高速飞行器共形窗口动态像差的双模自适应校正研究

The dynamic aberrations of conformal optical window in high-speed aircrafts are generated from the window surface shape and aero-optical effects. Adaptive optics (AO) technology can theoretically correct both types of aberration at the same time, which has great potential for application. However, traditional AO is confronted with bottlenecks such as too complex system structure and insufficient correction bandwidth. In this project, the multidisciplinary integrated simulation analysis is used to obtain the dynamic aberrations of the conformal window which are produced by the combined effect of many factors. The temporal and spatial variation characteristics of the aberration and its influencing mechanism on image quality are analyzed. Then the adaptive optical correction requirements are proposed. The adaptive optical system for correcting the dynamic aberration of the conformal window employs a wavefront sensorless design and a dual-mode correction method. The first correction mode utilizes the specific mode modulation produced by the deformable mirror to estimate aberrations and then correct them, mainly correcting large dynamic range, low time frequency errors. The second correction mode performs linear phase retrieval and image restoration based on small-aberration assumption, mainly correcting small dynamic range, high time frequency errors. The research achievements can provide new technical approaches for the optimal design and aberration correction of conformal optical system for high-speed aircraft and further promote the practical use of AO technology in this field.

高速飞行器共形光学窗口的动态像差由窗口面形和气动光学效应共同作用产生。自适应光学技术理论上可以对这两类像差同时进行校正，应用潜力巨大。但传统自适应光学面临系统结构过于复杂，校正带宽不足等瓶颈。本项目通过多学科集成仿真分析获取多因素共同作用的共形窗口动态像差，对像差的时空变化特性和像质影响机理进行分析，提出自适应光学校正指标。校正共形窗口动态像差的自适应光学系统采用无波前传感器设计和双模校正方式。第一种校正模式利用变形镜产生的特定模式调制估计像差并校正，主要校正大动态范围、低时间频率误差。第二种校正模式基于小像差条件进行线性相位恢复和图像复原，主要校正小动态范围、高时间频率误差。研究成果可以为高速飞行器共形光学系统的优化设计和像差校正提供新的技术途径，进一步推动自适应光学技术在本领域的实用化发展。

本项目针对高速飞行器共形窗口动态像差引起的像质下降问题，在分析像差时空特性的基础上，开展了无波前传感器自适应光学快速校正方法、计算鬼成像方法、大动态夏克-哈特曼波前传感方法及图像复原方法等方面的研究。具体包括：（1）提出了基于模型的无波前传感器自适应光学同步校正算法和顺序校正算法，与传统算法相比，校正N阶像差模式所需图像数量分别从2N+1和3N减少到N+2和2N，大幅提高了校正速度。（2）提出了基于通用模型的无波前传感器自适应光学系统自标定方法和校正方法，该方法同时适用于点目标和扩展目标，无需使用波前传感器标定变形镜影响函数，简化了系统的应用条件; （3）将基于模型的无波前传感器自适应光学技术应用于成像光学系统自身像差的自动校正，具有计算简单、校正速度快、精度高的优点; （4）提出了一种分类-复原深度学习网络，用于克服湍流扰动对计算鬼成像像质的影响；（5）提出了一种扩大哈特曼-夏克波前传感器动态范围的迭代外推法并研制了基于FPGA的波前实时处理机。与合作单位共同完成了共形窗口绕流激波、层流、湍流以及头罩窗口光学效应仿真计算研究，建立了头罩窗口光学效应等效像差模型，开发了共形光学窗口动态像差分析软件。上述研究在自适应校正系统标定、大动态像差的快速探测和校正、快速图像复原等方面取得了进展，为共形窗口动态像差的校正提供了多种可供参考的技术途径，有利于促进自适应光学技术在相关领域的实用化发展。