多层共轭自适应光学(MCAO)等技术可以有效改善湍流大气的干扰。为了MCAO的设计和性能优化，需要对址点的大气湍流轮廓线进行测量。基于面源目标像相对运动方差或协方差统计反演的湍流轮廓线测量方法得到的反演线性方程组需要通过离散化获得，离散化误差导致观测视场不能太大。较小的视场使方程数目受限，导致湍流轮廓线反演结果受运动方差或协方差测量误差的影响。以PML方法为研究实例，提出了分层积分系数矩阵方法，该方法可以显著改善离散化误差，大幅增加观测视场范围，降低协方差误差对测量结果的影响，提高湍流轮廓线的反演精度。数值模拟结果表明，基于分层积分系数矩阵的PML方法的视场可达到400″，测量误差也可大幅度降低。

The multi-conjugate adaptive optics (MCAO) can effectively reduce the disturbance of turbulent atmosphere. For the design and performance optimization of MCAO technology, it is necessary to measure the atmospheric turbulence profiles of the site. The inversion linear equations of the methods based on differential image motion variance or covariance of extended object are obtained through discretization. The discretization error causes the effective field of view (FOV) not to be very wide. A small FOV limits the number of equations, resulting in turbulence profile inversion results being affected by motion variance or covariance measurement errors. In this paper, taking the PML method as an example, we propose a layer integral coefficient matrix method. This method obviously reduces the discretization error, greatly improves the range of observation FOV, reduces the influence of the covariance error on the measurement results, and improves the accuracy of the turbulence profile inversion. The simulation results show that the FOV of the PML method using layer integral coefficient matrix can reach 400″, and the measurement error is also greatly reduced.