Avaliação da qualidade fisiológica de sementes de milho por análise de imagem e espectroscopia no infravermelho próximo

Abstract

Maize is a highly expressive crop in Brazilian and world agribusiness, we are looking for more and more productivities and one way to achieve it is through the technology of seeds. The development of methodologies for the evaluation of seed quality is essential for quality control programs of the companies. With the advancement of technology, more and more automated techniques and equipment are being used to improve the detection of batches with different levels of vigor. Image analysis and Near Infrared Spectroscopy (NIR) are non-destructive, precise, fast and with little subjectivity techniques, and are increasingly used in the industry. The objective of this work was to verify the potential of using the technique of image analysis by the GroundEye® equipment through the characteristics of color and geometry and the technique of near infrared spectroscopy in the identification of lots of different physiological qualities of maize seeds. Two hybrids, whose seeds were artificially aged, were used for 0, 24, 48, 72, 96 and 120 hours. For the lots analysis, characterization tests (water content, germination, first germination count, cold test, field emergence, emergence speed index and seedling length) were performed. The design was a completely randomized design (DIC), with four replications in a 6 x 2 factorial scheme (lots of maize seeds and number of hybrids). Seed image analysis was performed on the Groundeye® equipment and color and geometry characteristics were analyzed. The NIR analysis was performed using whole and ground seed, in the whole seeds were obtained spectra of both sides of the seed (embryo and endosperm), the data obtained by the equipment were treated and analyzed by the PLS-DA method with the help of the software PIROUETTE. It is concluded that the image analysis did not allow to differentiate the seeds of corn of different physiological qualities through the characteristics of color and geometry and using the chemometric and the near infrared analyzes it was also not possible to separate the batches of different qualities.