Iron Deficiency Diagnosis Using Combined FMM-based Features and SVM on Lemon Leaf Images

Diagnosing iron deficiency in the lemon plant is one of the early ways to prevent and preserve the plant. Diagnosis of iron deficiency depends on the emergence of green veins in young lemon leaves (buds) as a result of a decrease in the leaf color of the plant (yellow leaf). In this research, it was proposed to use the technique of the Fuzzy Mathematical Morphology FMM and Histogram to extract the features of young lemon leaf images using a database collected from the Botanical Department, Faculty of Science, Misurata University. This database consisted of 210 images that were divided into three categories: leaves with iron deficiency, healthy leaves, and leaves with other diseases. These FFM-based extracted database features were classified using the Support Vector Machine SVM classifier, where 70% of the dataset features were used for training, and 30% for testing. The system's performance was further enhanced by incorporating the Gray Level Co-occurrence Matrix (GLCM) with the FMM-based features. Evaluations showed significant improvement: FMM alone achieved 83% accuracy, FMM with histogram analysis reached 95%, and the combined FMM/Histogram/GLCM approach reached the highest accuracy of 97%.