Paddy Field Pest Image Dataset for Deep Learning-Based Multi-Class Classification

This dataset consists of over 22108 high-resolution images of paddy field pests, collected from various locations in Bangladesh, specifically designed for deep learning applications in pest recognition and classification. The dataset includes six major pest species commonly found in South Asian rice fields: Brown Planthopper, Green Leafhopper, Leaf Folder, Rice Bug, Stem Borer, and Whorl Maggot. Each image has been carefully annotated by entomologists and categorized based on pest severity levels, including highly harmful, moderately harmful, moderately less harmful, and less harmful. The images were captured under real-world conditions using smartphone cameras (Redmi Note 10 Pro Max and iPhone XS) in varying environmental conditions.

This dataset is essential for the development and evaluation of automated pest detection systems using deep learning techniques such as Convolutional Neural Networks (CNNs). It supports the early and accurate identification of pests, contributing to sustainable agricultural practices and precision pest management in rice cultivation. The dataset also includes augmented images through methods like rotation, scaling, and brightness adjustment, enhancing the variability and robustness of deep learning models. This resource will be valuable for researchers, farmers, and practitioners working on pest monitoring systems and sustainable crop protection.