Epamig monitors water conditions of coffee plants with drones, field sensors and satellite images

The Agricultural Research Company of Minas Gerais (Epamig) has developed extensive work to help coffee producers detect, remotely and accurately, which areas of their crops have a water deficit and require differentiated management. The research “Spectral monitoring to estimate water conditions in coffee growing areas”, which is divided into five subprojects, began in 2020 and has brought innovations to the remote monitoring of rural properties.

Currently conducted in coffee growing areas in the Minas Gerais municipalities of São Sebastião do Paraíso and Três Pontas, the project is financed by the Brazilian Agricultural Research Corporation (Embrapa), through the Coffee Research and Development Program – Consórcio Pesquisa Café, and also has support from the National Institute of Coffee Science and Technology (INCT Café), Minas Gerais State Research Support Foundation (Fapemig), National Council for Scientific and Technological Development (CNPq) and the Coordination for the Improvement of Higher Education Personnel ( Capes).

Research has associated reflectance indices with the health of coffee plants in crops. When the sun emits energy and solar radiation, part is absorbed by the plant, to carry out photosynthesis, another part is emitted in the form of fluorescence, and a final part is reflected by the plant. It is this reflected portion that determines its reflectance index. A healthier, well-hydrated plant absorbs more energy for photosynthesis and, consequently, reflects less of this energy. Satellites and drones that monitor areas of the Earth's surface are capable of recording spectral images of this reflected energy and generating data that are correlated with the water potential of plants, whose variation is determined by images of different colors.

Epamig has been monitoring agricultural areas via satellite images since the beginning of the 2000s, however this is the first work that integrates this data, at the orbital level, with that generated at the leaf level, by sensors used directly in the field, and in canopy level, using Unmanned Aerial Vehicles (UAVs), popularly called “drones”.

The spectral data (which are those generated by measuring the color reflectance in plants) obtained at these three levels (orbital, canopy and leaf) are being computed and combined using Artificial Intelligence and multivariate statistical techniques, so that researchers can establish ideal mathematical models to measure the water conditions of coffee trees. In addition, the team also aims to improve the “Regador” cell phone application, initially developed in work carried out throughout 2014, in the municipality of Santo Antônio do Amparo (MG), in partnership with the professor at the Federal University of Itajubá (UNIFEI) , Vanessa Souza. The software, which was registered with the National Institute of Industrial Property (INPI) in 2022, assesses the water status of coffee plants with satellite spectral images and is in the improvement phase.

“In the current work, we are using high-resolution images generated by a European Space Agency (ESA) satellite and applying several spectral vegetation indices, as we saw that the best known of them, the NDVI, does not guarantee the precision we are looking for. We saw that these other indices are capable of translating in more detail, into images, the conditions of the plants and can, therefore, contribute to improving the application already developed”, explains the general coordinator of the project, and researcher at Epamig Sul, Margarete Volpato.

According to her, the information generated by the program will help coffee producers adjust the management of their crops in cases of water deficit, changing irrigation, choosing spacing and coffee tree varieties, among other actions. “The data will also help producers to protect themselves against diseases and pests, as they will know, precisely, which plots are most vulnerable. Furthermore, the integration of remote monitoring into field work reduces the time needed to analyze the water conditions of plants, taking just a few minutes to generate results for a large area, which traditionally can take weeks or even months”, adds Margarete.

From the sky to the leaves

One of the subprojects linked to the research is entitled “Unmanned aerial vehicles for monitoring water conditions in areas”, which has contributed with data generated from UAV images in flights at approximately 100 m altitude over coffee properties. “The application of this equipment is advantageous, as the images generated are not as affected by the presence of clouds in the sky, as is the case with satellites. So, this allows us to carry out flights when we need to, even on cloudy days, for example”, says the subproject coordinator, and researcher at EPAMIG Headquarters, Marley Lamounier.

He reports that he goes to each location three to four times a year to obtain specific information about the rainy and dry seasons. It has used an eBee SQ UAV (fixed wing), a Matrice 210 UAV (multirotor), as well as a GNSS Geodetic RTK model R8 ground positioning system, used for georeferencing and offering millimeter precision. “The drones’ multispectral cameras allow us to see the plant with different wavelengths, which go beyond the capabilities of the human eye, generating a large sequence of images. Subsequently, all of them are juxtaposed to form a single image, which we call a mosaic”, he highlights.

For the information generated to reach maximum precision, it is also necessary to integrate data on the plants' plant physiology, collected using equipment used directly on the soil and leaves of coffee trees. This is the work carried out within the subproject “Leaf reflectance patterns of coffee trees as indicators of water status”, coordinated by Epamig Sul researcher and CNPq Productivity in Technological Development and Innovative Extension fellow, Vânia Aparecida da Silva.

With the help of soil compaction and chlorophyll meters, as well as a mini leaf spectrometer and a Scholander Pump, the team is able to calculate the water potential of plants in crops. “The water potential directly reflects the free energy of the water that the plant has and which will be used for growth, development and production of coffee. The more hydrated it is, the healthier and more developed the plant will be, which has a positive impact on its productivity”, says Vânia.

She explains that spectral indices are also obtained at leaf level, and later computed with those obtained using drones and satellites. “This reflectance pattern, which we observed, changes according to the water status of the plant. By integrating the three levels of analysis, from the sky to the leaves, we will establish correction factors for efficient and accurate results”, argues the researcher.

The other subprojects are: “Remote sensing for monitoring water conditions in coffee growing areas”, coordinated by Embrapa Café researcher, Helena Maria Ramos Alves; “Study of the spatial variability of soil moisture using sensors”, coordinated by EPAMIG Sul researcher, Vanessa Castro Figueiredo; and “Modeling spectral variables from orbital and suborbital sensors to determine the water conditions of coffee trees”, coordinated by professor at the Federal University of Lavras (UFLA), Danton Diego Ferreira.

According to the coordinators, at the end of the research, Epamig will seek to establish partnerships with companies and institutions in the field of computer science to enable the implementation of the “Regador” application at a commercial level and make it available to the general public.