Principais Coleções do Repositório
Submissões Recentes
A methodology for synthetic data generation and validation applied to agricultural engineering
(Universidade Federal de Lavras, 2026-07-09) Brandão Junior, Luiz Carlos; Magalhães, Ricardo Rodrigues; Ferreira, Danton Diego; Ferreira, Danton Diego; Lima, Danilo Alves de; Lacerda, Wilian Soares; Vitor, Giovanni Bernardes
The scarcity of high-quality experimental data, coupled with the high costs of laboratory and sensory analyses, are significant barriers to the development of computational modeling and artificial intelligence in the field of Agricultural Sciences. This thesis introduces a reverse engineering framework for generating synthetic data based on Cholesky decomposition, which can reconstruct multivariate distributions while preserving statistical properties and structural dependence between variables. As a case study, the technique was employed in a model for the sensory quality of Arabica coffee using a public dataset from the Coffee Quality Institute (𝑛 = 207 samples). The method was divided into four stages: (i) data extraction and preprocessing (ii) statistical synthesis preserving first and second order moments (iii) multiple validation of statistical fidelity and (iv) evaluation of predictive power via the Train Synthetic, Test Real (TSTR) protocol. The results showed high structural fidelity, obtaining a correlation matrix similarity of 96.20% and a Kolmogorov-Smirnov (KS) test similarity of 85.81%; all variables exceeded the technical acceptance threshold of 70%. In RF regression, only the models trained with synthetic data achieved R² = 0.9498, a Mean Absolute Error (MAE) of only 0.216, and 100% accuracy within a margin of error of ±10%. For the ordinal classification of sensory quality, the approach achieved 89.37% accuracy and a Weighted Quadratic Kappa Coefficient of 0.9558, showing almost perfect agreement with the standards of the Specialty Coffee Association (SCA). As its main scientific contribution, this thesis establishes a reverse engineering framework for reconstructing synthetic databases from aggregated statistics, demonstrating that the simultaneous preservation of statistical fidelity and predictive utility can be achieved without the use of deep generative models. The proposed approach expands the possibilities for secure data sharing, reduces dependence on costly experimentation, and offers a computationally efficient and easily reproducible solution with high potential for technology transfer to applications in Agricultural Engineering, Agronomy, and other areas that depend on reliable experimental databases.
Cryopreservation of Coffea spp. seeds: thermal, hygroscopic, oxidative stress and targeted metabolomics characterization, associated with the physiological quality
(Universidade Federal de Lavras, 2026-08-05) Gontijo, Gabriela Ribeiro; Rosa, Sttela Dellyzete Veiga Franco da; Vilela, Ana Luiza de Oliveira; Alonso, Ana Paula; Nery, Marcela Carlota; Martins, Maria Alice; Coelho, Stefânia Vilas Boas; Veiga, Patrícia de Oliveira Alvim
The conservation of Coffea genetic resources remains challenging because the seeds exhibit intermediate to recalcitrant storage behavior, limiting the use of conventional conservation strategies in seed banks and increasing reliance on field collections. This study investigated the effects of seed water content, packaging type, the calorimetric properties of water in seed tissues, and targeted metabolomics on the physiological response of Coffea seeds subjected to immersion in liquid nitrogen. Seed physiological quality was evaluated by tetrazolium viability, electrical conductivity, and germination tests. Biochemical analyses included antioxidant responses, lipid composition, soluble sugars, cell wall carbohydrates, and biomass fractions. The results showed that successful cryopreservation depends on a narrow range of seed water content and is strongly affected by packaging type, which influences rewarming efficiency and tissue integrity. The absence of detectable freezable water was not sufficient to predict postcryopreservation performance, indicating that water behavior alone does not fully explain seed survival. Physiological recovery was associated with specific biochemical profiles, including soluble compounds, structural carbohydrates, lipid composition, and protein preservation. Compounds such as verbascose and galactose were consistently associated with higher viability and germination, indicating their potential as biochemical indicators of physiological quality. These findings show that cryopreservation tolerance in Coffea seeds is related to the interaction between water status, packaging conditions, and the intrinsic biochemical characteristics of the seeds.
Seletividade de herbicidas aplicados em pré-emergência e uso de fitorreguladores de crescimento na cultura do feijão-comum
(Universidade Federal de Lavras, 2026-02-27) Carvalho, Antonio Henrique Fonseca de; Moreira, Silvino Guimarães; Melo, Christiane Augusta Diniz; Silva, Amilton Ferreira da; Peixoto, Devison Souza; Pimentel, Guilherme Vieira
In the present study, an initial literature review is presented emphasizing the importance of pre-emergence herbicides for weed management in common bean (Phaseolus vulgaris L.) cultivation, as well as the limited number of active ingredients currently registered for this crop. Furthermore, the potential use of different compounds acting as growth regulators and physiological modulators in common bean production systems is discussed. Subsequently, the two scientific articles constituting the core of this thesis are presented. The first article aimed to evaluate the selectivity of pre-emergence herbicides in common bean and to quantify the yield benefits associated with their application. The experiment was conducted in Lavras, Minas Gerais, Brazil, under a randomized complete block design comprising 30 treatments and three replications, during two growing seasons (winter and summer), using the cultivar IAC Netuno. Phytotoxicity symptoms, weed control efficiency, yield components, and grain yield were assessed. Data were subjected to analysis of variance, and treatment means were grouped using the Scott–Knott test at the 5% probability level. The highest and most stable grain yields, in both winter and summer growing seasons, were obtained with the treatments containing S-metolachlor + fomesafen and flumioxazin + S-metolachlor, in addition to the weed-free control treatment. Conversely, diclosulam, sulfentrazone, and their respective combinations exhibited high phytotoxic potential to common bean plants, resulting in significant grain yield reductions. These findings demonstrate that, for the cultivar evaluation, only the mixtures S-metolachlor + fomesafen and flumioxazin + S-metolachlor exhibited satisfactory selectivity. The second article evaluated the effects of biostimulants and fungicides on the physiological management of common bean plants, associated with different application timings throughout the crop cycle. The experiments were carried out in three cultivation environments located in the municipalities of Carrancas (Environment I) and Lavras (Environments II and III), using the carioca bean cultivar IAC 2051 and the black bean cultivar IAC Netuno. The experimental design consisted of randomized complete blocks arranged in an 11 × 3 factorial scheme (ten treatments plus the untreated control combined with up to three application timings), with three replications. The total number of branches, yield components, and grain yield were evaluated. Benzyladenine and prohexadione-calcium demonstrated the greatest potential for increasing grain yield across both cultivars and cultivation environments. Benzyladenine exhibited greater responsiveness to an increased number of applications, whereas prohexadione-calcium showed greater efficiency when applied during the early developmental stages of the crop. Gibberellic acid exhibited limitations, particularly in black bean cultivars, due to excessive stimulation of vegetative growth, which negatively affected grain yield. Overall, the results demonstrate that growth regulators significantly influence yield components and grain productivity in common bean, with plant responses being dependent upon the product applied, cultivar, environmental conditions, and application frequency. Keywords: Phaseolus vulgaris; chemical control; growth modulators; common bean yield.
Characterization of structural changes in coffee plants after mechanized harvesting using airborne LiDAR
(Universidade Federal de Lavras, 2026-07-10) Velásquez Rubio, Santiago; Ferraz, Gabriel Araújo e Silva; Pérez Ruiz, Manuel; Sanz Uribe, Juan Rodrigo
The three-dimensional representation provided by the light detection and ranging sensor (LiDAR) offers possibilities for analyzing the structure of the coffee canopy, especially under conditions in which mechanized harvesting may be associated with changes in canopy organization. This study aimed to characterize structural changes in the coffee canopy associated with mechanized harvesting using metrics derived from LiDAR point clouds. LiDAR surveys were conducted before and after harvesting in an experimental Arabica coffee plot, followed by point cloud preprocessing, surface model generation, planting row segmentation, and extraction of structural metrics by segment. Subsequently, the data were analyzed using paired statistics, principal component analysis (PCA), and k-means clustering, in addition to the generation of vertical and transverse canopy profiles. The results indicated reductions in metrics associated with height, structural occupancy, and apparent canopy volume, accompanied by increased canopy openness and a greater relative presence of returns in lower strata. Heterogeneous spatial patterns of alteration were also identified along the planting rows, with different intensities and directions of structural response among segments. Overall, the profiles revealed vertical redistributions of the canopy after harvesting. Thus, this study contributes to expanding the understanding of how mechanized harvesting may be reflected in the structural organization of coffee plants and reinforces the potential of LiDAR as a support tool for spatial and vertical analysis of the crop.
Sujeitos e dispositivos de disciplinamento no cotidiano da escola
(Universidade Federal de Lavras, 2026-03-24) Araújo, Rafaela Silva; Betlinski, Carlos; Coelho, Allan da Silva; Carvalho, Alexandre Filordi de
The aim was to investigate and analyze the issues that arise in the daily life of the school and to understand how disciplinary processes occur. Based on this premise, this dissertation adopts the general objective of analyzing student behavior within the context of power relations in the school institution. Specific objectives include: identifying the main forms of manifestation of power relations between students and teachers; investigating how norms influence student behavior; analyzing the impact of power relations on the construction of student identity; and evaluating the role of pedagogical and administrative practices in the school institution, broadening the understanding of the subject. To this end, this research is specified through the production of two articles and an educational product. The first article consists of the theoretical framework, where a discussion anchored in the works of Foucault is carried out, offering valuable insights into the processes of subjectivation, allowing for a more comprehensive and rich analysis of power dynamics in the school environment. The second article consists of exploratory qualitative research, with data collected in three schools in southern Minas Gerais. A questionnaire containing twenty-four structured questions was used. Sixty students from three state schools participated, totaling an average of 20 students per school. A discussion was held to understand and identify the disciplinary processes existing in the school institutions. For the analysis of the results, we will use the rules presented by Bardin. Within this analytical framework, we detail the power relations existing in the aforementioned schools, highlighting and describing the devices and constituents of subjectivities. This research is justified by the need to understand how the subjects that make up the school (students, teachers, supervision, and administration) behave and relate to each other, seeking discussions to perceive their importance in the processes of socialization of knowledge, ethical and political values for students. In the third part, we present an educational product, which is a training proposal for the school's pedagogical team, based on the concept This educational product, a methodological guide for a teacher training workshop entitled "Power, Dialogue, and Pedagogical Practice," is aimed at teachers, specialists, and school administrators. It was developed to foster critical and collaborative reflections on power relations present in daily school life and their implications for the construction of pedagogical practices. This guide proposes to be an instrument for dialogue, knowledge sharing, and the re-signification of teaching as a political, ethical, and transformative act. The lack of training related to this theme often generates social reproduction, which causes class prejudice, gender bias, oppressive relationships, and authoritarianism. These are, therefore, propositions that permeate the development of this research, and we hope to contribute to knowledge about the school institution and its formative and political function in contemporary society.