Mapping sugarcane quality using artificial intelligence tools and high-resolution imagery

• Grant number: 23/12986-0
• Support Opportunities: Regular Research Grants
• Start date: May 01, 2024
• End date: April 30, 2026
• Field of knowledge: Agronomical Sciences - Agricultural Engineering - Agricultural Machinery and Implements
• Principal Investigator: Tatiana Fernanda Canata
• Grantee: Tatiana Fernanda Canata
• Host Institution: Faculdade de Ciências Agrárias e Veterinárias (FCAV). Universidade Estadual Paulista (UNESP). Campus de Jaboticabal. Jaboticabal , SP, Brazil

Abstract

Anticipating the sugar content of sugarcane stalks is essential to developing data-driven solutions to determine the optimal moment of the mechanized harvesting. However, traditional laboratory measurements are destructive and limited in scale, delaying the harvest planning on an industrial scale. The objective of the project is to characterize the spatial variability of sugarcane quality attributes through the integration of multiple sources of data using artificial intelligence. Samples will be collected in commercial sugarcane areas in the state of São Paulo to measure qualitative attributes in the laboratory during two consecutive seasons. Multispectral images will be acquired on the same field evaluation dates to calculate vegetation indices and obtaining plant canopy reflectance values. Meteorological data provided by a local meteorological station will also be used as a manner to increase the robustness of the predictive models. These models will be developed using artificial neural networks and random forest methods from the integration of ground-truth data and multispectral images of high-resolution. The development of this project will allow the characterization of the plant's spectral response according to its sugar content and, therefore, will support to prioritize the fields in ideal conditions for harvesting. Non-destructive estimation of sugarcane quality has potential to improve the use of inputs in the sector's production chain, in addition to mitigating the environmental impact resulting from mechanized operations. The results of this project will make possible to fill a national gap in relation to the development of technological innovations adapted to tropical agriculture, as well as enable the adoption of sustainable practices related to the early monitoring of the fields. (AU)