53 orthomosaics processed from Unoccupied Aerial Vehicle (UAV) image data of lake shorelines sampled during a field campaign in Central and Eastern Yakutia, Siberia in 2021 (RU-Land_2021_Yakutia)

During the RU-Land_2021_Yakutia summer field campaign in August and September 2021 in the Verkhoyansk Mountain Range in Eastern Yakutia and in the Central Yakutian Lowland, multispectral drone-based images were acquired over 53 selected lakes to analyse the vegetation and shallow lake waters along shores and to record the current lake shorelines. The images were taken in the course of further investigations of the lakes during that summer expedition. Baisheva et al. (2022) gives an overview of the lakes studied and the corresponding hydrochemistry. In addition, we published datasets including water isotope data of the lake (Stieg et al. 2022) and vegetation surveys of the lakeshores (Stieg et al. 2022).This dataset includes the orthomosaics (in raw data format (DN), and for the good-quality acquisitions normalised to surface reflectance) and the processing reports of the 53 sampled lakes. The event list gives an overview of the relevant lake information, which can be found here: https://doi.pangaea.de/10.1594/PANGAEA.955723 We used a consumer-grade, lightweight Unoccupied Aerial Vehicle (UAV) set up in combination with a D-RTK Station (GNSS antenna). The multispectral images were taken by a DJI Phantom 4 quadcopter UAV including an imaging system with one Red-Green-Blue RGB sensor and five spectral channels, able to capture both colour and narrow band images (5 bands: Blue, Green, Red, Red-edge, Near-infrared).A standardized flight plan was used to capture the shoreline of the lakes whenever possible using the DJI GS Pro app. If a preliminary route planning was not possible due to non-high-resolution map material or external circumstances (limited view, wind, rain), it was flown manually. For both flight procedures, multispectral images were taken automatically every 2 seconds. Speed was set to 4 m/s and altitude above ground level to circa 55 - 60 m.The produced UAV images were processed to construct the orthoimages using the software Agisoft Metashape Professional, version: 1.7.5. build 13229 (64 bit). Orthoimages are geometrically corrected images that are georeferenced to the topography (the relief) and vegetation (the top-of-canopy elevation). The orthomosaics were constructed from the multiple overlapping pictures from different camera viewpoints which make it possible to create a photogrammetric point cloud reconstruction and constructing the orthomosaics using structure from motion/multi-view stereo (SfM-MVS) techniques. There was no preselection of the images before the processing. Regular settings of the software were used, the processing parameters are listed in the individual processing report of each orthomosaic. The Micasense DLS2 illumination-sensor data, measured in parallel during each acquisition, was used to normalize the Digital Number (DN) of the orthomosaics to surface reflectance.

2021年8月至9月，于雅库提亚东部维尔霍扬斯克山脉及雅库提亚中部低地开展RU-Land_2021_Yakutia夏季野外考察活动期间，研究团队针对选定的53个湖泊采集了多光谱无人机影像，用于分析湖岸带植被与浅湖水体，并记录当前湖岸线。该影像为本次夏季科考期间对湖泊开展进一步调查的成果之一。Baisheva等人（2022）概述了本次研究涉及的湖泊及其对应的水化学特征。此外，团队还发布了包含湖泊水同位素数据（Stieg等，2022）及湖岸植被调查数据（Stieg等，2022）的配套数据集。
本数据集包含53个采样湖泊的正射镶嵌图（原始数据格式为数字量（Digital Number，DN），优质采集数据已归一化为地表反射率）及处理报告。相关湖泊信息的概述可通过事件列表查阅，详情见：https://doi.pangaea.de/10.1594/PANGAEA.955723。
本次采集采用消费级轻量化无人飞行器（Unoccupied Aerial Vehicle，UAV），搭配D-RTK基站（GNSS天线）。多光谱影像由DJI Phantom 4四旋翼无人机搭载的成像系统采集，该系统包含1个红绿蓝（RGB）传感器与5个光谱通道，可同时获取彩色影像与窄波段影像（5个波段：蓝、绿、红、红边、近红外）。
研究采用标准化飞行计划，尽可能通过DJI GS Pro应用程序完成湖岸线影像采集。若因缺乏高分辨率地图素材或外部因素（视野受限、大风、降雨）无法开展预规划航线，则采用手动飞行模式。两种飞行模式下，均自动每2秒拍摄一张多光谱影像，飞行速度设定为4 m/s，离地高度约为55-60 m。
采集的无人机影像通过Agisoft Metashape Professional软件（版本：1.7.5 build 13229，64位）处理以生成正射影像。正射影像是经过几何校正的地理配准影像，其配准依据为地形起伏与植被冠层高程。正射镶嵌图由不同拍摄视角的多张重叠影像构建而成，通过运动恢复结构/多视图立体（SfM-MVS）技术完成摄影测量点云重建并生成正射镶嵌图。处理前未对影像进行预筛选，采用软件默认参数，具体处理参数详见每张正射影像的独立处理报告。每次采集时同步获取的Micasense DLS2光照传感器数据，被用于将正射影像的数字量（DN）归一化为地表反射率。