Weed discrimination based on the spectral response of the corn crop, Manabí, Ecuador

Abstract

Precision agriculture allows to gain efficiency, sustainability and profitability, since it provides great benefits in reducing the environmental impact of agriculture, economic risks and at the same time contributes to controlling the vigor of crops and improving the quality of their yield. The objective of this research is to discriminate weeds within the corn crop, based on their spectral response. For this, the advanced EBEE SQ agricultural drone was used, with which multispectral images were captured through its Parrot Sequoia camera. The images were processed with software in Geographic Information Systems (GIS). With the multispectral bands, different vegetation indices were calculated such as NDVI, NDVIAS, NGRDI, NDRE, GNDVI, using map algebra tools in specialized programs. A supervised classification was applied to the different indices to discriminate the different land covers, which obtained a precision of 93% and a Kappa index of 0.93. The results allowed to clearly differentiate the coverage of crops, weeds and bare soil. The data showed that both early-growing and developed weeds occupy 38% of the crop area. With this information, it is possible to improve the planning of agronomic practices, adding the herbicide at the specific site of the weeds.