RiceLeaf-06-2777: A Rice Leaf Disease and Pest Damage Image Dataset

Description

This dataset comprises 2,777 rice leaf images distributed across six categories representing different rice leaf conditions: bacterial leaf blight (blb), blight, brown spot, healthy leaves, pest damage, and tungro. The collection is intended to support research in machine learning, deep learning, computer vision, and automated rice leaf disease classification. It provides image data for developing and evaluating AI-based approaches to plant disease recognition and agricultural monitoring. Dataset Composition The dataset contains the following six categories: BLB (831 images): Images assigned to the blb category. Blight (505 images): Images classified under the blight category. Brown Spot (517 images): Images belonging to the brown spot category. Healthy (267 images): Images labeled as healthy. Pest Damage (379 images): Images categorized as pest damage. Tungro (278 images): Images labeled as tungro. The class distribution provides examples of all six categories, with varying numbers of images per class. Image Characteristics The collection contains JPEG images with varying spatial dimensions and aspect ratios. Image widths range from 435 to 4,080 pixels, while image heights range from 567 to 4,080 pixels. The total storage size of the image collection is approximately 6.001 GB. Individual image file sizes range from approximately 20 KB to 5.89 MB. The most frequently occurring image dimensions are 3072 × 4080 pixels (1,399 images), 1840 × 4080 pixels (486 images), and 2296 × 4080 pixels (299 images). Potential Applications This dataset can be used for: Rice leaf disease and pest damage classification. Convolutional neural network (CNN) development and evaluation. Transfer learning using pretrained deep learning architectures. Comparative evaluation of image classification models. Agricultural computer vision and AI-based crop health monitoring. Investigations into class imbalance and model generalization. Technical Summary Total images: 2,777 Number of classes: 6 Image format: JPEG (.jpg) Image width range: 435–4,080 pixels Image height range: 567–4,080 pixels Individual file size range: Approximately 20 KB–5.89 MB Total image storage: Approximately 6.001 GiB Capture-date range recorded in metadata: February 18–March 31, 2026 Location: Monirampur, Jashore. Organization: Class-specific directories This dataset provides a resource for researchers developing data-driven approaches to rice leaf condition recognition and intelligent agricultural applications.

Institutions

• Daffodil International University

  Dhaka Division, Dhaka

Categories

Licence