Morphological variation on tomato leaves due to different nitrogen contents

Abstract

Elliptic Fourier Analysis (EFA) is a method used to quantify shape differences.  It mathematically describes the entire shape of an object by transforming the contour into Fourier coefficients, used as variables for statistical analysis, and involving the fitting of some type of curve to the object outline.  Generally, the shape of agricultural products such as fruit, vegetables, grain and in addition other organs of plant is one of the most important factors for their classification and grading in relation to commercial quality and organoleptic properties.  The aim of this study is to quantify the morphological variation of the shape of tomato leaves in response to their different nitrogen (N) content using the EFA coefficients, the fractal geometry and the perimeter ratio in combination with the Partial Least Squares Discriminant Analysis (PLS-DA).  The analyses were realized on a tomato crop where each sample was chemically analyzed at the laboratory to establish the N content.  The leaves (168) were divided into 3 groups following different N concentrations.  Results suggest no relation between leaves lengths and N concentration is present following the Kruskal-Wallis performed with a p=0.735.  The PLS-DA performing on the EFA coefficients, fractal index and perimeter ratio shows a high sensitivity, sensibility, and reduced mean classification error (82.3%, 81.07% and 18.3% respectively).  The percentages of the correct classification in the model resulted to be 69.29% while the independent test equal to 56.1%.  This study demonstrated the relation between leaf shape and N content (expressed in 3 concentration groups).

Keywords: Tomato leaf, elliptic Fourier analysis, fractal index, perimeter ratio, partial least squares discriminant analysis