Classification of nematode infested areas based on a machine learning algorithm using Planet multispectral images

• Grant number: 23/11289-3
• Support Opportunities: Scholarships in Brazil - Scientific Initiation
• Effective date (Start): November 01, 2023
• Effective date (End): December 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Geosciences - Geodesy
• Principal Investigator: Aluir Porfírio Dal Poz
• Grantee: Maria Angélica Padovani Ederli
• Host Institution: Faculdade de Ciências e Tecnologia (FCT). Universidade Estadual Paulista (UNESP). Campus de Presidente Prudente. Presidente Prudente , SP, Brazil
• Associated research grant: 21/06029-7 - High resolution remote sensing for digital agriculture, AP.TEM
• Associated scholarship(s): 24/06421-2 - CLASSIFICATION OF AREAS INFESTED BY NEMATODES IN SOYBEAN CROPS USING MULTISPECTRAL IMAGES: COMPARISON OF MACHINE LEARNING ALGORITHMS, BE.EP.IC

Abstract

Accurately identifying the occurrence of nematodes in plantations is crucial to prevent losses in crop productivity. However, proper diagnosis demands a large number of soil and root samples across the cultivated area, resulting in high costs. The use of Remote Sensing and Photogrammetry enables precise detection and mapping of nematode infestations in soybean crops. Advanced Precision Agriculture technologies optimize and automate agricultural processes, offering benefits to both production and the environment. The study aims to investigate how Planet's multispectral imagery can be applied to detect nematode-affected areas in soybean plantations. Furthermore, the goal is to develop a method using machine learning techniques to classify these areas based on the level of nematode infestation. The study area is located in an experimental farm in Caiuá (SP), with a primary focus on soybean cultivar plots. Various vegetation indices will be employed, including Normalized Difference Vegetation Index, Normalized Difference Green Vegetation Index, and the Phytomatodes and Free-Living Nematodes Index will be tested. The Random Forest algorithm will be applied for area classification, and performance metrics will be used to assess the proposed model. The outcome will be a thematic map that visually indicates areas with different infestation levels, providing valuable insights for crop management.