Banana Leaf Disease Image Dataset 2026
Description
Title: Banana Leaf Disease Image Dataset 2026 About data collection: The image data was collected from two major banana-growing regions in the Chittagong division of Bangladesh: Rangunia Upazila, Chittagong Hathazari Upazila, Chittagong. Data Acquisition: Device Used: Realme Speed Edition 9 smartphone camera Image Format: All collected images were stored and subsequently resized to 224 × 224 pixels resolution, suitable for input into deep learning architectures. Capture Conditions: Multiple angles and distances were used to capture both early-stage and advanced-stage fungal infections, along with healthy plant samples for class balance. Dataset Classes: The dataset comprises images representing: Healthy banana plants (no visible fungal infection) Fungal-infected banana plants (including common diseases such as Sigatoka leaf spot, Panama wilt, dead leaf, insect spot) The collected raw images were organized and labeled according to disease category before undergoing preprocessing. Data Preprocessing: Resizing: Standardized to 224 × 224 pixels Purpose of Data Collection: This dataset supports the development of an explainable deep learning framework (SHAP-enabled) for accurate, transparent, and trustworthy disease detection in smart agricultural systems.
Files
Institutions
- Southern UniversityChittagong, Chittagong