Banana Leaf Image Dataset for Classification of Sigatoka, Cordana, and Healthy Leaves
Description
This dataset contains 2,354 banana leaf images collected under natural field conditions across four agricultural sites in the San Martín Region, Peru: Calzada and Habana in Moyobamba Province, Zapatero in Lamas Province, and Shima in Huallaga Province. The images are organized into three classes: Sigatoka, Cordana, and healthy leaves. The dataset is distributed across three folders: • sigatoka: 995 images labeled as leaves affected by Sigatoka. • cordana: 815 images labeled as leaves affected by Cordana. • sanas: 544 images labeled as healthy leaves. Each folder name indicates the class assigned to its images. The dataset is intended to support the development and evaluation of machine learning and deep learning models for banana leaf image classification, including the distinction between diseased and healthy leaves.
Files
Institutions
- Universidad Nacional de San MartínSan Martín Department, Tarapoto