Modeling of the Leaf Area of Maytenus obtusifolia Mart. from Scanned Images

The leaf has a vital role in the functions of the plant, being responsible for photosynthesis and gas exchange. Thus, the objective of this study was to fit a mathematical equation model to estimate the leaf area of Maytenus obtusifolia Mart. through the linear dimensions of the leaves. For that, six hundred and fifteen healthy leaves were collected from plants belonging to the Federal University of Espirito Santo, Sao Mateus Campus, in the municipality of Sao Mateus, located in the north of the State of Espirito Santo, Brazil. All leaves were digitized and the images processed using the ImageJ® software, obtaining the measurements of the maximum length of the main midrib (L), the maximum width of the leaf blade (W) and the real leaf area (RLA) of each sheet. Subsequently, the product of length and width multiplication (LW) was also obtained. 500 sheets were randomly separated for the generation of models of mathematical equations and their respective coefficient of determination (R2), where RLA was used as dependent variable as function of L, W or LW as independent variable. Based on the models generated, a 115 leaf sample was used for validation, where the L, W and LW values of this sample were replaced in the adjusted equations, thus obtaining the estimated leaf area (ELA). A comparison of the means of RLA and ELA was performed by Student’s t test at 5% probability. We also calculated the mean absolute error (MAE), the root mean square error (RMSE) and the Willmott index (d). The best equation was defined by the following criteria: non-significant values of RLA and ELA averages, R2 and index d closest to unit, and MAE and RMSE values with greater proximity to zero. The quadratic model equation represented by ELA=0.18122798+0.72847767(LW)+0.00002789(LW)2 generated by multiplying the length with the width (LW) is the most suitable for the estimation of the leaf area of Maytenus obtusifolia Mart., in a fast, safe and non-destructive way.