Summary of Leaf-based plant disease detection systems: A compilation of systematic study findings to classify the leaf disease classification schemes

Abstract

Food production is the mainstay of every economy, and relies on the yield of the agricultural producers. This yield is often affected by micro and macro diseases that occur throughout a given fruit-bearing plant's growth. For example, some of the diseases affecting banana leaf include Fusarium oxysporum, Mycosphaerella musicola, Gloeosporium musae, Erwinia Carotovora, Pseudomonas Solanaceanim, Pentalonia nigronervosa, Erionota thrax, BSV and BBM Virus. Such pathogens are countless, so advances in image processing are deemed necessary to categorize and suggest remedial measures for those viruses. A sizeable proportion of work conducted in the area aims at linear approaches where classification, attribute extraction, and definition are conducted to detect aesthetically detectable ailments. However, these interventions are not equipped for bigger and more diverse sets of data, so machine learning and artificial intelligence-based approaches such as Q-learning, re-enforcement learning, etc. take responsibility for formulating diseases which were invisible to the human eye in every component of the processing layers. Because of such a wide variety of diseases and such a large number of processing algorithms, system designers are often misleading as to which algorithms combination should be used to classify which kind of diseases. To eliminate this confusion, this paper compare and objectively evaluate some of the latest existing techniques in this field, and evaluate the best fusion of algorithms that can be used to develop a highly accurate classification system for leaf disease. This finding suggests plenty more research directions which must be undertaken to enhance system efficiency.

Topics & Concepts