EVALUATION OF PLANT HISTOLOGY BY AUTOMATIC CLUSTERING BASED ON INDIVIDUAL CELL MORPHOLOGICAL FEATURES

Abstract

A procedure has been developed for the automatic clustering of plant cells observed by confocal microscopy. The contribution of cell morphological features to reveal histological regions has been investigated. Several adjacent images were acquired to visualise a representative region of the sample and a mosaic image was built. The cell size and shape and the cell wall thickness were quantified. The extracted features were used to automatically classify the cells into morphological groups. The technique made it possible to split the cell population into 8 groups mainly corresponding to histological regions of beet root.