Identification of Leaf Constant Values in Manalagi and Madu Mango Cultivars Based on Digital Image Processing

Accurate leaf area measurement is crucial in plant physiology studies; however, conventional methods are often impractical and costly. This study aims to identify the leaf constant values of Manalagi and Madu mango cultivars using a digital image processing approach. A descriptive quantitative study was conducted at the Agrotechnology Laboratory of Pekalongan University using 40 healthy leaf samples from each cultivar. Leaf area was measured using ImageJ software, while leaf length and width were manually measured to calculate the leaf constant based on the Montgomery equation. The measurement data were analyzed using descriptive statistics, boxplots, linear regression analysis, and accuracy validation through RMSE, NRMSE, NSE, and Willmott’s index of agreement (d). The results showed that the average leaf constant value was 0.706 for Manalagi and 0.779 for Madu, with homogeneous data distribution and no outliers. The correlation between the measured leaf area and the predicted leaf area was very strong, with R² values of 0.9947 for Manalagi and 0.9992 for Madu, along with very low prediction errors (NRMSE of 0.015 for Manalagi and 0.009 for Madu). Moreover, the NSE and Willmott’s index values approached 1, indicating excellent model performance. These findings demonstrate that the derived leaf constants can be used as practical references for field leaf area estimation, contributing to more efficient agronomic research and horticultural crop management.