SLIF-Brinjal: An In-Field Leaf Dataset for Disease Recognition in Precision Agriculture

Description

This dataset contains 8987 images of brinjal (eggplant) leaves collected under real-time agricultural field conditions in Sri Lanka. Images were captured with a smartphone camera under natural light, without controlled backgrounds, to reflect realistic field environments across diverse agroclimatic zones. The dataset includes eight classes: Bacterial Blight, Bacterial Leaf Spot, Bacterial Wilt, Cercospora Leaf Spot, Little Leaf, Mosaic Virus, Powdery Mildew and Healthy leaves. All images are stored in JPG format at a consistent resolution and were manually labelled and cross-validated by agricultural pathology experts based on visible disease symptoms. The dataset is created to support research and development in machine learning, deep learning, and computer vision, particularly for brinjal leaf disease detection and precision agriculture applications under real-world conditions. Users of this dataset are required to cite the following publications. George, R., Nishankar, S., Thuseethan, S., Pakeerathan, K., Ragel, R.G., Pavindran, V., 2026. SLIF-Brinjal: An In-Field Leaf Dataset for Disease Recognition in Precision Agriculture. Scientific Data.

Steps to reproduce

1. Download and extract the dataset file, which contains three subfolders: Raw Dataset (1459), Phase I – Dataset (4377), and Phase II – Dataset (8987). 2. Organize images according to the provided class folders for each dataset. 3. Use standard image preprocessing techniques such as resizing and normalization. 4. Split the dataset into training, validation, and test sets as required. 5. Train machine learning or deep learning models using the prepared data for brinjal leaf disease classification.

Institutions

• University of Peradeniya

  Central Province, Kandy
• Charles Darwin University

  Northern Territory, Darwin

Categories

Licence