【作者】 李伟雄；

【导师】 徐抒岩；

【作者基本信息】 中国科学院研究生院（长春光学精密机械与物理研究所） ， 光学工程， 2012， 博士

【摘要】 随着空间相机分辨率的不断提高，各空间技术强国除了追求相机的高地面分辨率外，对相机的图像采集效率也有了较为严格的要求，并且提出更多的图像应用方向，敏捷成像应运而生并成为了重要的研究方向之一。本文分析了不同敏捷成像模式对高分辨率空间相机像移补偿的要求，在此基础上研究适应飞行器大侧摆和大俯仰姿态下的敏捷成像的像移补偿方法。首先，参照国外敏捷成像卫星，规划出单条带成像、多轨连续条带成像、同轨连续条带成像、多轨立体成像、同轨立体成像以及同轨多目标成像六种在飞行器姿态三轴姿态稳定时敏捷成像模式，研究每种成像模式的特点和作用。按照每种敏捷成像模式对飞行器姿态要求，将六种模式分为侧摆成像、俯仰成像以及侧摆俯仰成像三类。其次，建立瞬态几何成像模型，用简单直观的方法定性分析侧摆成像和俯仰成像两类敏捷成像模式对像移补偿的要求。根据齐次坐标变换矩阵及景点到像面的变换过程，建立了像移速度矢计算模型，并推导了像移速度的精确计算公式。通过精密的像移矢计算模型以及传递函数的评价方法定量分析三类敏捷成像模式对像移补偿的要求。再次，对传统补偿方法下的像移速度相对误差和偏流角误差的误差源进行归纳。通过对传统补偿方法在敏捷成像时的像移速度和偏流角匹配残差、像移速度和偏流角估值误差以及偏流角累积误差的分析，确定传统补偿方法对敏捷成像模式的适应能力，进而提出了适应大侧摆和大俯仰的敏捷成像模式需要解决的三个问题，侧摆成像问题、持续成像问题以及俯仰成像问题。然后，研究敏捷成像的三个问题的解决方法：分析异速匹配和实时偏流调整方法的原理，并通过地面实验和在轨测试的完成两个方法验证。建立了实时偏流调整中的机构调整附加误差计算的数学模型，对实时偏流调整原理以及偏流角误差进行了定量分析,从理论证明了实时偏流调整的可行性。根据横向像移图像的生成原理，建立了横向像移图像的数学模型。提出了组建完成方程组的横向像移图像复原的方法，完成了地面仿真验证和算法精度分析。最后，根据异速匹配、实时偏流调整以及构建完整方程组方法的实施条件，设计了适用于敏捷成像的高分辨率空间相机的敏捷像面，并且提出了基于敏捷像面的适应于飞行器大侧摆和大俯仰姿态的高分辨率空间相机的敏捷成像方案，为高分辨率空间相机敏捷成像的像移补偿技术的发展，提供了理论依据。更多还原

【Abstract】 With the continuous improvement of space camera resolution, in addition to thepursuit of the camera’s high ground resolution, the images’ collection efficiency ofcamera has also been more stringent required by countries that have powerful spacetechnology, and more aspects of application of image were promoted, then agileimaging emerged and became an important one of research directions. Requirementsof the different agile imaging modes to the image motion compensation of highresolution space cameras were analyzed，and method of image motion compensationwas researched to adapt the large angle of roll and pitch attitudes of agile aircraftin this paper.First, with reference to foreign agile imaging satellites, single band imaging,multi-track continuous band imaging, same-track continuous band imaging,multi-track stereo imaging, same-track stereo imaging and same-track multi-targetimaging modes was planed when the three-axis attitudes of aircraft are steady，andthe characteristic and good of each imaging mode was studied. The six modes ofagile imaging were also divided in to three kinds: roll imaging, pitch imaging andboth roll and pitch imaging in according to the requirements to the attitudes ofaircrafts.Secondly, the establishment of instantaneous-state geometric imaging model was established, and used to qualitatively analyze the requirement of the roll imagingand pitch image to the image motion compensation. According to the homogeneouscoordinate transformation matrix and the transformation process from the scene tothe image plane, an image motion velocity vector computational model was built,and a precise formula of image motion velocity was derived. The precise imagemotion vector model and transfer function evaluation method were applied toquantitatively analyze the requirement of three kinds of agile imaging mode to theimage motion compensation.Thirdly, error sources of image motion velocity’s relative error and drift angle’serror were summarized when the conventional compensation method of imagemotion was applied to agile imaging. Matching error, estimation error andcumulative error of drift angle and matching error and estimation error of imagemotion velocity were analyzed to determine adaptability of the conventionalcompensation method to agile imaging, and then three pivotal problems that need tobe solved were discovered in order to fit agile imaging mode of large roll imagingand pitch imaging.The three problems are problem of roll imaging，problem ofpersistent imaging and problem of pitch imaging.Fourthly, three issues of agile imaging solutions were studied. The principle ofmethods of the multi-velocity match of image motion velocity and real-timeadjustment of drift angle were analyzed, and that were validated by experiment onground and test in orbit. A mathematical model of the additional error of drift’sframework was modeled when drift angle was revolving. A theory of creating imageof transverse image motion was promoted by reaching the imaging process of TDICCD. A method of composing integrated group of equations was brought up torecover image of transverse image motion. The method was validated in theemulation environment, and the arithmetic precision was analyzed.Finally, based on the methods of multi-velocity match of image motion velocity,real-time adjustment of, and composing integrated group of equations, a agile imageplane of was designed for agile imaging，and a blue print of agile imaging of high-resolution space cameras was put forward to adapt to large angles of roll andpitch attitudes of aircraft based on the agile image plane，that provide academic baseto the development of image motion compensation of space high resolution cameras’collecting scene agilely.更多还原