Deep Transfer Learning Technique for Multimodal Disease Classification in Plant Images

Area of Science:

• Agricultural Science
• Computer Vision
• Machine Learning

Background:

• Rice (Oryza sativa) is a major Indian crop, crucial for agriculture and GDP.
• Accurate disease detection in rice is vital for crop yield and food security.
• Image recognition technologies offer promising solutions for agricultural challenges.

Purpose of the Study:

• To provide a comprehensive survey of methodologies for rice plant disease detection.
• To analyze classifiers and strategies used in identifying rice crop illnesses.
• To propose an enhanced Convolutional Neural Network (CNN) model for disease detection.

Main Methods:

• Systematic review of research papers from the last decade on rice plant diseases.
• Analysis of various classification techniques and their effectiveness.
• Development and proposal of an enhanced CNN model for image-based disease identification.

Main Results:

• Identified and categorized diverse approaches for rice disease detection.
• Evaluated the performance of different classifiers.
• Demonstrated the potential of deep neural networks, specifically CNNs, in image classification for plant disease recognition.

Conclusions:

• Deep neural networks show significant success in image classification tasks for plant disease recognition.
• The proposed enhanced CNN model offers a promising approach for accurate rice disease detection.
• Further research can build upon these findings to enhance agricultural monitoring systems.