Real-Time Detection of Crop Leaf Diseases Using Enhanced YOLOv8 algorithm

Abstract

Agriculture is a mainstay of the Moroccan economy and the breadwinner for millions of farmers, offering them a broad array of crop varieties. However, a lack of resources and expertise renders it difficult for farmers to properly diagnose plant diseases. Consequently, valuable time and resources are often wasted trying to save diseased crops. To tackle this issue, a revolutionary method that uses the most recent breakthroughs in computer vision and deep learning technology to identify plant diseases in real-time has been presented. Our system employs YOLOv8 (You Only Look Once), an advanced object identification approach that analyses leaf images at a rate of 70 FPS (Frames Per Second). This innovative image analysis technique consists of splitting an image into numerous grid cells and uses a single neural network to predict the bounding box coordinates and class probabilities. This leads to a more efficient and accurate image assessment, resulting in quicker and more exact disease identification. Our results demonstrate the exceptional efficiency of YOLOv8, which outperforms conventional object identification algorithms in both speed and accuracy. This solidified YOLOv8's position as a leading solution for object detection and recognition in real-world scenarios.

Topics & Concepts