Mathematical models for the estimation of leaf chlorophyll content based on RGB colours of contact imaging with smartphones: A pomegranate example

Abstract

The objective of this study was to develop a mathematical model for the non-destructive, fast estimation of the leaf chlorophyll (Chl) content of pomegranate trees. For this reason, contact images of the leaf samples were firstly captured with smartphones and the RGB colours of the images were used for the estimation of the leaf Chl contents. Here, different methods were used for the contact imaging. In the present study, two closed boxes with a small hole (equal to the dimensions of a smartphone camera) on each were formed. Samples were inserted into the hole; and a red LED light and white LED light, separately, were passed through the hole and the leaf. Furthermore, a series of models were tested to best estimate the leaf chlorophyll content of the pomegranate trees by using the RGB colours of contact imaging. Results showed that the use of red LED light sources, instead of white LED light sources, during contact imaging, provides a better estimation of the leaf Chl content. Results also suggest that colour values are highly related to the total weight of the contact imaging area. According to the results obtained, the best estimation of the leaf Chl content (of a given area) is possible by using both the G and B colour values with multiple regression models. It is also found to be important to use the weight of the sampled area for the estimation of the leaf chlorophyll content in mg ∙ g−1.

Keywords

• colour index
• computer programs
• fast estimation
• multiple regression analysis
• plant health
• sustainable farming