Development of Efficient Deep Convolutional Neural Network for Prediction of Leaf Disease Identification in Agriculture

Agriculture is the real backbone of an economic growth, plays a critical part in any country's economic growth and development. In many cases, agriculture production may be reduced, thereby reducing the nation's economy, because crops may be harmed by diseases. Diseases generally affect the leaves of the crop and should be identified early on so that the quality and quantity of the produce can be increased. The Deep convolutional neural network model is being implemented on a Kaggle benchmark dataset of tomato plant leaves. Gamma correction, images flipping, noise injection, scaling, color augmentation, & rotation were the five data augmentation methods used. Our experiments with two datasets, augmented and non-augmented, and found that data augmentation can improve model performance. The proposed model's training contains batch size, training epochs and dropouts. When using validation data, the suggested model outperforms other common learning algorithms. After extensive simulation, the proposed model has achieved 97.49% classification performance. The precision of the suggested task exceeds that of typical machine learning methodologies. Consistency and dependability are also examined for the suggested model.