Rose Leaf Disease Detection Dataset

Roses are cherished worldwide as symbols of love, purity, and affection, and they also serve as an important commercial crop. However, various leaf diseases significantly impact their aesthetic value and economic profitability. To address this challenge, we developed a specialized dataset of rose leaf images, aimed at advancing research in machine learning for disease identification.

This dataset was collected through field visits to Golaap Gram, Allah Vorsha Nursery, and Daffodil Nursery. All images were captured under natural daylight to ensure clear and detailed visuals. The dataset is organized into four distinct categories, with both original and augmented versions:
Without Augmentation:
● Black Spot: 432 images
● Downy Mildew: 390 images
● Insects Infected: 431 images
● Healthy: 795 images
Total: 2,048 images
Augmented Classes:
● Black Spot: 4,320 images
● Downy Mildew: 3,900 images
● Insects Infected: 4,310 images
● Healthy: 7,950 images
Total Augmented: 20,480 images
To maximize its usefulness for research purposes, the dataset has undergone comprehensive preprocessing, including image classification, noise reduction, background removal, and augmentation. The augmented dataset significantly enhances the diversity and size of the original dataset, providing a robust foundation for training and testing machine learning models.