QQ客服 官方微博 反馈留言

题名	空间相机振动参数检测及降质图像复原
作者	刘海龙
学位类别	博士
答辩日期	2015-05
授予单位	中国科学院大学
导师	韩诚山
关键词	空间相机 振动参数检测 降质图像复原
其他题名	Space Camera Vibration Parameters Detection and Blurred Image Restoration
学位专业	机械电子工程
中文摘要	遥感卫星在轨工作期间，在太阳光压、地球引力梯度、空间微小陨石碎片等外部因素，及反作用轮、太阳帆板和其他运动部件的机械与热效应等内部因素的影响下，卫星平台的空间姿态会受到扰动，传递到遥感器平台，表现为平台的振动现象，这会导致空间相机成像过程中，景物在像面的投影产生相对移动，从而降低图像质量。随着遥感器光学系统设计及制造水平的进步和光学成像器件性能指标的提升，空间相机角分辨率在不断的提高，导致相机对振动也越来越敏感，振动开始成为影响高分辨率遥感器成像质量的重要因素。检测空间相机振动参数，对研究卫星平台振动规律，提升遥感器性能具有重要意义，也为振动抑制及降质图像复原提供了数据基础。论文的主要研究工作如下：从光学系统调制传递函数出发，分析了振动对于MTF的影响，推导了MTF和振动参数间的计算公式，并通过仿真实验，使用评价参数直观地展现了振动幅度、振动频率和积分级数等参数对于TDICCD成像质量的影响。针对焦平面为机械拼接结构的TDICCD推扫相机，提出了一种基于拼接区域对同一景物重复成像的振动检测方法，达到了不增加额外设备，仅利用相机自身结构精确测量相机振动参数的目的；同时为了提高检测方法的适用性，基于卷帘快门CMOS图像传感器自相关成像，提出了一种在推扫相机和凝视相机上均能有较好应用的振动参数检测方法，实现了利用低帧频图像序列检测高频振动的目的，大幅度降低了检测算法的数据传输及处理压力，为检测算法的星上嵌入式实现提供了可能。根据传感器成像原理，提出了一种利用全局快门相机模拟TDICCD和卷帘快门CMOS成像的方法。基于该方法设计了振动参数检测方法验证实验，并搭建实验平台、设计实验参数，分别对本文提出的两种振动参数检测方法进行验证实验。实验结果表明两种检测方法均有较好的检测精度：对于基于机械拼接结构TDICCD相机的振动参数检测方法周期检测相对误差小于2%，振幅检测绝对误差小于2个像元；对于基于面阵CMOS自相关成像的振动参数检测方法，周期相对误差小于1%，振幅绝对误差小于1个像元。根据振动检测方法验证实验中检测到的振动参数，对振动降质图像进行复原。提出一种振动点扩散函数离散化方法，对任意已知运动函数的振动，通过这种像素级插值离散化算法，均可快速计算其点扩散函数。利用得到的离散化点扩散函数对降质图像进行逐行复原，并使用评价参数客观评价复原前后图像质量，评价结果表明复原效果明显，证明了本文提出的两种振动参数检测方法的正确性。
英文摘要	During on-orbit working process of the remote sensing satellite, the gesture of the satellite platform will be disturbed under the influence of external factors, such as sunlight pressure, gravity gradient, spatial micrometeorites debris, etc., or internal factors, such as, solar panels, reaction wheels, the mechanical and thermal effects of other moving parts, etc. The disturbance shall be passed to the remote sensor platform, assumed as phenomena of the platform vibrations. During the integration of space camera imaging process, the vibrations will lead to relative movement of the scene projection, thereby reducing the image quality. With the progress of the remote sensing optical system’s designing and manufacturing level and the enhancement of the optical imaging device’s performance, the angle resolution of the spatial camera has continuous improvement, also lead the camera more sensitive to vibrations. Vibrations start to become an important factor affecting the quality of high-resolution remote sensing device. Detecting space camera vibration parameters is fairly significant to research of the vibration law of the satellite platform and enhancement of the remote sensing performance. But also provides data basement for the vibration suppression and restoration of the degraded image. The main work of this paper is as follows: From the optical system’s modulation transfer function, we will analyse the influence of vibrations to the MTF, derive the formulas between MTF and vibration parameters, and with simulation experiment show the vibration influence of the vibration amplitude, frequency and other parameters to TDICCD image quality visually by using evaluation parameters. For focal plane mechanical splice structure TDICCD scanning camera, we propose a method for detecting vibration parameters, by utilizing spliced region which imaging the same scene repeatedly. Reaching the purpose of no additional facilities, using only the camera self structure to accurately measure the parameters of camera vibrations. For more application, based on a rolling shutter CMOS image sensor autocorrelation imaging, we propose a vibration parameters detecting method, which can be preferably used at both scanning camera and staring camera, achieving the purpose of utilizing image sequences of low frame rate to detect high frequency vibrations, greatly reducing the image transmission and processing pressure, making the detection algorithm satellite embedded transplant possible. According to the sensor imaging principle, presents an analog imaging methods which simulate TDICCD and rolling shutter CMOS imaging by using global shutter camera. Based on the method, we design a checking experiment to validate the vibration parameters detecting methods, and then construct experimental platform, design experimental parameters, prove that both of the two vibration parameter detection methods to be true. Validation results showed that both detection methods are accurate: for detection method based on focal plane mechanical splice structure TDICCD, the cycle detecting relative error is less than 2%, the amplitude detecting absolute error is less than 2 pixel; for detection method based on plane-array CMOS autocorrelation imaging, the cycle detecting relative error is less than 1%, the amplitude detecting absolute error is less than 1 pixel. Through the vibration detection method validation experiments, we get the parameters of the detected vibration, then restore the vibration degraded image. Presents a vibration point spread function discretization method. For any vibration that known function of movement, through this pixel-level discrete interpolation algorithm can quickly calculate the point spread function. Restore the blurred image by utilizing the discrete point spread function and evaluate the image quality objectively by using evaluation parameters that before and after the restoration, the evaluation results show that the restoration effect is obvious, indirect prove the correctness of two proposed vibration parameter detection method.
公开日期	2015-12-24
内容类型	学位论文
源URL	[http://ir.ciomp.ac.cn/handle/181722/48872]
专题	长春光学精密机械与物理研究所_中科院长春光机所知识产出
推荐引用方式 GB/T 7714	刘海龙. 空间相机振动参数检测及降质图像复原[D]. 中国科学院大学. 2015.

个性服务

查看访问统计

相关权益政策

暂无数据

收藏/分享

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。