Uso de imagens digitais na determinação quantitativa de fósforo, potássio, matéria orgânica, cálcio e magnésio em solo

Abstract

Soil fertility monitoring is essential for optimizing and increasing agricultural productivity. Therefore, this work aimed to evaluate the use of digital images for the quantification of phosphorus, potassium, organic matter, calcium, and magnesium in soil, through the RGB – red, green, and blue – color channels. The images, obtained with a smartphone, of 72 samples, were correlated with the experimental parameters of fertility through multivariate calibration. Principal component analysis (PCA) did not show a trend towards grouping or separating the samples as a function of the concentrations of each analyte. Partial least squares (PLS) regression with 4 latent variables did not show satisfactory results due to the low correlation between predicted and measured values.