Accurate and Fast Tomographic Wavefront Reconstruction for NextGeneration Multi Object Adaptive Optics System

为了实现下一代大型地基望远镜宽视场中多个物体的同步自适应光学（AO）校正，我们研究了层析成像波前重建的新方法。我们发现了一种新算法，它使用先前时间步长的测量结果作为附加信息，可以显着提高断层扫描估计的准确性并显着拓宽可用视野。为了通过 GPGPU 并行处理实现更快的层析估计，建议通过改变层析重建中使用的共轭梯度法的初始猜测并增加并行处理的数量来实现必要的速度。

In order to realize simultaneous adaptive optics (AO) corrections for multiple objects in wide field view of next generation large ground based telescopes, we examined new methods of tomographic wavefront reconstruction. We found a new algorithm, which uses measurements from previous time step as additional information, can significantly improve the accuracy of the tomographic estimation and significantly widen the available field of view. For faster tomographic estimation with parallel processing with GPGPU, it is suggested that the necessary speed can be achieved by changing the initial guess for the conjugate gradient method used in the tomographic reconstruction and by increasing the number of parallel processing.