Improving Wheat Yield Protection with Deep Feature Fusion for Estimating Yellow Rust Severity Using CNN Models

Wheat is one significant staple crop that is under severe threat from diseases like yellow rust, which might cause it to lose a significant portion of its yield. Early detection and severity evaluation of the illness are critical to the effective use of insecticides and disease management. This article presents a method for estimating the severity of yellow rust in wheat leaves by combining deep characteristics from two convolutional neural network (CNN) models. To create deep features from photos, VGG16 and MobileNetV2 are first trained. After using generic fusion rules to fuse these characteristics, support vector machines (SVM) are employed for classification. The Yellow-Rust-19 dataset is used in experiments to show that combining deep features increases accuracy by about 0.4% when compared to using separate CNN models. The findings highlight how deep feature fusion may improve disease severity estimation, which will help with crop protection and disease management.