Convolutional Neural Networks for the Identification and Forecasting of Potato Leaf Disease

In order to diagnose potato leaf diseases, this research suggests a hybrid model that employs a Random Forest classifier for sample class classification and a CNN for photo attribute extraction 1885 samples of various potato leaf diseases brought on by fungi, viruses, pests, bacteria, Phytophthora, nematodes, and healthy leaves are included in the dataset that Kaggle was able to collect. Resizing the photos to a uniform size and splitting the dataset into equal training, testing, and validation sets are essential steps in a comprehensive preparation procedure. Softmax activation is used for classification in the CNN model, while ReLU activation is used for normalization. Its five layers have increasing filter diameters. For the final label categorization, CNN features are fed into a Random Forest classifier, which has 100 decision trees. The suggested model's accuracy and effectiveness in classifying illnesses show promise. A 77% overall accuracy rate and a 44% to 96% precision range for several disease categories indicate that the hybrid model suggested for potato disease categorization is producing encouraging results.