Classifying Corn Leaf Diseases using Ensemble Learning with Dropout and Stochastic Depth Based Convolutional Networks

Widespread plant illnesses have a negative influence on crop production, so it is important to diagnose them timely otherwise there is possibility that they can spread speedily and dramatically lower crop output. Corn is the second-most prevalent cereal grain grown for human use, and several societies throughout history have depended on it. Therefore, discovering corn leaf disease too late can result in significant loss. In our work, we looked into how deep learning works and made a convolutional ensemble network to help the model find corn lesion features better. To better recognize plant disease categories, we employed ensemble learning to ensemble three convolutional neural networks (CNN) which has a custom CNN with dropout, a custom CNN with Stochastic Depth, and DenseNet201. The ensemble technique combines the power of three different networks, resulting an optimal performance. On the corn leaf diseases images of Plant Village dataset, the proposed technique averaged 98.36% accuracy. The experimental results validate the suggested approach and show that it outperforms the current state-of-the-art.