Java Plum Leaf Dataset from Bangladesh
Description
The Java Plum Leaf Dataset is a comprehensive collection of high-quality images of java plum leaf, systematically gathered from Amil Model Town in Ashulia, Savar, Dhaka, Bangladesh, between October 01, 2025, and October 20, 2025. The dataset consists of ten distinct classes: Abnormal, Anthracnose, Cutting, Dry, Healthy, Leaf_Blight, Leaf_Spot, Sooty_Mold, Ant_Affect and Powdery_Mildew which represent various stages of both diseased and healthy conditions in java plum leaf. The total dataset contains 2,940 images, with class-wise distribution as follows: Abnormal (320 images), Ant_Affect (335 images), Anthracnose (338 images), Cutting ( 356 images) ,Healthy Leaf (297 images), Dry Leaf (328 images), Leaf_Blight(336 images), Powdery_Mildew (316 images), Sooty_Mold (211 images) and Leaf Spot (106 images). The images were captured using the iphone 15 pro max smartphone in natural lighting conditions, ensuring high resolution and clarity. The original images were captured in jpg format at a resolution of 6048 × 8064 pixels, later resized to 480 × 640 pixels with a resolution of 72 dpi to standardize the dataset. Although the dataset does not include annotations, it serves as a valuable resource for researchers and machine learning practitioners focused on plant disease classification and early disease detection in java plum trees. This dataset is particularly relevant for applications in precision agriculture and automated plant pathology analysis, where understanding leaf conditions can play a critical role in improving crop health management strategies.
Files
Institutions
- Daffodil International University