A Machine Learning Dataset for Classification of Common Coffee Leaf Diseases in Uganda.
Description
This dataset provides a well-structured collection of 3,312 labeled images of coffee leaves captured from farms in Uganda. The images are categorized into three main classes: Healthy, Coffee Leaf Rust , and Phoma disease. Each class is stored in separate folders to facilitate easy retrieval and processing. All images are in JPEG format with a resolution of 256 × 256 pixels. The Healthy folder contains 1,179 images of disease-free coffee leaves, the CLR folder holds 1,023 images of leaves affected by Coffee Leaf Rust, and the Phoma folder contains 1,110 images showing Phoma disease symptoms. Image augmentation techniques, including rotation, flipping, and brightness adjustment, were applied to address class imbalance and increase dataset diversity for machine learning tasks. This dataset is valuable for research in computer vision applications like image classification and disease detection in coffee plants.
Files
Steps to reproduce
The coffee leaf dataset was collected from farms in Uganda using a smartphone camera under varying lighting conditions (daylight and low light). Leaves were sampled at early, medium, and advanced growth stages. Images were categorized into healthy, Coffee Leaf Rust , and Phoma disease classes. Data cleaning involved removing duplicates and enhancing image diversity through augmentation techniques like rotation, flipping, and brightness adjustments. All images were labeled and stored in distinct folders for structured processing.