Plants Disease Classification using Deep Learning

Abstract

Early detection of plant disease is useful in reducing its rapid spread; however similar visual appearances of different plant diseases make it a challenging problem. In the proposed approach, we improve the performance of plant disease detection by learning the fine differences in the visual appearances of these different diseases. We used pre-processing, data augmentation, and deep learning for the classification of different categories of diseases in plants. The representation of minority classes with fewer images is improved using DC-GAN. Different CNN based deep learning techniques are applied for classification. The performance comparison of the proposed approach with existing approaches on a publicly available plant village dataset shows its superior performance with an accuracy of 97.2% and an F1 score of 0.97 for incorrect predictions of different plant diseases.