Biometria de feijão caupi por meio de análise de imagem digital

Abstract

Seed biometrics is a widely used variable in distinguishing the physiological quality of different species, however, conventional biometrics methods require time and human resources and are susceptible to measurement errors. This characterization can be achieved through digital images, involving the extraction of information about objects captured in a digital image based on certain characteristics such as color, texture, etc. Therefore, the objective of this study was to characterize the biometrics of cowpea (Vigna unguiculata L.) seed lots using digital image analysis. The experiment utilized five lots of cowpea varieties. Vigor assessment included germination tests, first Count of germination, seedling emergence, seedling length, dry mass weight, 1000-seed weight, and seed biometrics (length, width, and length-to-width ratio). Image acquisition was performed using an inverted flatbed scanner (HP Scanjet 2004®). Data collection from the images was conducted using a plugin (SeedsAnalyser) belonging to the ImageJ software, evaluating parameters of morphological characteristics. The plugin enabled the extraction of data for all evaluated parameters, which were consistent with manually obtained data. Thus, with the use of image analysis, it was possible to perform the biometrics of cowpea seed lots with good precision.