Mario Luiz Tronco

Spatial Big Data mining is often hindered by high computational complexity and the intrinsic autocorrelation of georeferenced records. To address these challenges, this study proposes an architectural optimization framework for the CHSMST+ algorithm, designated as CHSMST+MR. Rather than introducing a brand-new clustering paradigm, the framework focuses on a Distributed Spatial Cardinality Reduction (DSCR) layer that aggregates redundant spatial records before the core iterative mining logic begins. By transforming raw records into a weighted key-value representation within the Apache Spark environment, the proposed approach significantly mitigates the shuffling bottleneck common in distributed systems. Experimental validation using high-density biological datasets demonstrates an average execution-time reduction of 51.36%, with performance gains reaching up to 79.96% in specific high-redundancy scenarios. The results, obtained through controlled local emulation, confirm that this architectural optimization provides a scalable, deterministic, and lossless solution for accelerating spatial clustering. This work contributes a methodological path for enhancing the performance of iterative spatial mining algorithms in environments characterized by massive data density and coordinate redundancy.

Computer vision systems are essential for automating agricultural tasks such as disease detection and fruit defect identification. However, their application in coffee farming faces significant challenges due to environmental variability and the complex structure of coffee trees, which complicate image acquisition. Thus, this study addresses two key questions: 1) Can low-cost, user-friendly equipment adapt to crop conditions while ensuring image quality 2) Can a computer vision algorithm accurately count and classify coffee beans with over 80% accuracy using data from low-cost cameras To answer these questions, an image acquisition system was developed based on the phenological characteristics of coffee plants, ensuring focused and consistent image capture. Additionally, a novel algorithm was created, utilizing statistical analysis of color spaces to effectively separate fruits from the background, segment images, and count fruits. The algorithm achieved accuracy rates, when compared with a traditional approach, within the desired range for each coffee fruit class: green (83%), green-olive (79%), cherry (86%), and raisin (80%). These results demonstrate the potential of this approach for accurate and efficient fruit processing in coffee farming, particularly when images are captured directly from tree branches.

In peanut cultivation, the fact that the fruits develop underground presents significant challenges for mechanized harvesting, leading to high loss rates, with values that can exceed 30% of the total production. Since the harvest is conducted indirectly in two stages, losses are higher during the digging/inverter stage than the collection stage. During the digging process, losses account for about 60% to 70% of total losses, and this operation directly influences the losses during the collection stage. Experimental studies in production fields indicate a strong correlation between losses and the height of the windrow formed after the digging/inversion process, with a positive correlation coefficient of 98.4%. In response to this high correlation, this article presents a system for estimating the windrow profile during mechanized peanut harvesting, allowing for the measurement of crucial characteristics such as the height, width and shape of the windrow, among others. The device uses an infrared laser beam projected onto the ground. The laser projection is captured by a camera strategically positioned above the analyzed area, and through advanced image processing techniques using triangulation, it is possible to measure the windrow profile at sampled points during a real experiment under direct sunlight. The technical literature does not mention any system with these specific characteristics utilizing the techniques described in this article. A comparison between the results obtained with the proposed system and those obtained with a manual profilometer showed a root mean square error of only 28 mm. The proposed system demonstrates significantly greater precision and operates without direct contact with the soil, making it suitable for dynamic implementation in a control mesh for a digging/inversion device in mechanized peanut harvesting and, with minimal adaptations, in other crops, such as beans and potatoes.

In digital farming, the use of technology to increase agricultural production through automated tasks has recently integrated the development of AgBots for more reliable data collection using autonomous navigation. These AgBots are equipped with various sensors such as GNSS, cameras, and LiDAR, but these sensors can be prone to limitations such as low accuracy for under-canopy navigation with GNSS, sensitivity to outdoor lighting and platform vibration with cameras, and LiDAR occlusion issues. In order to address these limitations and ensure robust autonomous navigation, this paper presents a sensor selection methodology based on the identification of environmental conditions using sensor data. Through the extraction of features from GNSS, images, and point clouds, we are able to determine the feasibility of using each sensor and create a selection vector indicating its viability. Our results demonstrate that the proposed methodology effectively selects between the use of cameras or LiDAR within crops and GNSS outside of crops, at least 87% of the time. The main problem found is that, in the transition from inside to outside and from outside to inside the crop, GNSS features take 20 s to adapt. We compare a variety of classification algorithms in terms of performance and computational cost and the results show that our method has higher performance and lower computational cost. Overall, this methodology allows for the low-cost selection of the most suitable sensor for a given agricultural environment.

In the context of the new industrial revolution - Industry 4.0, the smart factory concept brought, to manufacturing, the idea of using large amounts of data acquired from a machining process and a set of mathematical techniques, discovering correlations, patterns, or trends in this database. Thus, machine tools are the focus of research in order to monitor and analyze the quality of the machining process based on data from embedded sensors. Based on this strategy, a signature process was created, that consists in capturing behavior patterns of a machine, such as machining conditions, machining quality, or tool wear. This article deal with a comparison between three Multinomial Logistic Regressions: the first using only time domain data, the second using only frequency domain data, and finally, the third using time and frequency domain data to identify the pattern of feed rate, depth of cut, and material being machined. It was observed that the methods had a precision of 96.25%, 37.92%, and 99.58%, respectively, showing that this methodology has great predictive efficiency and could be used to monitor the cutting parameters and material studied in this paper.

The advancement of technology and the needs imposed by the market, such as high surface quality and reduction of manufacturing costs, have made machine tools the focus of numerous researches, aimed at improving their performance in different industrial processes. Some of these researches have as objective the instrumentation of the machine, aiming at the monitoring of some variables related to the machining process. To carry out such monitoring, systems with sensors are used that allow the measurement of physical quantities that are closely linked to the cutting parameters (feed, cutting speed and cutting depth). These parameters are of fundamental importance for the process and help in the good machinability of the materials, an error in the configuration of these parameters, before or during the process, can cause wear or even breakage of the tool. The main quantities used in monitoring machine tools are mechanical vibration (correlated with the rotation of the workpiece and the movement and wear of the tool), temperature (which varies mainly with the plastic deformation at the workpiece-tool interface and the friction at the workpiece-tool and chip-tool interfaces) and sound vibration (related to stiffness and friction between materials). This article presents the development of a system that monitors acceleration, temperature, and sound emitted during an external turning process, using an embedded system (data acquisition and processing boards) with low cost and high computational power. To determine the process signature that relates temperature, vibration and audio to the machined material and to the cutting parameters (feed and depth of cut), an Artificial Neural Network (ANN) was used, which obtained an accuracy rate of approximately 80%, indicating that the monitoring system is capable of generating data that allows the determination of the process signature.

In agriculture, intelligent systems applications have generated great advances in automating some processes in the production chain. To improve the efficiency of those systems is proposes a vision algorithm to estimate the amount of fruits in sweet orange trees. This study proposes a computer vision system based on the capture of thermal images and fuzzy image processing. A bibliographical review has been done to analyze the state-of-the-art of the different systems used in fruit recognition, and also the different applications of thermography in agricultural systems. The algorithm developed for this project uses the intensification operator to contrast-enhanced and the fuzzy divergence for segmentation and Hough transform for fruit identification. It estimates the numbers of fruits in the tree, a task that is currently manually performed. In order to validate the proposed algorithm a database was created with images of sweet orange acquired in the Maringá Farm. The validation process indicated that the variation of the tree branch and the fruit temperature is not very high, making it difficult to segment the images using a temperature threshold. Errors in the segmentation algorithm could mean the increase of false positives in the fruit-counting algorithm. Recognition of isolated fruits with the proposed algorithm presented an overall accuracy of 93.5 percent and grouped fruits accuracy was 80 percent. The experiments show the need of other image hardware to improve the recognition of small temperature changes in the image.

Data mining algorithms to find association rules are an important tool to extract knowledge from databases. However, these algorithms produce an enormous amount of rules, many of which could be redundant or irrelevant for a specific decision-making process. Also, the use of previous knowledge and hypothesis are not considered by these algorithms. On the other hand, most existing data mining approaches look for patterns in a single data table, ignoring the relations presented in relational databases. The contribution of this paper is the proposition of a multi-relational data mining algorithm based on association rules, called TBMR-Radix, which considers previous knowledge and hypothesis through the using of the Templates technique. Applying this approach over two real databases, we were able to reduce the number of generated rules, use the existing knowledge about the data and reduce the waste of computational resources while processing. Our experiments show that the developed algorithm was also able to perform in a multi-relational environment, while the MR-Radix, that does not use Templates technique, was not.

The progressive growth in the volume of digital data has become a technological challenge of great interest in the field of computer science. That comes because, with the spread of personal computers and networks worldwide, content generation is taking larger proportions and very different formats from what had been usual until then. To analyze and extract relevant knowledge from these masses of complex and large volume data is particularly interesting, but before that, it is necessary to develop techniques to encourage their resilient storage. Very often, storage systems use a replication scheme for preserving the integrity of stored data. This involves generating copies of all information that, if lost by individual hardware failures inherent in any massive storage infrastructure, do not compromise access to what was stored. However, it was realized that accommodate such copies requires a real storage space often much greater than the information would originally occupy. Because of that, there is error correction codes, or erasure codes, which has been used with a mathematical approach considerably more refined than the simple replication, generating a smaller storage overhead than their predecessors techniques. The contribution of this work is a fully decentralized storage strategy that, on average, presents performance improvements of over 80% in access latency for both replicated and encoded data, while minimizing by 55% the overhead for a terabyte-sized dataset when encoded and compared to related works of the literature.