Image assessment tool for plant health and growth tracking with structural similarity and vegetation index

Abstract

Software supporting the measurements of leaf area using digital photographs requires a frame reference such as a manual ruler scale to derive the surface areas. Two wellknown image software are ImageJ and EasyLeafArea. Each software uses a thresholdbased approach in pixel count, coupled with calibration to avoid perspective distortion, before deriving the surface areas. However, both are unable to provide an assessment of the leaf’s health. Assessing the health of the leaves is done through the observation of any discolouration. In this study, an assessment tool is created to calculate the plant’s surface area as well as monitor the health of the leaves. Aruco markers are used as reference to calculate leaf surface area. Perspective Distortion errors present in the images are taken into consideration through the usage of a clear and consistent pattern. Images are processed and transformed from Red, Green, Blue (RGB) to Hue, Saturation and Value (HSV) colour space. The health of the leaves is monitored by detecting discolouration using HSV. Finally, the assessment tool can track the growth of the leaves through contour detection and provide information to users on the conditions of the plant’s health.