一种新的星载光学遥感器光轴指向运动滤波方法  

A New Filter Method for Optic Axis Pointing Motion of Spaceborne Optical Remote Sensors

作　　者：侯帅 于飞[1] 李超[1] 刘成 黄刚[1] 侯丹[1] 赵丽婷[1] HOU Shuai;YU Fei;LI Chao;LIU Cheng;HUANG Gang;HOU Dan;ZHAO Liting(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1期47-55,共9页Spacecraft Recovery & Remote Sensing

基　　金：国家重大科技专项工程。

摘　　要：针对应用在动中成像的星载光学遥感器光轴指向运动进行准确的运动估计问题,需要在测量出相机光轴的运动信息后对噪声进行滤波,以获得准确的相机光轴指向偏移量。文章对低通滤波、最小均方自适应滤波、卡尔曼滤波和无迹卡尔曼滤波等四种常见方法进行了试验分析,明确了低通滤波方法有更好的效果和适应性,但是性能还需要进一步提升。对相机光轴运动的随机性与所受噪声的统计特征不可预知性,提出了一种根据指向偏移与噪声幅频特性差异自确定截止频率的二次双向低通滤波方法,实现了截止频率的在线估计,解决了常见低通滤波的相位失真问题,滤波效果较四种常见方法更好,满足实际工程应用需求。Aiming at the problem of accurate estimation of the optical axis pointing motion of spaceborne optical remote sensors imaging in motion state,it’s necessary to filter the noise from the measurement information to obtain accurate pointing offset of the optical axis.Four common methods,i.e.low-pass filter,least mean square(LMS)filter,Kalman filter and unscented Kalman filter,are analyzed,and the results show that the low-pass filter method is better in efficiency and adaption,and has potential in improving performance.Considering the randomness of camera optical axis motion and the unpredictability of statistical characteristics of the received noise,a two-way low-pass filter method with self-determined cut-off frequency is proposed according to the difference of amplitude frequency characteristics between the camera optical axis pointing offset and the noise.The experimental results show that this method realizes on-line estimation of cut-off frequency and solves the problem of phase distortion in low-pass filter method,with filtering effect better than the traditional four methods,meeting the needs of practical engineering application.

关 键 词：动中成像 光轴指向运动 运动估计 运动滤波 光学遥感器 遥感卫星 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]