Skip to main navigation menu Skip to main content Skip to site footer

Coffee Crops Analysis Using UAVs Equipped with Multispectral Cameras

Abstract

The use of Unmanned Aerial Vehicles (UAVs) equipped with spectral cameras has increased in recent years, especially in the agricultural sector, because it allows farmers and researchers to analyze the state of a crop, i.e., health, nutrients, growth, epidemics, among other parameters. In Colombia, the coffee production sector faces several challenges, such as the need to increase the productivity, the yield, and the quality of coffee. This work estimated the health status of a Castilla variety crop located in San Joaquín, Tambo, Cauca to support the decision-making of coffee growers. For this, chlorophyll data were measured in the field with the CCM-200 plus device, multispectral images were captured with the MAPIR SURVEY 3 camera airborne on a SOLO 3DR UAV, and synthetic data were generated to increase the data set. Six vegetation indices were set, which—together with the chlorophyll values—were modeled through the implementation of simple and multiple linear regressions, decision trees, vector machines, random forests, and k-nearest neighbors. The model with the best performance and the lowest mean square error was disorder with the support vector machine. Likewise, the best performance indices in the models were CVI, GNDVI, and GCI, which are widely used in agriculture to estimate the chlorophyll of plants.

Keywords

agriculture, coffee, multispectral images, synthetic data, vegetation index, UAV

PDF XML

Author Biography

Natalia Arteaga-López

Rol: Conceptualization; Analysis; Writing – original draft.

Carlos Delgado-Calvache

Roles: Conceptualization; Analysis; Writing – original draft.

Juan-Fernando Casanova

Roles: Conceptualization; Methodology; Writing – edit and review.

Cristian Figeroa

Rol: Conceptualization; Methodology; Writing – edit and review.


References

Downloads

Download data is not yet available.

Similar Articles

You may also start an advanced similarity search for this article.