CNN-based leaf disease detection for rooftop gardening using multi-species image segmentation

Urban Farm DB summary

This study developed a machine learning-based leaf-disease detection system to help rooftop gardeners and small-scale growers, aiming to address plant-disease spread and delayed diagnosis that leads many gardeners to abandon increasingly popular rooftop gardening in space-limited urban areas. Using image segmentation as a preprocessing step followed by a classification model, and trained and tested on leaf images from five commonly cultivated species -- guava, jamun, lemon, mango and pomegranate -- the system achieved an overall accuracy of 1.00, with lemon and mango disease classification reaching accuracies of 0.995 and 0.991 respectively and F1-scores exceeding 0.88.