MAVSD：A Multi-Angle View Segmentation Dataset for Detection of Solidago Canadensis L

MAVSD (Multi-Angle View Segmentation Dataset) is a comprehensive multi-angle aerial dataset specifically designed for detecting and segmenting Solidago canadensis L., a widespread invasive plant species. The dataset features high-resolution RGB imagery captured by DJI M300 UAV from four distinct viewing angles (30°, 45°, 60°, and 90°), providing comprehensive coverage of the target plant's morphological characteristics in natural environments. The dataset encompasses 13 semantic categories, including the invasive species, various vegetation types, and common terrain features. All images are provided with dual annotation formats: VOC format classification labels and polygon-based semantic segmentation annotations (JSON+PNG format). The dataset is divided into training and testing sets, with corresponding file lists for each viewing angle. MAVSD's multi-angle acquisition strategy and detailed semantic annotations make it particularly suitable for developing deep learning models for invasive species monitoring and ecological surveys in real-world applications.

This dataset aims to address the challenges in automated vegetation monitoring by providing diverse perspectives and high-quality annotations, facilitating the development of more robust and accurate detection systems for invasive species management.

MAVSD（多视角分割数据集，Multi-Angle View Segmentation Dataset）是一款专为加拿大一枝黄花（Solidago canadensis L.，一种广泛分布的入侵植物物种）的检测与分割任务打造的综合性多视角航空数据集。该数据集采用大疆M300无人机（DJI M300 UAV）采集的高分辨率RGB影像，涵盖4种不同拍摄视角（30°、45°、60°与90°），可全面覆盖自然环境中目标入侵植物的形态学特征。数据集共包含13个语义类别，涵盖该入侵物种、多种植被类型以及常见地形特征。所有影像均提供双标注格式：VOC格式分类标签，以及基于多边形的语义分割标注（JSON+PNG格式）。数据集已划分为训练集与测试集，并为每种拍摄视角配备了对应的文件列表。MAVSD的多视角采集方案与精细化语义标注，使其尤其适用于开发面向实际应用的入侵物种监测与生态调查深度学习模型。

本数据集旨在通过提供多样化视角与高质量标注，解决植被自动化监测中的现存挑战，助力开发更鲁棒、更精准的入侵物种管控检测系统。