Classification of Plant Leaf Diseases Based on Improved Convolutional Neural Network

Area of Science:

• Agricultural Science
• Computer Science
• Artificial Intelligence

Background:

• Accurate plant leaf disease identification is crucial for agriculture.
• Manual identification is time-consuming, labor-intensive, and prone to errors.
• Deep learning offers a promising automated solution for disease classification.

Purpose of the Study:

• To develop an efficient and accurate deep learning model for plant leaf disease identification.
• To address limitations of traditional convolutional neural networks, such as long training times and large model parameters.
• To improve the accuracy and reduce computational cost in plant disease classification.

Main Methods:

• An improved convolutional neural network (CNN) architecture was proposed.
• The model integrates Inception modules for multi-scale feature fusion and Squeeze-and-Excitation (SE) modules for feature recalibration.
• A global average pooling layer replaced fully connected layers to reduce model parameters.

Main Results:

• The proposed deep learning model achieved a high accuracy of 91.7% on the test dataset.
• Significant reductions in model parameters and training convergence time were observed compared to traditional CNNs.
• The model demonstrated effective feature extraction for classifying diseased plant leaf areas.

Conclusions:

• The developed deep learning method is feasible and effective for plant leaf disease identification.
• The enhanced CNN architecture offers a superior alternative to traditional methods for agricultural disease diagnosis.
• This approach contributes to more efficient and accurate crop monitoring and management.