FF-PCA-LDA: Intelligent Feature Fusion Based PCA-LDA Classification System for Plant Leaf Diseases

Crop leaf disease management and control pose significant impact on enhancement in yield quality to fulfill consumer needs. For smart agriculture, an intelligent identification system is inevitable for efficient crop health monitoring. In this view, a novel approach proposed using feature fusion PCA-LDA classification (FF-PCA-LDA). Handcrafted hybrid deep features are extracted from RGB images. TL-ResNet50 used extract the features. Fused vector obtained by combining handcrafted After fusing image features, PCA employed select most discriminant LDA model development. Potato as case study validation of approach. The developed experimentally validated potato benchmark dataset. It offers high accuracy 98.20% unseen dataset which was not during training process. Performance comparison technique with other approaches shows its superiority. Owing better discrimination learning ability, overcomes segmentation step. may be automated tool monitoring, control, can extended types.