Zooming Algorithm Based Edge Detection Techniques for Plant Leaf Disease Identification using Convolutional Neural Networks optimized with Battle Royale Optimization

Abstract

Agriculture is a very important sources for every living things but it is affected with various issues like plant diseases. To predict the plant leaf diseases in an early stage is very necessary in the agriculture field, which improve the productivity. Thus, to improve the effectiveness of the plant leaf disease identification process and classification, a Deep Learning (DL) based optimization approach was developed. In this manuscript, a Zooming Algorithm based edge detection model is used for identification and Convolutional Neural Networks optimized with Battle Royale Optimization (CNN-BRO) is proposed for classification.  The implementation of this work is done by MATLAB and the parameters are calculated. Thus, the proposed CNN-BRO approach was attained 25.34% ,15.45% high accuracy, 28.5% , 11.46%  high specificity, 27.97% ,10.25%  high sensitivity, 30.22% ,13.95%  high recall, 16.38% , 9.1%% high precision, 25.27%, 13.21%  high F-score and  59.99%, 42.86%  lower system error than other existing methods like  Support Vector machine with exponential spider monkey optimization (SVM-SMO) and Bacterial Foraging optimization based Radial Basis Function Neural Network (BRBFNN-BFO).