Bayesian optimized CNN ensemble for efficient potato blight detection using fuzzy image enhancement.

Area of Science:

• Agricultural Science
• Computer Science
• Machine Learning

Background:

• Potato blight causes significant agricultural and economic losses.
• Accurate and early detection of potato blight is crucial for crop management.

Purpose of the Study:

• To develop a highly accurate deep learning model for potato leaf blight detection.
• To optimize Convolutional Neural Network (CNN) models using Bayesian optimization and ensemble learning.

Main Methods:

• Trained multiple CNN architectures (ADAM, SGD, RMSProp, ADAMAX) and evaluated individual performance.
• Applied data augmentation and fuzzy image enhancement to improve feature extraction and mitigate class imbalance.
• Utilized Bayesian optimization to determine optimal weights for a deep ensemble model, exploring 11 combinations.

Main Results:

• The final ensemble model (EDL7: DL1 + DL2 + DL3) achieved a top accuracy of 97.94%.
• The ensemble model demonstrated superior performance over individual CNN models.
• Achieved high precision (0.981), recall (0.983), and F1 score (0.982).

Conclusions:

• Bayesian-optimized ensemble learning significantly improves potato blight detection accuracy.
• The proposed method offers a robust and reliable solution for agricultural disease classification.
• This approach has the potential to minimize crop losses due to potato blight.