UAV-Based Aquatic Plant Dataset for Segmentation and Classification

This dataset contains UAV-captured images of aquatic plant species from wetland ecosystems in Bangladesh, including Bangladesh Agricultural University, Zinda Park, and Water Lily Lake. It supports both semantic segmentation and classification tasks. The segmentation subset includes 1,761 image–mask pairs with pixel-level binary annotations, while the classification subset contains 1,468 cropped images across 14 species. Images were acquired using a DJI Mavic 4 Pro drone under controlled conditions to ensure quality and consistency. The dataset is divided into training and testing sets and is suitable for computer vision research in environmental monitoring, biodiversity analysis, and precision agriculture applications.