Produtividade e caracterização fenotípica por meio de imagens multiespectrais de cultivares de morangueiro no Cerrado

Abstract

Strawberry is appreciated all over the world, being sensitive, mainly the photoperiod and temperature. Most strawberry cultivars planted in Brazil from genetic improvement programs in other countries. To optimize the program of genetic improvement, technologies are being developed, with the aid of remote sensing, to quantify morphophysiological attributes in the plant in a timely and fast way. The objective of the study was to evaluate the behavior of strawberry cultivars in terms of production and phenotypic characterization using multispectral images under environmental conditions in the Cerrado. The experiment was conducted in the field, at the Federal University of Uberlândia - Campus Monte Carmelo, MG. The experimental design adopted was in randomized blocks with six treatments (strawberry cultivars: San Andreas, Albion, PR, Festival, Oso Grande and Guarani) and four replications. The seedlings were transplanted to beds covered with double-sided plastic film with a drip irrigation system. The harvest was carried out weekly, starting at 40 days after planting (DAP) of the seedlings and ending at 164 DAP. The following were evaluated: number of fruits per plant; average fruit mass (g); total production (g plant-1 ); SPAD index, number of leaves, leaf length and width; soluble solids content (ºBrix); titratable acidity (% citric acid); vitamin C (100g-1 mg of fresh pasta); anthocyanins (mg 100g-1 of fresh pasta); fruit firmness (N); soil temperature under and over mulching, leaf temperature, air temperature and relative humidity. Multispectral images were obtained at 78 days after planting, at 12 o'clock, using an unmanned aerial vehicle, under which a Mapir Survey3W camera. With the mean reflectance values, the normalized difference vegetation index (NDVI) was calculated and Pearson's correlation was performed between the agronomic variables and the spectral bands. For the bands with the highest correlation, models were generated through the application of artificial intelligence, with the validation of the model determined by calculating the mean squared error. The obtained results allow to conclude that the temperature and the relative humidity of the air influenced the strawberry production. In cultivars with higher soil temperatures under and over mulching, and leaf temperature, the lowest yields were observed. The cultivar most adapted to the growing conditions was Festival. The use of multispectral images did not allow the estimation of the SPAD index and the average mass of the fruits. The number of leaves, leaf length, leaf width, production and number of fruits were highly correlated with the red, green and NDVI bands. It was possible to estimate agronomic variables through multispectral images at 78 DAP, which helps in the selection of new genetic materials.