Bangladesh Bean Leaf and Pod Disease Image Dataset for Multi-Class Classification

Description

This dataset presents a comprehensive collection of bean leaf and bean pod images gathered from real agricultural fields located in Bogura (B Block), Mymensingh (Ishwarganj), and Netrokona, Bangladesh. The dataset includes both healthy samples and disease-affected specimens representing several important diseases that commonly impact bean production. All images were captured under natural field conditions using smartphone cameras, preserving variations in lighting, background complexity, viewing angles, disease severity, and plant growth stages. Such diversity enhances the dataset's ability to reflect real-world agricultural environments and improves its suitability for developing robust artificial intelligence solutions. The dataset is designed to support research in plant disease recognition, computer vision, machine learning, deep learning, precision agriculture, and intelligent crop monitoring systems. It can be utilized for tasks such as disease classification, object detection, image segmentation, transfer learning, and automated decision-support systems for sustainable agriculture. By providing field-collected and carefully organized images of both bean leaves and bean pods, this dataset serves as a valuable resource for researchers, students, and agricultural technology developers working on automated plant disease diagnosis and smart farming applications.

Steps to reproduce

1. Download the dataset from Mendeley Data. 2. Extract all ZIP files containing the bean leaf and bean pod disease image classes. 3. Organize the images according to their class labels: Bean Leaf Anthracnose Bean Leaf Anthracnose (Advanced Infection) Bean Leaf Cercospora Leaf Spot Bean Leaf White Mold Bean Leaf Mosaic Disease Healthy Bean Leaf 4. Preprocess the images by resizing them to the desired input size (e.g., 224 × 224 pixels) and applying normalization techniques. 5. Split the dataset into training, validation, and testing sets according to the experimental design. 6. Apply data augmentation techniques, if required, to improve model generalization. 7. Train machine learning or deep learning models for bean disease classification. 8. Evaluate model performance using standard metrics such as accuracy, precision, recall, F1-score, confusion matrix, and ROC-AUC. 9. Compare the performance of different models and experimental settings to identify the most effective approach for bean disease recognition. 10. Use the trained models for automated bean disease detection and precision agriculture applications.

Institutions

• Daffodil International University

  Dhaka Division, Dhaka

Categories

Licence