NIRS and multivariate methods for discrimination of morning glory species at different growth stages

Morning glory species are weeds very common in tropical crops, where they cause direct and indirect damage. The management of these species primarily relies on the application of herbicides, disregarding the growth stage and spatial distribution. Studies addressing new techniques for identifying these species may contribute to the development of proximal sensors for carrying out specific and rational management. Thus, the objective of this work was to use near infrared spectroscopy (NIRS) and multivariate analysis to discriminate two species of morning glory in three growth stages. NIRS spectra were collected from Ipomoea hederifolia and Merremia aegyptia were collected at three different stages in the spectral range of 4.000 to 10.000 cm-1. PCA and PC-LDA were used to analyze the entire spectrum and specific bands. NIRS associated with PCA and PC-LDA were sufficient to discriminate I. hederifolia and M. aegyptia species and their growth stages. PCA allowed a proper segregation of stages and species when applied individually PC-LDA correctly classified between 90.93 to 100% of species and stages. The best discrimination results were observed in the NIR spectra ranges from 4.500 to 6.000 cm-1 and 4.500 to 6.000 + 6.500 to 7.750 cm-1. This study represents an advance in the research and implementation of NIRS technology to discriminate weed species for the future development of equipment to assist in the adoption and/or performance of a specific management of weeds, capable of contributing to the reduction in the use of herbicides in crops.