CLASSIFICATION OF AREAS INFESTED BY NEMATODES IN SOYBEAN CROPS USING MULTISPECTRAL IMAGES: COMPARISON OF MACHINE LEARNING ALGORITHMS

• Grant number: 24/06421-2
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: July 01, 2024
• End date: August 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Geosciences - Geodesy
• Principal Investigator: Aluir Porfírio Dal Poz
• Grantee: Maria Angélica Padovani Ederli
• Supervisor: Eija Maarit Honkavaara
• Host Institution: Faculdade de Ciências e Tecnologia (FCT). Universidade Estadual Paulista (UNESP). Campus de Presidente Prudente. Presidente Prudente , SP, Brazil
• Institution abroad: Finnish Geospatial Research Institute, Finland
• Associated to the scholarship: 23/11289-3 - Classification of nematode infested areas based on a machine learning algorithm using Planet multispectral images, BP.IC

Abstract

Accurately identifying the occurrence of nematodes in crops is crucial to avoid losses in crop productivity. However, proper diagnosis requires a large number of soil and root samples throughout the cultivated area, which results in high costs. The use of Remote Sensing and Photogrammetry enables the precise detection and mapping of nematode infestations in soybean crops. The advanced technologies used in Precision Agriculture promote the optimization and automation of agricultural processes, bringing benefits to both production and the environment. The objective of the study is to investigate how multispectral images from Planet can be applied to detect areas affected by nematodes in soybean plantations. Additionally, the study aims to develop a method using machine learning techniques to classify these areas according to the level of nematode infestation. The study area is located on an experimental farm in Caiuá (SP), with the main focus of this study being the plots with soybean cultivars. Various vegetation indices will be used, such as the Normalized Difference Vegetation Index, Normalized Difference Green Vegetation Index, and the Plant-parasitic Nematode and Free-living Nematode Index will be tested. In this BEPE proposal, the Random Forest, Support Vector Machine, and k-Nearest Neighbor algorithms will be applied for area classification, and performance metrics will be used to evaluate the proposed model. The result will be a thematic map that visually indicates areas with different levels of infestation, providing useful information for crop management.