SOM neural network for clustering plant and plant stress based on kinetic imaging of chlorophyll fluorescence

Chlorophyll fluorescence (ChlF) is a useful tool to detect plant stress in a non-destructive way. A method via SOM based on kinetic imaging of chlorophyll fluorescence is described, that allows to screen and image the dynamic change in chlorophyll fluorescence intensity by illuminating a dark-adapted leaf with blue-LED(460nm) and acquired by a EMCCD in 210fps, 20fps, 2fps individually. The shape of the fluorescence induction curve is characterized by several key inflection points: I-P — M'-S-M-T. We cluster and classify plant and plant stress by these key inflection points using a self organization map(SOM)neural network. Experiment show that this method can classify plant and plant stress.