Identification of Plant Leaf Disease Using a Novel Convolutional Neural Network

Abstract

A critical element in preventing a major outbreak is the detection of plant leaves. An important research issue is the automatic detection of plant diseases. For both human life and condition, a plant's dedication is essential. Like humans and other animals, plants do suffer the negative impacts of illnesses. A plant's normal development is influenced by the frequency of plant diseases that occur. The entire plant, including the leaf, stem, organic material, root, and flower, is affected by these diseases. Most of the time, if a plant's ailment is not treated, it dies or may cause leaves, blooms, organic products, and so forth to fall off. For accurate identification and treatment of plant diseases, appropriate determination of these disorders is necessary. Plant pathology is the study of plant infections, their causes, and methods for preventing, managing, and eradicating them. However, the current approach includes human inclusion for structure and identifying disease proof. This tactic is time-consuming and expensive. Instead of using the current method, a programmed division of diseases from plant leaf images utilising a delicate registration methodology may be more beneficial. In this study, we describe a method for identifying and characterising plant leaf diseases naturally called Bacterial Searching Improvement Based Radial Basis Function Neural Network (BRBFNN). We use bacterial search streamlining (BFO), which increases the speed and accuracy of the system to recognise and organise the regions contaminated by diverse illnesses on the plant leaves, to assign Radial Basis Function Neural Network (RBFNN) the proper weight. The location development calculation increases the system's efficiency by searching for and gathering seed focuses on typical traits for the highlighted extraction operation. To make progress against parasite diseases including early curse, leaf twist, leaf spot, late scourge, and basic, cedar apple, and leaf rust. The suggested approach achieves more accuracy in identifying evidence and characterizing infections.