Electrophysiological assessment of plant status in e-farming systems: AI and machine learning

Abstract

Agriculture has been an inseparable part of society, providing necessary sustenance for life. However, Engineering in agriculture has been focused on the efficiency of labor and not improve the health of the plant itself. The problem which this project aims to solve is the assessment of the growing status of plants. The growing status of the plant has always been obscure to us and hard to quantify. The objective of this project is to analyze the potential relationship between the electrophysiological property of plants with the growing status of plants. To achieve this goal, an experiment that measures the electrophysical property of plants that grow under different environments is conducted. The experiment used plants with a short life cycle that lasted for 1 month during which 4 measurements are conducted on a weekly basis. Preliminary analysis is conducted to explore the potential relationship by contrasting the mean value for each group that grows under different conditions. Comparison between each group is conducted in two aspects across 4 weeks, which are the magnitude and phase of the bio-impedance of the plants. In addition, the potential bias for the size difference of leaves is explored and considered in the analysis by conducting an experiment that measures bio-impedance on different leaves of different sizes. Finally, a database is constructed with measured data. And Linear SVC is used as the classification model which is trained to classify different growing conditions of the plants based on the collected data.