Sri Lankan Tea Leaf Dataset

This dataset comprises tea leaf images collected from various Sri Lankan tea estates, specifically curated for automated disease recognition. It consists of 1,791 raw images alongside an offline-augmented version totaling 5,968 images, distributed across six distinct categories: algal leaf spot, black blight, blister blight, gray blight, spider mites, and healthy leaves. The raw images capture natural visual variability under field acquisition conditions, incorporating diverse backgrounds, fluctuating illumination, varying leaf poses, symptom severity levels, and distinct growth stages. To mitigate class imbalance and enhance visual diversity for robust machine-learning model training, class-wise offline augmentation techniques were systematically applied to generate the expanded dataset.
Category Distribution:

Algal leaf spot:
Raw: 417 images
Augmented: 1240 images

Black blight:
Raw: 345 images
Augmented: 966 images

Blister blight:
Raw: 126 images
Augmented: 757 images

Gray blight:
Raw: 431 images
Augmented: 1245 images

Spider mites:
Raw: 138 images
Augmented: 788 images

Healthy:
Raw: 334images
Augmented: 972 images

Key Features
The dataset provides a region-specific image resource for tea leaf disease recognition. It includes multiple disease categories with visually diverse symptoms, such as lesion color variation, texture differences, margin irregularity, and disease spread patterns. The augmented version expands the original dataset using lesion-preserving transformations, including rotation, flipping, scaling, brightness adjustment, and related image transformations, while maintaining disease-relevant visual characteristics.

Purpose
This dataset is intended to support the development and evaluation of machine-learning and deep-learning models for tea leaf disease classification. It can be used for automated disease recognition, lightweight model benchmarking, data augmentation studies, transfer learning, and mobile-assisted crop health monitoring. The dataset may help researchers and agricultural practitioners develop early screening tools for tea leaf diseases and support more timely disease management in tea cultivation.