Use of aerial imagery and high-throughput phenotyping in soybean crop

Abstract

Asian soybean rust, caused by the fungus Phakopsorapachyrhizi, is the most severe disease in the crop and can cause yield losses of up to 100%. Another pathogen that deserves attention in the crop is powdery mildew(Erysiphediffusa), which has seen a resurgence in recent years. Therefore, in order to maintain high grain yields, the number of foliar applications of fungicides has greatly increased. Thus, the development of new resistant cultivars is a crucial step in preserving the potential productivity of the crop, and the use of multilines emerges as a promising strategy in the pursuit of durable resistance. The use of aerial imagery and high-throughput phenotyping in soybean crop is an innovative approach that has gained prominence in modern agriculture, as it allows for the rapid, efficient, objective, and cost-effective collection of detailed information on a large scale when compared to traditional methodologies. Thus, the objective was to correlate vegetation indices obtained from unmanned aerial vehicles with grain yield, severity of Asian soybean rust andmildew severity in soybean crops under tropical conditions. Six commercial soybean cultivars with INOX® technology (TMG 7060 IPRO, TMG 7063 IPRO, TMG 7262 RR, TMG 7062 IPRO, TMG 7363 RR, TMG 7067 IPRO), a multiline (a mixture of lines), and a susceptible control for Asian soybean rust and powdery mildew (M6410 IPRO) were evaluated. The experiments were conducted in Lavras, Ijaci, and Nazareno, in the state of Minas Gerais, Brazil, during the 2020/21 growing season. The experimental design used was a randomized complete block design, with treatments decomposed into split-plot strips, 4 x 8 (four fungicide application managements and seven cultivars + one multiline), with three replications, with the number of applications in the plots and cultivars in the subplots. The images were collected using two drones, a DJI Mavic Pro equipped with a red-green-blue (RGB) camera and a DJI Matrice 100 with a Parrot Sequoia (multispectral) camera. The evaluated traits included severity of Asian soybean rust, powdery mildew severity, defoliation index, grain yield, and the vegetation indices NDVI and MPRI. Joint analysis (multi-environment) and split-plot analysis over time were performed. Statistical analyses were conducted using the R environment. The NDVI and MPRI vegetation indices showed a high correlation, allowing for the use of both RGB and multispectral cameras in high-throughput phenotyping for evaluating grain yield and powdery mildew severity in soybean crops under tropical highland climates. There was no significant correlation between Asian soybean rust severity and the other evaluated traits, which complicates the use of aerial images for quantifying the severity of Phakopsorapachyrhizi in tropical conditions.