Mango Leaf Disease Detection and Classification Using Spatial Attention Enabled Ensemble Classification

Leaf diseases are growing to be a concerning threat which impacts both the biological conditions and economic conditions of human beings. The poor cultivation would result in world hunger as well as increase in cost of living since reduction in supply would lead to increase in demand especially with necessity commodities. Mango being a seasonal crop with high value would face the same issue as well. Therefore, this research work proposes an ensemble learning classifier which is based on spatial attention that incorporates two deep learning classifiers namely the VGG-16 model and EfficientNetV2-BO model for immediate identification as well as classification of various mango leaf diseases such as Cutting Weevil, Gall Midge, Die Back, Bacterial Canker, Powdery Mildew, Anthracnose and Sooty Mould. The proposed ensemble model classifies effectively resulting with high evaluation scores of 97.13%, 97.53%, 97.33% and 97.38% in terms of Accuracy, Precision, Recall and Fl-Score respectively. These results are better than ResN et50, TL-based diagnosis using AlexNet, ANN, AlexNet and VGG16 models.