空间光学遥感器多维关联分析指标推荐方法  

A Parameter Recommendation Method of Space Optical Remote Sensors Based on Association Analysis

作　　者：尚志鸣[1] 文高进[1] 李辰[1] 王洪民[1] SHANG Zhiming;WEN Gaojin;LI Chen;WANG Hongmin(Beijing Institute of Space Mechanics&Electricity,Beijing 100094,China)

机构地区：[1]北京空间机电研究所,北京100094

出　　处：《航天返回与遥感》2020年第4期111-117,共7页Spacecraft Recovery & Remote Sensing

摘　　要：目前空间光学遥感载荷的相关设计工作仍依赖专家知识,对积累的历史数据挖掘不足。文章提出基于关联分析的空间光学遥感器指标自动推荐方法。以谱段设计为例,首先依据多维数据特征,整理遥感载荷应用方向与监测谱段数据,构建事务数据库;基于Apriori算法,设计有效规则筛选指标,提取遥感应用方向与监测谱段的关联规则,实现由具体应用自主推荐谱段指标的功能;最后,定量分析了大气、海洋、陆地等各类遥感应用方向与光谱中心波长间有意义的关联,结果显示提出方法可为面向应用的遥感载荷指标设计提供有益的支持。At present,the design task of space optical remote sensing cameras relies on expert knowledge,and historical technical parameters data are insufficient in mining.The paper proposes an automatic parameter recommendation method for space optical remote sensor based on association analysis and takes the spectral design as an example.According to the characteristics of multi-dimensional data,the specific remote sensing tasks and its corresponding remote sensing sensor parameters are collected and normalized to construct the transaction database.Based on the Apriori algorithm,an effective rule selection method is designed.Then the credible association rules of the remote sensing tasks and the center wavelength of the spectrum segment are extracted.At last the function recommending the spectral parameters from the specific remote sensing task is obtained,and its reliability is given quantitatively.The results show that the proposed method can provide useful support for the design of application-oriented remote sensing cameras.

关 键 词：遥感应用 指标推荐 关联分析 关联规则评价 

分 类 号：V443.5[航空宇航科学与技术—飞行器设计]