Machine Learning based estimation of Chlorophyll and Flavonoid content in Bitter Leaf using Color and GLCM Texture Features

The prediction of plant biochemical content is critical for applications in both agriculture and pharmacology. This study presented a non-invasive approach that integrates Gray Level Co-occurrence Matrix (GLCM) texture analysis with Artificial Neural Networks (ANNs) to estimate chlorophyll and flavonoid content in Vernonia amygdalina (bitter leaf). Leaves at three developmental stages were analyzed using standard spectrophotometric methods for validation. GLCM features (contrast, correlation, energy, and homogeneity) were extracted from normalized leaf images, and feature selection techniques were employed to identify the most predictive variables. An optimized ANN, trained and validated using these selected features, demonstrated high predictive accuracy without relying on destructive sampling methods. This integrated GLCM-ANN framework offers improved efficiency, reduced costs, and enhanced scalability compared to conventional approaches. Its potential applications extend to precision agriculture enabling real-time monitoring of plant health, early stress detection, and optimized nutrient management as well as to the pharmaceutical industry for improving the quality control of herbal medicines. Future work will focus on expanding the dataset and incorporating multi-modal imaging to further refine model performance and advance non-invasive plant biochemical analysis. Keywords: Bitter leaf, photosynthesis, non-destructive biochemical analysis, hyperspectral imaging, regression models, Artificial Neural Networks, precision agriculture.