Plant Disease Classification Using Fusion Cauchy Reverse Learning Strategy Based Orca Predation Algorithm with DenseNet-121

Plant diseases represents specific spots with various colors and locations, that have a significant impact on leaves lead to affect the plants and reduced the crop yield. Thus, an accurate plant disease classification is essential corresponding to the distinct spot shape, size, etc. However, the existing Deep Learning (DL) models have several limitations that make the model difficult to distinguish between various plant diseases due to similar characteristics. To overcome this limitation, a Fusion cauchy reverse learning strategy based Orca Predation Algorithm (FOPA-DenseNet-121) is proposed to identify and classify the plant leaf disease accurately. The proposed DenseNet-121 model has the advantage of feature reusability that helps to learn the information about various diseases effectively. Also, the FOPA based hyper parameter tuning model adjusts the DenseNet-121 model parameters to learn the difference between various diseases which helps to differentiate the diseases and enhance performance of classification model. Experimental results of the proposed model achieved accuracy of 98.92% which is greater than the existing approach like Visual Graph Geometry -16 (VGG-16).