Repository logo

Aplicação de redes neurais convolucionais para medição de patinagem em tratores agrícolas

dc.contributor.advisorSilva, Rouverson Pereira [UNESP]
dc.contributor.authorFeih, Rafael Rodrigues [UNESP]
dc.contributor.institutionUniversidade Estadual Paulista (Unesp)
dc.date.accessioned2021-08-24T14:34:11Z
dc.date.available2021-08-24T14:34:11Z
dc.date.issued2021-06-18
dc.description.abstractAssim como na Indústria, a era 4.0 chegou na Agricultura. Técnicas, ferramentas, sistemas de tecnologia da informação e robótica passaram a contribuir ativamente para a digitalização da Agricultura. Além da evolução tecnológica, existe uma orientação à sustentabilidade da Agricultura e utilização racional dos recursos naturais renováveis e não renováveis. Os tratores agrícolas atuais são energeticamente mais eficientes e menos poluentes desde que estejam adequados a operação em curso. Caso o conjunto trator-implemento-solo (TIS) esteja mal equilibrado, haverá redução da eficiência operacional, aumento no consumo de combustível, incremento dos custos de manutenção, desgaste prematuro dos pneus entre outros efeitos na operação agrícola. O índice de patinagem é um dos entes a serem medidos e controlados para verificar o quão bem balanceado o TIS está. Nesse contexto misto de tradição e inovação, objetivou-se neste trabalho a verificação da aplicabilidade de redes neurais convolucionais para reconhecimento e contagem de padrões visuais para então, indiretamente, medir o deslocamento das rodas de um trator com e sem carga. De posse dessas informações, foi possível determinar os índices de patinagem de dois tratores agrícolas. Futuramente essa tecnologia poderá ser adaptada para aplicações em smartphones comerciais. De posse dos níveis de patinagem, os produtores agrícolas poderão fazer os ajustes necessários à operação e assim, usar mais racionalmente os recursos naturais, máquinas, implementos agrícolas e mão-de-obra com vistas à sustentabilidade do negócio.pt
dc.description.abstractThe 4.0 era has arrived to the Agriculture as it had happened into the Industry. Techniques, tools and information systems and robotics started contributing actively to the Agriculture digitalization. It is not only technological evolution, but also growing concern regarding sustainability and proper use of renewable and nonrenewable resources. In general, current farm tractors are more energetically efficient and less polluter if they are properly adjusted to the ongoing operation. When the set tractor-implement-soil (TIS) is misbalanced, there will be reduction of operational efficiency, increase in fuel consumption and maintenance costs, premature wear to tires among other negative effects to the operation. The slippage index is one of the entity to be measured and controlled to know how well TIS is balanced. In the context of innovation and traditional concepts, this study aimed to evaluate the applicability of convolutional neural networks (CNNs) to recognize and count marked patterns into the tire sidewall to then, indirectly, measure the displacement of the rear wheel with and without load. Based on these pieces of information, it was possible to measure, with certain limitation, the slippage index of farm tractors. In the future, this technology may be adapted to commercial smartphones as an application. By knows the slippage index at a glance, farmers will be able to do the necessary adjustments to the operation. As a result, there will be a more rational use of natural resources, machinery, implements and labor while keeping business sustainability.en
dc.identifier.capes33004102001P4
dc.identifier.urihttp://hdl.handle.net/11449/214155
dc.language.isopor
dc.publisherUniversidade Estadual Paulista (Unesp)
dc.rights.accessRightsAcesso abertopt
dc.subjectAgricultura de Precisãopt
dc.subjectInteligência Artificialpt
dc.subjectSustentabilidadept
dc.subjectArtificial Intelligenceen
dc.subjectSustainabilityen
dc.titleAplicação de redes neurais convolucionais para medição de patinagem em tratores agrícolaspt
dc.title.alternativeApplication of convolutional neural networks to measure agricultural tractor slipagge indexen
dc.typeDissertação de mestradopt
dspace.entity.typePublication
relation.isOrgUnitOfPublication3d807254-e442-45e5-a80b-0f6bf3a26e48
relation.isOrgUnitOfPublication.latestForDiscovery3d807254-e442-45e5-a80b-0f6bf3a26e48
unesp.campusUniversidade Estadual Paulista (UNESP), Faculdade de Ciências Agrárias e Veterinárias, Jaboticabalpt
unesp.embargo24 meses após a data da defesapt
unesp.examinationboard.typeBanca restritapt
unesp.graduateProgramAgronomia (Produção Vegetal) - FCAVpt
unesp.knowledgeAreaProdução vegetalpt
unesp.researchAreaNão constapt

Files

Original bundle

Now showing 1 - 2 of 2
Loading...
Thumbnail Image
Name:
feih_rr_me_jabo_par.pdf
Size:
18.05 MB
Format:
Adobe Portable Document Format
Description:
Loading...
Thumbnail Image
Name:
feih_rr_me_jabo_int.pdf
Size:
385.85 MB
Format:
Adobe Portable Document Format
Description:

License bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
license.txt
Size:
2.98 KB
Format:
Item-specific license agreed upon to submission
Description: