Application of Remote Sensing and Computer Vision Techniques in the Phenotyping of Sugarcane Families

• Grant number: 24/16005-6
• Support Opportunities: Scholarships in Brazil - Master
• Start date: February 01, 2025
• End date: July 31, 2026
• Field of knowledge: Agronomical Sciences - Agronomy
• Principal Investigator: Anete Pereira de Souza
• Grantee: Luan Affonso Montandon Pompeu
• Host Institution: Centro de Biologia Molecular e Engenharia Genética (CBMEG). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
• Company: Ministério da Agricultura, Pecuária e Abastecimento (Brasil). Empresa Brasileira de Pesquisa Agropecuária (EMBRAPA). Embrapa Sede
• Associated research grant: 22/04006-2 - Center for Plant Molecular Breeding (CeM²P), AP.PCPE

Abstract

Sugarcane is a crucial crop for meeting food demands and facilitating the energy transition in the coming years. The crop has seen significant productivity gains through genetic improvement during breeding programs. However, sugarcane breeding faces challenges such as the long time required for cultivar development and difficulties with conventional phenotyping, which can be imprecise and costly. This project proposes the development and application of a high-throughput phenotyping (HTP) strategy for sugarcane families, using unmanned aerial vehicles (UAVs) and image analysis through computer vision. The project aims to compare digital phenotypes with already assessed phenotypic traits to validate a less expensive and more accurate method for the breeding program of the Sugarcane Center at the Agronomic Institute of Campinas. High-resolution RGB and multispectral images will be captured by UAV during several flights throughout the development of both plant cane and ratoon cane. The images will be processed, and features such as vegetation indices, estimates, and computer vision statistics will be extracted. It is expected that the data analysis will demonstrate the proposed methodology's efficiency in identifying phenotypic variations among the different families obtained through crosses, with a correlation between remote sensing data and traditional evaluations. The results of this project have the potential to contribute a precise, cost-effective, and efficient method to the phenotyping process of a Brazilian sugarcane breeding program, significantly impacting productivity.