals and regional characteristics. For the classification step, we use Random Forest and SVM algorithm. The datasets include AIS data collected from the regions near Denmark, the Port of Santos, and the Gulf of Mexico, with 2 to 10 minutes of sampling intervals. In this work, compared to the previous work, an in-depth study was carried out on the results obtained in the trajectory prediction stage, showing that higher sampling frequency improves prediction accuracy, while regional characteristics influence model performance over extended prediction periods. These findings laid the foundation for the subsequent classification step, performed in this work, to match pairs of trajectories and identify continuous vessel paths despite altered identifiers. The classification step was directly affected by the prediction trajectory accuracy result in the first step. This twostage approach offers a promising solution for detecting AIS spoofing and can be applied to address a critical challenge in maritime traffic monitoring.

Multi-sensor-based surveillance systems with overlapping detection areas must address the challenge of associating objects detected by different sensors and accurately determining whether they correspond to the same real-world entity. Notably, different sensors may have distinct state vectors (e.g. AIS, radar, and passive sonar), and in some cases, information from systems such as AIS may be unavailable. Additionally, the area under surveillance may be vast and contain a large number of vessels, leading to a high computational cost for track association. This study aims to simulate the aforementioned scenario, which is commonly observed along the Brazilian coast today, and to apply a combination of association and fusion techniques to enhance problem-solving. Some models are classical, while others introduce contextual adaptations or modifications to their original conditions of use. Statistical results indicate that contextualization significantly reduces the execution time of multi-sensor association algorithms, while adaptations in classical algorithms improve target fusion across different sea conditions.

In precision viticulture, deep learning and computer vision have been increasingly employed to automate grape cluster counting, a crucial task for yield estimation and farm management. However, existing methods fail to address the unique challenges posed by the pergola vine training system - a widely used but understudied method, particularly in Brazil's Northeast, where viticulture plays a key role in economic and social development. To address these limitations, we propose a novel, lightweight approach that combines object detection (YOLOv8) and tracking (ByteTrack) to count grape clusters in handheld smartphone videos captured under diverse environmental conditions. Unlike previous works, our technique is applied to the unique layout of the pergola system, where occlusions, perspective variations, and irregular cluster distributions complicate automated counting. We evaluate our method on real-world vineyard footage, achieving$75.1 \% \text{mAP} {@} 50$and a 69.9 % F1-score for cluster detection, with a final cluster count only 27 % above the actual value-establishing the first baseline for this challenging task. By enabling cost-effective, accessible yield estimation, our work contributes to improving efficiency and equity in viticulture. We publicly release our implementation to encourage further research and practical adoption (https://github.com/artsasse/pergola-grape-count/blob/main/grape_cluster_count.ipynb).

Fishing activity matters to the entire world because it affects the economy, ecosystems, and human sustainability. Detecting and classifying fishing activity is a challenge that has been the focus of some studies over the years, but many of them are limited to using solutions without considering context information. These works show solid classification results but are limited to the classification task only. In general, most of them assume a fishing activity is in progress. Hence, these works have not explored the potential of false fishing detections, leading to misclassifications and forcing the fitting of detections into one type of fishing technique. Our hypothesis is that geographic context information can improve the detection of fishing activity and the classification of different types of fishing. Therefore, individual and collective information were extracted from a public labeled database used in related works. Individual information is the kinematics of each vessel, while collective information is the geographic fishing areas. The model adopts a stacking ensemble strategy, with the first level being a kinematic classifier and the second level a correlation model with geographic context. The solution presented effectively fills the identified gaps and demonstrates robust results.

Military systems require accurate technical solutions to support decision making. In naval warfare, an important problem to solve relies on a capability to detect higher level category artifacts, such as warship formations. This work aims to fill a gap observed in literature about this subject. In this paper, we proposed, implemented and tested a model to detect warship formations, capable of classifying the formation according to its type. We also investigate some published works related to this subject and point out the differences and gaps perceived. Using a Kohonen Network as classifier based on position and velocity of ships, this work describes the computational model used and the results obtained according to different number of ships and samples. Using synthetic data combined with real data, the results show, with different types of metrics, that the adopted solution has reliable and promising results and it is adequate to deal with a real-world problem.