Artificial Driving and Spectral Graph Theory for Automated Plant Leaf Disease Classification

Having a certified plant pathologist visually inspect the diseased leaves is now the gold standard for disease diagnosis in plants. This is because symptoms are present in the vast majority of plant illnesses. Because manual disease diagnosis is slow and results are highly dependent on the skill of the pathologist, this is an ideal domain for computer-aided diagnostic tools.Step one of the suggested methods are preprocessing, step two is feature extraction, and step three is training the model. Pictures are acquired, then processed, enhanced, resized, segmented, and finally converted from RGB to HSI as part of the preprocessing phase. The topic of this study is the extraction of numerous features for the purpose of leaf disease categorization. Hexagons of colors are retrieved for classification purposes. The model was trained using Spectral Graph Theory. On average, this state-of-the-art technique achieves a higher accuracy rate of 92.48% than CNN and SVM.