A Novel Deep CNN Model with L1 Regularization for Plant Disease Classification

Abstract

In the modern era of agriculture, the early detection and classification of plant diseases are crucial for ensuring crop health and maximizing yield. Traditional methods of disease detection, which rely heavily on manual inspection by experts, are time-consuming, expensive, and prone to human error. This research leverages the capabilities of deep convolutional neural networks (DCNN) and introduces a novel deep convolutional neural network named LDCNN. The proposed approach encompasses feature extraction, feature selection, and a DCNN. DenseNet-201 classifier is utilized for feature extraction, and L1 regularization is employed for feature selection. The selected features are used to train ResNet-9 through extensive experiments. The results demonstrate that LDCNN surpasses other methods in literature in terms of accuracy, precision, recall, and F1-score.