Plant water-stress detection by digital imaging using the entropy concept: application to menthe plant (Mentha viridis)

Abstract

The purpose of this work is to demonstrate the capability of digital imaging within the visible range of the electromagnetic radiation to detect early signs of water stress in plants. The approach is based on colour information and on using the entropy concept. The method is composed of two main steps. The first one classifies pixels of the image that contains plant and soil into plant and background pixels. The classification is performed by a hierarchical analysis method by using a two-dimensional (2D) histogram algorithm. The second one introduces a non-destructive water stress indicator using the entropy concept. The method is applied to detect the extent of modification in menthe plants occurring under water-stress conditions. A physiological study of the water status of the plant is undertaken in order to compare the results obtained by the proposed approach. The results show that the proposed methodology is able to detect the signs of water stress in the menthe plant.