Advancing machine learning for remote sensing applications in agriculture: Case study of the Brazilian Cerrado
Conference Year
January 2023
Abstract
To meet the challenge of sustainable agricultural growth, accurate inventory and estimation of crop types is important. As agricultural activities intensify, the land use and land cover of the region is reshaped. The first step towards understanding the consequences of these changes is to quantify land‐cover and land‐use changes (LCLUCs). The objective of this study is to expand the spatial and temporal crop types of datasets of the Brazilian Cerrado, a biodiversity hotspot and globally important region undergoing rapid agricultural expansion and climate changes from 2000-2020. Random Forest machine learning model was used for the classification. In terms of classification accuracy, a combination of parameters was tested and the highest accuracy classification with an overall accuracy was 80.1%
Primary Faculty Mentor Name
Gillian Galford
Status
Graduate
Student College
Rubenstein School of Environmental and Natural Resources
Second Student College
Graduate College
Program/Major
Natural Resources
Primary Research Category
Physical Science
Advancing machine learning for remote sensing applications in agriculture: Case study of the Brazilian Cerrado
To meet the challenge of sustainable agricultural growth, accurate inventory and estimation of crop types is important. As agricultural activities intensify, the land use and land cover of the region is reshaped. The first step towards understanding the consequences of these changes is to quantify land‐cover and land‐use changes (LCLUCs). The objective of this study is to expand the spatial and temporal crop types of datasets of the Brazilian Cerrado, a biodiversity hotspot and globally important region undergoing rapid agricultural expansion and climate changes from 2000-2020. Random Forest machine learning model was used for the classification. In terms of classification accuracy, a combination of parameters was tested and the highest accuracy classification with an overall accuracy was 80.1%