Paradigm of YOLO:

Abstract

Efficient along with precise identification of plant diseases can significantly enhance agricultural productivity and overall crop yield, particularly for lemon trees and their foliage.[24] Algal Leaf Spot, Black spot, Citrus canker, Citrus pest, Citrus Scab, Dry leaf, Greening, Leaf Curl, and yellow spot are diseases that can significantly impair the health of lemon trees and reduce their fruit yield. These diseases typically develop on the leaves, leading to reduced fruit quality and output, sometimes by more than half. Convolutional Neural Networks (CNN) have demonstrated promise in the field of lemon tree disease identification.[25] Nevertheless, there is a lack of research specifically focused on the identification of diseases in lemon trees, particularly in regions such as Bangladesh where lemon cultivation is widespread. Previous experiments in this field have primarily based on secondary data, limiting the depth of knowledge and practical application. In order to fill these knowledge gaps, our research focused on collecting lemon disease leaves from well-known lemon-growing regions, such as Khagan and Ashulia in Bangladesh. I utilized advanced Convolutional Neural Network (CNN) techniques, specifically YOLOv8, to assess its effectiveness in identifying diseases. The trials I conducted highlighted the benefits of YOLOv8 in terms of its speed and accuracy, demonstrating its practicality for identifying diseases in agricultural settings. Furthermore, our review of current literature uncovered a variety of machine-learning methods applied to publicly available datasets for lemon leaf disease diagnosis. By doing our experiments, I obtained a precision confidence curve score of 1.00, indicating a high level of confidence in the accuracy of our data. Various metrics, including precision, accuracy, and recall, were utilized to examine the performance of the model.