An Efficient Plant Disease Identification Using Intense Transfer Learning Empowered Efficient Net with Kernel Extreme Learning Machine

Abstract

Mechanized detection and cataloging of plant diseases still remain a dynamic research period. A quick and accurate plant disease identification model can support to progress in economic crop fortification on a small scale and food safety on large measure. In addition, computer vision and deep learning (DL) methods pave the way for monitoring the health state of the plants and detect the diseases during early period. In this part, this study enterprises an intelligent plant disease diagnosis using deep transference learning based EfficientNet with Kernel Extreme Learning Machine (KELM), named DTLEN-KELM model. The projected DTLEN-KELM model includes the design of median filtering (MF) and contrasts limited adaptive histogram equalization (CLAHE) as a preprocessing method. Besides, the DTLEN-KELM model contains an EfficientNet B0 based feature extractor to derive optimal feature vectors which are then classified using KELM model. The deliverable of the DTLEN-KELM technique is validated utilizing a standard dataset and the experimental results highlighted the improved disease diagnostic performance over the current methods.