https://repositorio.ufu.br/handle/123456789/30498 2026-08-29T06:52:49Z 2026-08-29T06:52:49Z https://repositorio.ufu.br/handle/123456789/49733 2026-08-22T06:25:08Z 2025-11-28T00:00:00Z

Title: Análise do ganho de produtividade de soja inoculada com microrganismos solubilizadores de fósforo por meio de informações geoespaciais Abstract: One of the major challenges of soybean cultivation in tropical regions is the low soil fertility, which limits phosphorus availability and leads to the intensive use of synthetic chemical fertilizers. This increases production costs, creates environmental impacts, and promotes dependence on imported inputs. In modern agriculture, beneficial microorganisms have emerged as alternatives for improving nutrient use efficiency, especially those capable of solubilizing low-solubility inorganic phosphates. This represents a promising, environmentally sustainable, and economically viable approach that contributes to increasing the availability of this nutrient to plants. The interaction between inoculants and plants, in addition to being complex, can manifest subtly. In such cases, classical statistical methods are generally unable to reveal productivity differences between treated areas, particularly when considering the full dynamics involved in the various developmental stages of the crop. In this context, geostatistical tools have been used to estimate spatial relationships between agricultural yield and different crop management parameters. The aim of this study is to evaluate soybean yield gains associated with inoculation using phosphorus-solubilizing microorganisms through statistical and geostatistical analysis of georeferenced yield data and, additionally, to determine whether spectral variables derived from multispectral images can explain and predict this yield using machine-learning algorithms. The results showed that the Random Forest algorithm provided the best soybean yield estimates with the lowest error, supported by multispectral images that incorporated information about the crop’s phenological dynamics, generating accurate estimates with reduced physical effort. With a large amount of training data, the kriging technique made it possible to generate a surface showing an average yield increase of 3.93%, validating the effectiveness of the P-solubilizing agent.

2025-11-28T00:00:00Z https://repositorio.ufu.br/handle/123456789/48962 2026-07-23T06:22:19Z 2026-02-23T00:00:00Z

Title: Discrimination of coffee diseases and pests using multispectral imaging Abstract: A identificação precisa de doenças e pragas em cafeeiros é crucial para o manejo dirigido da lavoura e para evitar perdas de produtividade. Neste estudo, utilizaram-se imagens multiespectrais para monitorar um cafezal irrigado por gotejamento da cultivar Topázio MG 1190, implantado em 2016 na Universidade Federal de Uberlândia – Campus Monte Carmelo, com o objetivo de discriminar sintomas causados por Leucoptera coffeella (bicho mineiro), Cercospora coffeicola (cercosporiose) e Hemileia vastatrix (ferrugem do cafeeiro). As imagens, capturadas mensalmente de fevereiro a novembro de 2024 com um drone DJI Phantom 4, foram analisadas por meio de modelos de classificação supervisionada de aprendizado de máquina (Redes Neurais e Random Forests) desenvolvidos no software Weka. Paralelamente, foram avaliados 4 pares de folhas por planta (terço médio) para analisar a incidência de cada doença/praga. Os dados de entrada para as classificações consistiram nas bandas espectrais originais do sensor, nos índices de vegetação derivados e na classe designada por planta. Posteriormente, a metodologia foi testada em uma lavoura comercial em julho de 2025. Os resultados demonstraram que os modelos de classificação atingiram até 85% de acurácia e índice Kappa de 0,8, evidenciando o potencial do uso de imagens multiespectrais para o monitoramento contínuo das principais doenças e pragas da cafeicultura. Além disso, os mapas gerados pelos modelos permitiram analisar a distribuição espacial dos problemas fitossanitários estudados, mostrando-se uma ferramenta de apoio a práticas de manejo localizado. Desse modo, a pesquisa evidenciou que essa ferramenta pode ser uma alternativa para a promoção da cafeicultura sustentável, contribuindo para o controle mais eficiente de doenças e pragas.

2026-02-23T00:00:00Z https://repositorio.ufu.br/handle/123456789/48923 2026-07-21T06:27:39Z 2026-01-20T00:00:00Z

Title: Guiafertil: software de análise exploratória e predição do limiar de produtividade da cultura de soja com técnicas de inteligência artificial Abstract: Soybean cultivation plays a fundamental role in Brazilian agribusiness, requiring increasingly efficient management practices to ensure high levels of productivity and sustainability. However, the criteria currently used by agronomists to recommend fertilizer and soil amendment rates exhibit considerable variability, mainly due to the lack of regional calibration studies that simultaneously consider specific soil attributes, climatic conditions, altitude, planting systems, and other variables that influence crop development. This limitation hinders technical decision-making in the field and reduces the accuracy of agronomic recommendations. In this context, understanding the relationship between these factors and crop productivity becomes essential to enable more robust estimates and support management decisions. Advances in Artificial Intelligence (AI) techniques offer new opportunities to model such relationships and identify complex patterns that are often difficult to detect using traditional approaches. Therefore, this study presents the development of Guiafertil, a computational tool designed for exploratory analysis of agronomic data and soybean yield threshold prediction. Guiafertil integrates a database management system, data preprocessing routines, statistical analysis, graphical visualization, and machine learning algorithms, including Linear Regression, Support Vector Regression (SVR), Random Forest, and Artificial Neural Networks. Using real-world data obtained from the “Champion Cases” database of the Brazilian Soybean Strategic Committee (CESB), the system enables the evaluation of soil, climate, and management attributes, model training, performance analysis, and productivity estimation based on practical scenarios, such as different fertilization rates. Among the main results, the Multi-Layer Perceptron (MLP) Neural Network achieved the best predictive performance, reaching an R² value of 0.56 for soybean yield prediction bags/ha, whereas Linear Regression showed significant predictive limitations, obtaining an R² value of -0.60. Furthermore, the case study demonstrated that adjustments in fertilizer application rates affect production costs but have limited impact on predicted productivity, highlighting the importance of more accurate decision-making processes. The software quality and usability were evaluated by agronomists, who assigned average scores above 9.0 for most of the assessed criteria. Therefore, Guiafertil proved to be a promising tool for supporting agronomic recommendations, contributing to more efficient use of agricultural inputs, cost reduction, and greater sustainability in soybean production. Additionally, this work opens new opportunities for future research involving other crops, expansion of the database, and integration with precision agriculture technologies.

2026-01-20T00:00:00Z https://repositorio.ufu.br/handle/123456789/48770 2026-07-10T18:57:15Z 2026-03-06T00:00:00Z

Title: Diversidade espacial e prospecção de fungos fitopatogênicos para o controle biológico do cipó preto (Adenocalymma impressum) Abstract: The black liana (Adenocalymma impressum) poses a significant challenge to comercial Eucalyptus plantations in the eastern Amazon owing to its rapid growth and resistance to conventional herbicide-based control. This study aimed to prospect and characterize phytopathogenic fungi associated with A. impressum for potential use in biological control. Fungal isolates were collected from symptomatic plants in Maranhão, Brazil, resulting in 124 isolates, of which 24 were pathogenic. Among these, Lasiodiplodia theobromae SC-35 exhibited rapid symptom development, extensive tissue colonization, and host specificity, while remaining nonpathogenic to eucalyptus clones. Environmental factors, including temperature and native forest remnants, influenced fungal diversity and aggressiveness. These findings demonstrate the potential of SC-35 as candidate for development of a sustainable and effective bioherbicide for managing black liana, highlighting the value of endemic phytopathogens in integrated forest weed management strategies.

2026-03-06T00:00:00Z