セマンティックセグメンテーションによる衛星画像を用いたUAVの自己位置推定の精度向上

DOI Web Site

書誌事項

タイトル別名
  • Improving The Accuracy of Visual Localization for UAV Using Satellite Imagery by Semantic Segmentation

抄録

<p>In this paper, we propose a method to accurately estimate the position of a UAV on satellite images by using the region information of buildings extracted by semantic segmentation for UAV pose estimation. The localization of UAV using satellite images is based on the similarity between images features using intensity information. However, the estimation accuracy of intensity information alone is degraded by changes in the seasons and moving objects. Therefore, in addition to the intensity information, the proposed meathod uses information on building areas extracted by semantic segmentation from overhead images to improve the accuracy of position estimation.</p>

収録刊行物

キーワード

詳細情報 詳細情報について

問題の指摘

ページトップへ