Pisum sativum Image Dataset: Healthy and Disease-Affected Cases

Early and accurate detection of plant diseases is vital for sustainable agriculture and optimizing crop yields. This dataset, titled the Pisum sativum Image Dataset, presents collection of images specifically curated to support machine learning research in plant disease detection and classification.

Focusing on Pisum sativum (pea)—a widely cultivated legume and critical food source—the dataset captures a broad range of visual symptoms associated with fungal, viral, and pest-induced diseases. It contains 12,096 labeled images, distributed across 10 distinct classes:

Anthracnose (645 images)

Ascochyta blight (3,839 images)

Botrytis blight (1,457 images)

Downy mildew (315 images)

Fusarium wilt (640 images)

Pea leaf roll virus (40 images)

Pod borer damage (891 images)

Powdery mildew (529 images)

Stem rot (41 images)

Healthy samples (3,699 images)

Captured under diverse conditions—varying in lighting, background, and severity—the dataset provides a challenging and representative benchmark for developing and evaluating computer vision algorithms.

This resource enables both binary classification (healthy vs. diseased) and multiclass classification tasks. It is well-suited for experimentation with convolutional neural networks (CNNs), transfer learning, and data augmentation strategies. Researchers can leverage the dataset to improve the robustness and generalization of AI models for real-time disease detection.

The Pisum sativum Image Dataset offers a valuable foundation for research in precision agriculture, agricultural informatics, and AI-powered crop monitoring. It is intended to support both academic researchers and industry professionals working toward scalable and accurate solutions for plant health management.