Lightweight Neural Architectures to Improve Sugarcane Production in Brazil

• Grant number: 24/00202-7
• Support Opportunities: Scholarships abroad - Research
• Start date: December 09, 2024
• End date: March 08, 2025
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: João Paulo Papa
• Grantee: João Paulo Papa
• Host Investigator: Hermano Igo Krebs
• Host Institution: Faculdade de Ciências (FC). Universidade Estadual Paulista (UNESP). Campus de Bauru. Bauru , SP, Brazil
• Institution abroad: Massachusetts Institute of Technology (MIT), United States
• Associated research grant: 13/07375-0 - CeMEAI - Center for Mathematical Sciences Applied to Industry, AP.CEPID

Abstract

Brazil is the world's largest sugarcane producer and sugar supplier, an important economic activity. Weed control must be adequate, for herbicides are likely to damage the culture. The problem worsens on perennial plants, e.g., sugarcane, since not cleaning the soil after a harvest is a widespread practice. Plants and weeds often get mixed, making their automatic differentiation a severe problem. This research project proposes to evaluate and devise new lightweight deep neural architectures that can effectively detect weeds in sugarcane cultures and be efficiently applied in real-time and as a field monitoring device.