Potato Leaf Disease Dataset
Description
This dataset contains annotated potato leaf images collected under real agricultural field conditions for automated plant disease recognition and precision agriculture research. The images were captured at the Bangladesh Agricultural Research Institute (BARI), Agricultural Research Station, Khulshi, Chattogram, Bangladesh, using an Apple iPhone 15 smartphone camera under natural lighting and background conditions. The dataset includes both healthy and diseased potato leaves affected by viral, bacterial, and fungal pathogens. A total of 804 original images were collected and categorized into six classes: Healthy, Bacterial Soft Rot, Fungal Late Blight, Viral Leaf Roll, Viral PVX, and Viral PVY. To improve class balance and enhance model generalization, several data augmentation techniques such as rotation, flipping, zooming, shearing, shifting, and brightness adjustment were applied, resulting in a final dataset of 2400 images with 400 images per class.