Recovery of satellite data and knowledge discovery by discrete optimization and time series analysis

通过离散优化和时间序列分析恢复卫星数据和知识发现

• 批准号: 22650029
• 负责人: TSUCHIYA Takashi
• 金额: $ 1.7万
• 依托单位: National Graduate Institute for Policy Studies
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Challenging Exploratory Research
• 财政年份: 2010
• 资助国家: 日本
• 起止时间: 2010-04-01 至 2013-03-31
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-22650029/
• 关键词: 人工衛星; データ解析; 最適化; 統計; データ復元; 事前情報; 惑星科学; 線形計画; Geitail; プラズマ粒子; 時系列解析; 整数計画

Geotail is a satellite to observe plasma of the magnetsphere of the earth. The distribution of observed low energy plasma particles are summarized as a three dimensional histogram before sent to the ground. Researchers conduct analysis based on a color chart called ET-spectrogram drawn from the three-dimensional histogram. However, a complete three dimensional histogram data are sent only eight hours per day, and for remaining period a marginalized two-dimensional histogram and ten statistics of the original data are sent to the ground. In this research, we made an attempt to recover the original three dimensional histogram from the compressed two dimensional histogram and 10 statistics to draw an ET-spectrogram. We developed a LP model which incorporates with appropriate prior information to recover the original three dimensional histogram and ET-spectrogram successfully.

Geotail是一颗观测地球磁层等离子体的卫星。观测到的低能等离子体粒子在发送到地面之前的分布被总结为三维直方图。研究人员根据三维直方图绘制的et谱图进行分析。然而，一个完整的三维直方图数据每天只发送8小时，其余时间发送一个边缘的二维直方图和10个原始数据的统计量。在本研究中，我们尝试从压缩后的二维直方图和10个统计量中恢复原始的三维直方图来绘制et谱图。我们建立了一个融合了适当先验信息的LP模型，成功地恢复了原始的三维直方图和et谱图。

项目成果

衛星観測プラズマデータの復元

卫星观测等离子体数据恢复

• 发表时间: 2010
• 作者: 田辺亮平;木下順三郎;河野佑亮;池上敦子;上野玄太;土谷隆
• 通讯作者: 土谷隆

最適化による人工衛星データの情報復元と知識発見

卫星数据优化信息修复与知识发现

• 作者: 池上敦子;村田裕介;土谷隆;上野玄太
• 通讯作者: 上野玄太

事前知識と線形計画法を用いた人工衛星観測データの復元

利用先验知识和线性规划重建卫星观测数据

• 发表时间: 2014
• 作者: 馴松晃司;土谷隆;上野玄太;池上敦子
• 通讯作者: 池上敦子