Caracterización de imágenes hiperespectrales para identificar cambios debido a agentes contaminantes en plantas mediante análisis de datos

Abstract

In recent years, there has been increasing interest in the characterization of physicochemical properties in plants, with the purpose of understanding and improving the specific needs of various crops. However, a disturbing issue is the accumulation of contaminants in plants due to industrial activity; therefore, in response to this concern, the proposal to use technologies based on image and data analysis to detect contaminants in crops by observing morphological changes in contaminated plants has been proposed. Multispectral imaging (MSI) and hyperspectral imaging (HSI) approaches are used to obtain accurate spectral information of objects, which allows determining unique optical properties assigned to each pixel. The present work focuses on the normalization of morphological changes, specifically color variations in Calendula Officinalis and Arvensis plants when exposed to the fungi Sphaerotheca pannosa and Entyloma serotinum , which cause changes in the color of plant leaves. To carry out this characterization, an algorithmic method was developed and applied to discriminate and analyze the information in images (HSI and RGB) using colorimetric parameters, such as chroma, hue and brightness. This made it possible to categorize and differentiate the contaminated sections of marigold plants.