Related Experiment Video
Updated: Aug 4, 2025
03:31
Author Spotlight: Enhancement of Salient Object Detection for Smart Grid Applications
Published on: December 15, 2023
592
Brinjal leaf diseases detection based on discrete Shearlet transform and Deep Convolutional Neural Network.
S Abisha1, A M Mutawa2, Murugappan Murugappan3,4,5
1Department of Electronics and Communication Engineering, Rohini College of Engineering and Technology, Nagercoil, India.
Plos One
|April 5, 2023
Summary
This study introduces a novel method using Deep Convolutional Neural Networks (DCNN) and Radial Basis Feed Forward Neural Networks (RBFNN) for accurate brinjal leaf disease identification. The DCNN model achieved a high accuracy of 93.30% in classifying various plant diseases.
Area of Science:
- Agricultural Science
- Computer Science
- Plant Pathology
Background:
- Accurate plant disease diagnosis is crucial for crop yield.
- Manual identification of plant diseases is time-consuming and often relies on microscopic symptoms.
- Early detection of diseases in crops like brinjal is essential for effective management.
Purpose of the Study:
- To develop an automated system for identifying and classifying brinjal leaf diseases.
- To compare the performance of Deep Convolutional Neural Networks (DCNN) and Radial Basis Feed Forward Neural Networks (RBFNN) for disease classification.
- To improve the accuracy and efficiency of plant disease diagnosis in agriculture.
Main Methods:
- Collected 1100 images of diseased brinjal leaves (five species) and 400 healthy leaf images.
- Applied image preprocessing techniques including Gaussian filtering and Expectation-Maximization (EM) segmentation.
- Utilized discrete Shearlet transform for feature extraction (texture, color, structure) and DCNN/RBFNN for classification.
Main Results:
- Deep Convolutional Neural Networks (DCNN) achieved a mean accuracy of 93.30% with feature fusion.
- Radial Basis Feed Forward Neural Networks (RBFNN) achieved 87% accuracy with feature fusion.
- DCNN outperformed RBFNN in classifying brinjal leaf diseases, especially with integrated features.
Conclusions:
- The proposed DCNN and RBFNN models offer an efficient and accurate method for brinjal leaf disease classification.
- Automated image analysis using DCNN shows significant potential for early disease detection in agriculture.
- Feature fusion techniques enhance the performance of neural networks in identifying plant diseases.
