Jujube Leaf Disease and Health Image Dataset: Anthracnose, Powdery Mildew, Insect Damage, Yellowing, and Healthy Classes

This dataset contains RGB images of jujube (Ziziphus jujuba) leaves collected for research on automatic plant disease recognition, stress detection, and leaf health assessment. The images are grouped into five phenotypic classes:
1.Anthracnose
2.Healthy
3.Powdery Mildew
4.Insect Damage
5.Yellow Leaf
In total, the dataset includes 1,260 images, with 300 images per class for Anthracnose, Healthy, Insect Damage, and Yellow Leaf, and 60 images for Powdery Mildew. The lower number of Powdery Mildew samples reflects its lower prevalence during the data collection period and is intentionally preserved to represent realistic field conditions.
Images were acquired under natural field conditions with varying backgrounds, illumination, and leaf orientations to improve the robustness and real-world relevance of computer vision and deep learning models trained on this dataset. All images are provided in their original RGB format without synthetic augmentation; users may apply their own augmentation strategies (e.g., rotation, flipping, color jitter) and class-weighting methods to address class imbalance in downstream analyses.A key strength of this dataset is the expert verification of disease and syndrome information:

The presence or absence of disease for each symptomatic class (Anthracnose, Powdery Mildew, Insect Damage, Yellow Leaf) and the Healthy class was verified and documented through a formal disease verification certificate issued by qualified experts.In addition, a separate syndrome and medicine verification certificate is provided, describing the characteristic visual syndromes (leaf symptoms) of each disease class and confirming the recommended treatment/management options (e.g., appropriate pesticides, fungicides, or other control measures) as validated by subject-matter specialists.

Scanned copies of both the disease verification certificate and the syndrome and medicine verification certificate are included as supplementary files in this repository. These documents provide users with transparent evidence of diagnostic validity and treatment recommendations, enhancing the reliability of the dataset for scientific, agronomic, and decision-support applications.

This dataset is intended for, but not limited to:
Training and benchmarking image classification and deep learning models for jujube leaf disease detection.Research on plant stress phenotyping and early disease diagnosis.
Development of mobile or edge-based decision support systems for farmers and extension workers.
Educational purposes in plant pathology, precision agriculture, and agricultural informatics.
Users are encouraged to cite this dataset when using it in publications and to note the presence of the attached verification certificates as part of the dataset’s quality assurance.