Disease Detection Techniques in Plants: Transition from Manual to Automation

Abstract

India is an agricultural land with substantial agricultural poverty. The plant diseases impose severe threat to the crop yield, productivity and sustainability each year. In early days, observational methods were adopted by experts and with the evolution of techniques, cultivators started sharing images of plants over distance and the distribution of knowledge was speeded with limited access. The rapid and accurate estimation of plant diseases is still an unmet need of agriculture and led researchers to train and test computer-aided deep learning techniques like ANN, CNN. These techniques are fast, precise and consistent with quantitative information. Plant epidemiologists are adopting and preferring automated disease detection techniques over previous techniques to save cultivators from stressful, time-consuming, and laborious disease detection methods. This is the very first review where we have shown how and why the transition of plant disease detection techniques from naked-eye to automatic detection techniques happened? In this review, we have tried to compile the most recent literature available with special emphasis on deep-learning techniques w.r.t. plant diseases.