Agri-Vision4: A Comprehensive Multi-Crop Leaf Disease Dataset of Tomato, Papaya, Zucchini, and Bottle Gourd from Bangladesh
Description
Agri-Vision4 is an expert-validated, multi-crop leaf disease dataset designed to advance computer vision and machine learning research in precision agriculture. Collected from diverse agricultural regions in Bangladesh using a full-frame SONY ALPHA 7 II camera, this dataset integrates four major crops—Tomato (Solanum lycopersicum), Papaya (Carica papaya), Zucchini (Cucurbita pepo), and Bottle Gourd (Lagenaria siceraria)—comprising a total of 28,000 images (5,266 original and 22,734 augmented) across 28 distinct disease and healthy classes. Strictly adhering to the Data in Brief (June 2025) policy, all images were standardized to 512x512 pixels, verified by senior agronomists to ensure diagnostic accuracy, and expanded using Python-based augmentation techniques (rotation, flipping, noise injection) to address class imbalance. Organized into separate directories for raw and augmented data, this high-resolution resource captures real-world variability in lighting and texture, serving as a robust benchmark for developing automated plant disease detection models.
Files
Institutions
- Daffodil International University