MACHINE VISION TECHNIQUES AND MULTIVARIATE CLASSIFIERS FOR NITROGEN FERTILIZATION DOSES DISCRIMINATION IN WHEAT

摘要

abstract_textpSidedress nitrogen fertilization is currently discussed throughout the world due to its economical and environmental implications. The cereal crops strongly respond to N application and there is a lack of current methods to determine N availability on the soil. The aim of this work was to evaluate the discrimination among three nutritional levels in wheat crop using digital images and a portable chlorophyll meter. Data were collected in plots with three levels of N (0; 30 and 60 kha(-1)) in three dates (8; 14 and 20 days after sidedress fertilization). The images were processed using nine spectral indices and elaborated multivariate classifiers based on the mean pixel values. The chlorophyll data and leaf nitrogen concentration were used in univariate classifiers. The classification using the machine vision techniques were better than the chlorophyll meter (SPAD) at 8 DAF, since the Kappa coefficient was better than a random classification. At 14 and 20 DAF there were no statistical differences between this type of data and the data from images. Using digital images it was possible discriminate the nutritional levels eight days after sidedress fertilization./p/abstract_text