Enhanced pearl millet mildew disease detection using ensemble deep learning methods

Millet crops play a crucial role in global food security, providing sustenance to millions of people worldwide. Mildew disease poses a significant threat to pearl millet, a staple crop in many regions, impacting both its quality and yield. Detecting diseases in millet crops is crucial for maintaining both the quality and quantity of agricultural yields. However, limited labelled data and the expense of manual data labelling pose significant challenges in this domain. To address these issues, we suggest a deep learning ensemble framework that utilises the potential of multiple models for enhanced disease detection accuracy. Ensembles integrate the strengths of individual deep-learning models to improve overall performance and robustness. DenseNet121 and ResNet50, two deep-learning models, were selected as the base models in our ensemble. Preliminary experimental results demonstrate the effectiveness of our ensemble approach, with an impressive accuracy of 96.6%.