天文目标低阶自适应光学校正和图像事后处理联合研究

Wavefront aberrations caused by atmospheric turbulence affect the image quality seriously for the optical imaging of astronomical objects under dynamic turbulences. It is necessary to correct wavefront aberrations and process the imaging afterwards. Real-time requirements of adaptive optics system and image post-processing are high in order to adapt to the dynamic nature of the turbulence, which results in technology difficulty and high development costs. Regarding general point spread functions of imaging system at different time under turbulence as different channels, we put forward an image post-processing method by introducing a multi-channel blind identification theory. The post-processing method has no real-time requirements . Considering low-order wavfront distortions are the main part of the total aberration caused by atmospheric turbulences, we intend to establish an integrated and low-cost imaging method for the optical imaging of astronomical objects under dynamic turbulences by combing a low-order adaptive optics system with the multi-channel blind identification signal recovery. The research includes the basic theory analysis, the control algorithms design, simulations, experiments and results analysis of the wavefront sensorless mode-based adaptive optics system. Based on the optical imaging mode of astronomical objects under dynamic turbulences, a linear model of blind multi-channel identification will be constructed. Regularization methods, convex optimization solving and deconvolution algorithm will be studied. Finally, we will focus on overall optimization of imaging correction and image post-processing. The integrated method in the project will provide the theoretical basis of low-cost imaging correction and post-processing method for astronomical objects under turbulent conditions.

天文目标光学成像时，大气湍流引起的波前像差严重影响成像质量，需要进行校正和恢复。为适应湍流动态性，对自适应光学和图像事后处理的实时性要求高，技术实现难度大。将湍流条件下不同时刻的成像系统综合点扩散函数看作不同通道，引入多通道盲识别理论，提出无实时性要求的天文目标图像事后处理方法。由于大气湍流引起波前畸变中低阶像差占大部分，将基于模型的低阶无波前探测自适应光学系统和多通道盲识别图像恢复方法结合起来建立适应大气湍流变化的低成本天文目标光学成像校正和图像事后处理的联合系统。具体包括基于模型的快速无波前探测自适应光学系统用于天文目标成像校正的闭环控制算法理论研究、控制算法设计、仿真和实验研究；根据天文目标成像的特点建立多通道线性估计模型、研究正则化参数选取方法、凸优化问题求解及解卷积算法；成像校正和图像事后处理方法的整体优化等，最终为湍流条件下天文目标低成本成像校正和高分辨率图像恢复提供理论基础。

相比常规自适应光学，无波前探测自适应光学系统更为广泛的应用环境，如暗弱目标成像、扩展目标成像及湍流较强的空间光通信系统等等，但收敛速度慢是该类系统最大的缺陷。本项目研究了基于模型的无波前探测自适应光学系统理论基础、设计了控制算法、进行了软件模拟和实验研究。并将湍流条件下不同时刻的成像系统综合点扩散函数看作不同通道，分析了多通道盲识别方法在自适应光学图像后处理中的理论和应用可行性。.重要结果一：以变形镜本征模作为基函数时，可以使系统获得变形镜接近理想校正时的校正效果，系统的收敛速度只跟变形镜的单元数有关。重要结果二：本项目基于光学和数学理论严格证明了成像焦平面上的掩模强度分布的变化与畸变波前的微分平方和成正比，并将其推广到扩展目标成像。重要结果三：噪声情况下点目标成像和扩展目标成像校正结果表明基于模型的无波前探测自适应光学系统具有较强的抗噪能力。重要结果四：搭建实验平台验证了成像掩模探测器信号与畸变波前的微分平方和之间的线性关系，结果表明不论取哪种掩模，成像掩模探测器信号MDS和波前相位的平均梯度平方和MSG之间总是具有线性关系，且系数非常接近理论值。重要结果五：将多通道盲识别理论用于成像复原的研究结果表明大气湍流条件下，不同时刻的点扩散函数之间存在互质关系。复原图像与原始目标之间的均方误差在十万分之一量级，研究结果为湍流退化图像的高分辨率恢复提供了理论基础。重要结果六：以97单元MEMS变形镜为基础搭建了基于模型的无波前探测自适应光学系统，测量了影响函数并完成了闭环校正实验。闭环校正结果表明，基于模型的无波前探测自适应光学系统有一定校正效果，系统可以达到收敛、光斑能量趋于集中。但在校正速度上与理论结果有一定差距，需要一定次数的迭代才能收敛。以上重要研究成果为基于模型的无波前探测自适应光学系统应用及AO校正后图像复原应用提供了重要理论和技术支撑。