面向高精度星敏感器姿态测量的高动态星图处理方法研究

As the space missions become increasingly complex and precise, requirements for attitude measurement system of the spacecraft are significantly growing. High accuracy, driftless and absolute output attitude are main characteristics of the star tracker, for which the star tracker gains great attention and development, and it is considered to be one of the most promising optical attitude measurement devices. From present international research situations, the dynamic performance is becoming the principal technology barrier and key topic, due to the restriction of the optical imaging system. This project fully considers the characteristics of the star spot under dynamic conditions. The proposed methods include: dynamic star image pre-processing and background noise estimating method based on correlation filtering and mathematical morphology operation; dynamic star energy restoration method based on nonlinear filter and established motion model; region distinction of star spot based on partial image differential and optical flow analysis applied to low dynamic and high dynamic range respectively. The signal-to-noise ratio and the position accuracy of the star spot will be improved. Regions of star spot can be distinguished effectively so that tracking problem of star spot under conditions of large angular velocity and large angular acceleration can be solved. The experiment methods include dynamic star simulation, dynamic astronomical observation and on-orbit star image verification. The dynamic range of the star tracker can be enhanced with arc-second attitude accuracy, which has significance for the applications of the star tracker in high-resolution earth observation, deep space exploration and intercontinental missile system with more demanding conditions.

随着航天任务逐渐复杂化和精密化，航天器对姿态测量系统的要求不断提高。以高精度、无漂移、可输出三轴绝对姿态为主要特征的星敏感器得以极大关注和发展。由于光学成像系统的制约，星敏感器的动态性能是目前的主要壁垒。其中，高动态星图的处理是主要制约因素和研究重点。本项目充分挖掘运动情况下星点的特征，提出类星点分布模板相关滤波结合大尺度结构算子数学形态学背景估计算法；提出基于星图泊松分布噪声和运动模型的非线性星点能量复原方法；提出分别适用于低动态和高动态的局部微分星点区域辨别方法和连续星图光流分析方法。有效提高星图中星点的信噪比和星点位置精度，确定星点窗口范围，解决大角速度和大角加速度情况下跟踪失效的问题。项目拟采用多种实验手段进行验证。将突破星敏感器动态性能的瓶颈，保证高动态环境下角秒级姿态测量精度和稳定性，对星敏感器在高分辨率对地成像、深空探测以及运动条件更为苛刻的洲际导弹系统中的应用具有重要意义。

本项目面向敏捷成像、视频拍摄等微纳遥感成像过程对高动态星敏感器技术的迫切需求，开展了高精度高动态星敏感器小型化关键技术研究，针对高动态下星图信噪比低/星点目标提取困难、跟踪模式中大角速度大角加速度情况下跟踪失效、在轨运行中太阳光/地气光等杂散光带来极强干扰的问题，进行低信噪比微弱点目标提取方法研究、星敏感器鲁棒跟踪与快速识别方法研究、基于能量信息挖掘的杂散光抑制方法研究，有效提高动态星图中星点信噪比和星点位置精度，确定跟踪窗口，解决高动态情况下跟踪失效及易受噪声干扰的问题。实现星敏感器精度5″，动态性能5°/s，更新率＞30Hz，参与研制的星敏感器已应用于高分辨率对地成像、小行星深空探测、低轨全球覆盖通信、夜光成像等航天任务中，有效提升我国高精尖航天装备的自主研制能力与水平，未来在微小卫星及高精度遥感卫星领域有特别的优势和广阔的应用前景。