Unmanned aerial vehicles to determine soybean plant injury caused by pre-emergence herbicides

Pedro Eduardo Rampazzo1*, Carlos Henrique de Lima e Silva1, Victor Hugo Moraes1, Alice Maria Albert1, Tavvs Micael Alves1, Adriano Jakelaitis1, and Jorge Luis Tejada2
Images from unmanned aerial vehicles (UAV) can serve as a baseline for studies in weed science, complementing observations obtained in the ground. The objective of this work was to determine soybean (Glycine max (L.) Merr.) plant injury caused by pre-emergence herbicides in sandy and clayey soils, using a low-cost UAV. The experiment was conducted in a randomized complete block design, with four replicates and seven treatments consisted of herbicides (diclosulam, chlorimuron, sulfentrazone, flumioxazin, and S-metolachlor), hand weeded and untreated treatments. Ground-based evaluations were carried out to assess soybean crop injury, plant stand, leaf chlorophyll content, plant height, canopy distance and grain yield. Images were taken using a UAV equipped with an RGB (red green and blue) camera. Soybean plants sprayed with diclosulam had lower plant reflectance in the R (98.9), G (147.1) and B (74.3) range than the other treatments in sandy soil. In clayey soil, hand weeded treatment had higher plants (30.8 cm) and untreated favored smaller plants (24.9 cm) compared to herbicide treatments. In sandy soil, soybean yield of all treatments was similar, however in clayey soil, soybean yield treated with chlorimuron and flumioxazin was higher than 5000 kg ha-1 and better than the others treatments. The nutrient-poor soil (sandy) may have aggravated the plant injury caused by herbicides and explain the lower yield observed compared to clayey soil. It was determined soybean plant injury caused by pre-emergence herbicides with the use of unmanned aerial vehicles, providing complementary results to ground-based measurements, indicating the potential of this technology for low-cost evaluations in weed science.
Keywords: Digital weed management, Glycine max, low-cost RGB sensor, post-application evaluation, spectral responses.
1Instituto Federal Goiano Campus Rio Verde, Rodovia Sul Goiana km 01 Zona Rural, CEP 75901-970, Rio Verde, Goiás, Brasil.
*Corresponding author (pedrorampazzo@gmail.com).
2Universidade Estadual Paulista Júlio de Mesquita Filho, Faculdade de Ciências Agrárias e Veterinárias, Via de Acesso Prof. Paulo Donato Castellane s/n, CEP 14884-900, Jaboticabal, São Paulo, Brasil.