Related Experiment Video
Updated: Jun 6, 2025
08:14
LeafJ: An ImageJ Plugin for Semi-automated Leaf Shape Measurement
Published on: January 21, 2013
28.3K
Early detection of plant leaf diseases using stacking hybrid learning
1Department of Computer Science, College of Engineering and Computer Science, Jazan University, Jazan, Saudi Arabia.
Plos One
|November 22, 2024
Summary
Early crop disease detection is challenging. This study uses ensemble deep learning on leaf images, achieving 99.75-100% accuracy for automatic disease identification, improving upon manual methods.
Area of Science:
- Agricultural Science
- Computer Science
- Plant Pathology
Background:
- Manual crop disease and pest identification is time-consuming, costly, and environmentally impactful due to pesticide overuse.
- Current methods like sticky traps require manual analysis, posing limitations for large-scale field monitoring.
Purpose of the Study:
- To develop an automated system for early crop disease identification using image processing and machine learning.
- To enhance the accuracy and efficiency of disease detection compared to traditional manual methods.
Main Methods:
- Utilized a hybrid learning approach combining scratch and transfer learning strategies.
- Employed an ensemble of Convolutional Neural Networks (CNNs) trained on the Plant Village dataset.
- Images of fruit plant leaves, including healthy and diseased samples, were processed for classification.
Main Results:
- Achieved high accuracy rates for automatic disease identification, ranging from 99.75% to 100%.
- Demonstrated the effectiveness of ensemble CNN models in categorizing diseases based on image attributes.
Conclusions:
- Automated crop disease detection using image processing and ensemble CNNs offers a highly accurate and efficient solution.
- This approach can significantly aid in early pest and disease identification, potentially reducing environmental impact from pesticide application.
