Space camera always generates smear fuzzy because multi-factors lead to flutter of platform, which will reduce system transfer function and resolution ratio. as a result imaging system can't reach the ability that the system designs. This project puts forward a fast restoration method of remote sensing image based on gyroscope vibration detection and coded exposure. Double-shaft fiber-optic gyroscope is used to detect the vibration of satellite platform, and imaging camera obtains remote sensing image by means of coded exposure at the same time. Processing system calculates PSF according to vibration data obtained by fiber-optic gyroscope and time code value in camera exposure process. Then the process eliminates influence of platform vibration and acquires clear remote sensing image fast by deconvolution. By advantage of coded exposure, imaging restoration can use method of inverse filtering, and the problem that inverse filtering can't be used due to spectrum value may be zero and noise is amplified infinitely is inexistent. Fiber-optic gyroscope can achive high sample frequency and measure vibration data from low frequency to high frequency accurately. What's more, it can measure arbitrary direction of two-dimensional vibration, which fits the vibration condition of remote sensing satellite platform. This method has absolute advantage in processing speed and image restoration result, and it is significant for getting high resolution remote sensing images fast.
在空间相机成像过程中,多种因素导致成像模糊,其中卫星平台的颤振使图像产生拖影模糊是一个重要因素,使系统无法达到预定的设计指标。 本项目首次提出一种基于陀螺颤振探测和编码曝光的遥感图像快速复原方法。采用双轴光纤陀螺实时探测卫星平台的振动过程,同时成像相机编码曝光获得遥感图像,处理系统根据光纤陀螺测量获得的颤振数据和成像相机曝光过程的时间编码值计算得到图像退化点扩散函数,对图像反卷积消除平台颤振影响,快速获得清晰的遥感图像。 由于采用了编码曝光,图像复原过程可以使用快速的逆滤波算法,彻底解决频谱值为零而无法逆滤波或者噪声被无限放大的问题。采用光纤陀螺测量平台颤振,可以有很高的采样频率,精确测量得到从低频到高频的宽频颤振数值,而且可测任意二维方向的颤振,符合遥感卫星平台实际颤振情况。该方法用直接逆滤波,在处理速度和复原图像效果方面具有绝对的优势,对于快速获得高分辨率遥感图像具有重要意义。
本项目研究了编码曝光的理论基础,验证提出方法的可行性,并对于编码曝光的码字选取进行研究,得到码字的选取准则和码字的搜索方法。开展基于双轴光纤陀螺的任意复杂颤振实时探测技术研究,得到实时探测颤振方法,利用探测到的颤振数据来估计模糊核,得到快速准确的PSF估计方法。构建了基于编码曝光和颤振探测的系统框架,搭建了实验装置。编写程序,进行仿真实验,对不同的复原方法进行比较,同时将本项目方法在不同的条件下进行验证,得到实验结果;搭建实验平台,进行实际实验,设置不同的颤振条件和拍摄场景,得到实验结果,并在主观和客观方面进行比较。最后编写软件,实现基于编码曝光和颤振探测的遥感图像的仿真和快速恢复。
{{i.achievement_title}}
数据更新时间:2023-05-31
基于 Kronecker 压缩感知的宽带 MIMO 雷达高分辨三维成像
内点最大化与冗余点控制的小型无人机遥感图像配准
基于ESO的DGVSCMG双框架伺服系统不匹配 扰动抑制
宽弦高速跨音风扇颤振特性研究
采用深度学习的铣刀磨损状态预测模型
基于编码曝光序列图像获取的运动模糊复原理论与方法研究
基于测度学习的高光谱遥感图像目标探测方法研究
遥感图像缺失信息的恢复方法研究
目标振动相干激光遥感探测方法研究