Estimation of water content on plant leaves using a spectroscopic camera

Abstract

In agriculture, it is important to detect and prevent plant diseases before they develop.This study proposes a method for estimating the moisture content of plant leaves using spectral images for the purpose of non-contact and automatic measurement of plant leaf wetness over a wide area within a farm. In the experiment, we conducted indoor observations using a prototype compact spectroscopic camera, a visible light camera, and a moisture sensor on several plant leaves controlled for the presence of water droplets. The experimental results suggest that the blue values of the visible light camera have a strong correlation with the measurements of the moisture sensor regardless of the angle of the leaf, suggesting that the camera may be suitable for evaluating the water-wet condition of the leaf. Furthermore, when applied to fixed-point observation images in an actual open-field cultivation, there was a possibility of estimating the moisture content of leaves over a wide area.