Análise de imagens na avaliação de morfologia de sementes e desempenho de plântulas de feijão-azuki (Vigna angularis Willd.)

Abstract

The objective of this study was to study the morphology of azuki bean (Vigna angularis Willd.) Seeds by means of image analysis and to associate them with their physiological potential. The morphological characteristics of two hundred seeds of the 2015 crop were studied and stored in the active germplasm bank (BAG) of cowpea of the Federal University of Ceará (UFC). For the capture of the images, the seeds were distributed in a similar position, in numerical order and equidistant in blue background and made by inverted scanner in aluminum box for later analysis. Soon after the recording of the images, the seeds were submitted to the germination test, the first count being done on the fourth day and the final count on the tenth day after the beginning of the test, for both days the images of the seedlings were captured by the same tool used for seed images. The images of the seeds and seedlings were analyzed in the ImageJ® software, being evaluated characteristics of area, width, length, color, perimeter and aspect ratio for the seeds and the thread. In order to evaluate the vigor of the seeds, the length of the aerial part and the root system were measured, obtaining the total length of the seedlings, later the seedlings were taken to the greenhouse and the mass of dry matter of their structures was weighed. To evaluate the viability of the seeds, the Tetrazolium test was performed. At the end of the measurements, four classes were formed from the seed area, generating the means of the observed characteristics. The experimental design was the completely randomized unbalanced and the data were submitted to analysis of variance (ANOVA) and comparison of means by means of the Tukey test at 5% probability level. In view of the results obtained, in which class IV obtained seeds with larger area, but class III presented bigger and vigorous seedlings with higher total dry matter mass. It is concluded that it is possible to perform image analysis of seed and seedlings morphology by ImageJ software, evaluating quantitative and qualitative characteristics and, in this way, correlating with their physiological potential.