Exploring the Potential of Convolutional Neural Networks in Automatic Diagnosis of Dragon Fruit Diseases from Plant Photographs

Abstract

Numerous instances of dragon fruit diseases over the years have caused an estimated 14% annual crop loss globally, affecting untold millions of people. The study of dragon fruit illnesses, or pathology, aims to increase plants’ chances of surviving in the presence of pathogenic parasite microbes and poor climatic conditions. Environment-related elements that influence the development of dragon fruit diseases include temperature, pH, humidity, and moisture. Misdiagnosis can result in chemical overuse that costs money, environmental imbalance and pollution, and the establishment of pathogen strains that are resistant to treatment. The process of diagnosing diseases nowadays is time-consuming, expensive, and dependent on human scouting. Compared to the current method, automatic disease segmentation and diagnosis from dragon fruit photos can be somewhat useful. This study uses Convolutional Neural Network (CNN) models, to accurately identify four kinds of apple plant diseases from photographs of apple plant leaves. “Healthy,” “scab,” “rust,” and “many illnesses” are among the classifications.