A Review on AI-Driven Intelligent Decision Support Systems for Fruit Disease Diagnosis Based on Taxonomic and Phytochemical Analysis

Abstract

The global agricultural sector faces persistent economic threats from fruit pathogens, which compromise both yield stability and produce quality. Traditional diagnostic protocols largely dependent on manual visual inspection or resource-heavy laboratory testing often fail to provide the rapid, scalable interventions required by small-scale producers. While recent computational models have introduced automated detection, most current frameworks rely solely on image-based patterns, neglecting the biological depth offered by botanical taxonomy and biochemical signaling. This study introduces an integrated Decision Support System (DSS) that bridges this gap by synthesizing machine learning architectures with taxonomic classification and phytochemical profiling. By utilizing a hybrid dataset of multi-spectral imagery and chemical biomarkers, the proposed model identifies disease markers that are often invisible to standard computer vision. Evaluation across diverse orchard environments indicates that this multi-modal approach significantly enhances diagnostic reliability compared to unimodal systems, achieving high scores in precision-recall and F1-metrics. Ultimately, the integration of phytochemical indicators provides a more transparent, biologically grounded pathway for sustainable disease management, offering a robust tool for real-time agricultural decision-making.