Leaf Disease Classification of Various Crops Using Deep Learning Based DBESeriesNet Model

Chilli is one of the world’s most extensively used spices. Numerous diseases affect chilli leaves as a result of climatic and environmental changes, which limit crop output. Chilli leaf disease drastically reduces the quality and yield of the chilli crop. However, after conducting a thorough analysis, we discovered that there are not many studies that compare the classification performance of chilli leaf disease using artificial intelligence (AI) approaches. In this proposed work, five main leaf diseases of chilli crops are classified using a Deep Batch-normalized ELU SeriesNet (DBESeriesNet) model, which provides better accuracy with less number of parameters. Various optimization functions, activation functions, and mini-batch sizes are used to further assess the method on a dataset of chilli leaves. With 3.4M parameters, the DBESeriesNet model generates an accuracy of 99.80% for the dataset of chilli leaves. To further test the proposed model’s efficiency, the various crop datasets from PlantVillage data site were used and it produced 100% accuracy for the categorization of apple, cherry, grape, peach and potato datasets. Additionally, the accuracy of the datasets maize, pepper and tomato are, respectively, 98.85%, 99.60%, and 99.36%. With an accuracy of 99.62%, the proposed DBESeriesNet classifier could able to classify 44 distinct classes of plant leaf datasets.