<p>CapeHallettUAVortho is a multi-temporal unmanned aerial vehicle (UAV) orthomosaic dataset developed for an Adélie penguin (<i>Pygoscelis adeliae</i>) colony at Cape Hallett, northern Victoria Land, Ross Sea, Antarctica. The dataset comprises six austral-summer UAV campaigns conducted between 2018 and 2025 during the early breeding period in mid-to-late November. A total of 4,471 nadir RGB images were acquired using multiple UAV platforms and survey configurations. A high-resolution 2018 orthomosaic acquired with a DJI Matrice 600 and a Canon EOS 5DS camera served as the primary reference mosaic, and orthomosaics acquired between 2021 and 2025 with consumer-grade DJI UAVs were co-registered using an automated image-to-image registration workflow. The workflow combined deep learning-based and conventional feature-matching approaches and evaluated four geometric transformation models. The released orthomosaics have ground sampling distances ranging from approximately 0.7 to 2.3 cm pix<sup>−1</sup>, with median co-registration residuals of 2.02–2.38 m based on independent tie-point validation. The dataset enables colony-scale analyses of boundary changes, spatial occupancy patterns, and variability in guano distribution and provides a high-resolution reference dataset for satellite validation and AI-based ecological monitoring studies.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Multi-temporal UAV orthomosaics of an Adélie penguin colony at Cape Hallett, Ross Sea, Antarctica (2018–2025)

  • Yongsik Jeong,
  • Jinku Park,
  • Jong-U Kim,
  • Youmin Kim,
  • Younggeun Oh,
  • Sin-Young Kim,
  • Jeong-Hoon Kim

摘要

CapeHallettUAVortho is a multi-temporal unmanned aerial vehicle (UAV) orthomosaic dataset developed for an Adélie penguin (Pygoscelis adeliae) colony at Cape Hallett, northern Victoria Land, Ross Sea, Antarctica. The dataset comprises six austral-summer UAV campaigns conducted between 2018 and 2025 during the early breeding period in mid-to-late November. A total of 4,471 nadir RGB images were acquired using multiple UAV platforms and survey configurations. A high-resolution 2018 orthomosaic acquired with a DJI Matrice 600 and a Canon EOS 5DS camera served as the primary reference mosaic, and orthomosaics acquired between 2021 and 2025 with consumer-grade DJI UAVs were co-registered using an automated image-to-image registration workflow. The workflow combined deep learning-based and conventional feature-matching approaches and evaluated four geometric transformation models. The released orthomosaics have ground sampling distances ranging from approximately 0.7 to 2.3 cm pix−1, with median co-registration residuals of 2.02–2.38 m based on independent tie-point validation. The dataset enables colony-scale analyses of boundary changes, spatial occupancy patterns, and variability in guano distribution and provides a high-resolution reference dataset for satellite validation and AI-based ecological monitoring studies.