UAV Multispectral Image Dataset for Urban Mistletoe (<i>Struthanthus interruptus</i>) Detection — Mexico City

This repository provides the complete multispectral UAV-based dataset used in the study “Multitemporal UAV Multispectral Dataset with Semantic Segmentation and Data Augmentation for Urban Mistletoe (Struthanthus interruptus) Monitoring." The images were captured in an urban park in Mexico City, targeting S. interruptus, a hemiparasitic plant that affects different species of trees. The collection contains preprocessed and derived products generated after geometric correction, radiometric correction, RGB recomposition, semantic segmentation, and several augmentation pipelines, including texture (GLCM), vegetation indices, and chromatic transformations. All data originate from two acquisition UAV flights: T1 — September 2022 (GSD = 3.6 cm/pixel) T2 — May 2023 (GSD = 2.5 cm/pixel)The dataset contains 1,324 multispectral tiles (250 × 250 pixels) spatially aligned across all modalities: multispectral, RGB composite, augmentations (GLCM texture descriptors, vegetation indices, and color spaces), and expert-validated binary segmentation masks with the presence or absence of mistletoe. The dataset includes a total of 289,956 images across all data types. This dataset is described in the associated Data Descriptor currently under review at Scientific Data.