A Deep Learning Based Model for the Classification of Cotton Crop Disease

Abstract

Plant disease is major factor on the subject of crop yield. Plant health needs daily basis update. Agriculture yield increases with good results if healthy and timely disease diagnosis performed. Artificial intelligence revolution in smart agriculture arise many challenges and opportunities for researchers. Deep learning methods considered smart disease detection and classification technology in Modern agricultural field. The sub-field of Artificial intelligence (Deep Learning) can structure algorithms in layer for development of ANN that can learn and make intelligent decision termed as Deep learning. Deep transfer learning attempts to solve problem of learned model that perform task, modify experimental model for solution of another task. Agricultural crop disease diagnosis becomes easy with several soil, weather text data and plant parts images. In computing the terms, “Smart agriculture or Smart Agro-industry or industry 4.0” termed interchangeably used in scientific community. This study is about detailed deep learning, smart agriculture and related research. Study covers Deep learning applications in agriculture, architecture model and some important model depiction. EfficientNet Model proposed for single label image patches. The data set obtained from different cotton crop in the region. Total no:of leaves images 20000 included for the experimentation. EfficientNet model used for disease classification, the common evaluation matrices such as precision, recall and F1-score used. The results shows recall, precision and F1-score of 0.88%, 0.89% and 0.89%, respectively.