Comparison of non-destructive methods for estimating the leaf area of Cinchona officinalis L. using digital image processing

Abstract

Cinchona officinalis it is an important plant species and was the only treatment for malaria for over three centuries. The aim of this study was to compare the accuracy of four non-destructive digital image processing methods (LeafArea and three ImageJ algorithms)for estimating the leaf area of young C. officinalis plantations under two establishment conditions: forest stands and enrichment strips. Leaves were photographed at a distance of 8 cm using a 24 MP smartphone and processed with the evaluated methods. Statistical analysis included box and whisker plots, Pearson correlation, and Friedman test. The results showed that ImageJ methods M3 and M4 had the highest accuracy (r = 0.99), with no significant differences between them, and overestimations detected in M1 and M2. It is concluded that M3 and M4 are fast, low-cost, and highly accurate options for foliar monitoring of C. officinalis in the field.