为了减少波前探测斜率噪声对基于哈特曼－夏克波前传感器的Zernike复原模式的影响，采用对波前复原矩阵斜率噪声影响的相关分析和线性解耦变换的方法，得到一组基于哈特曼－夏克传感器斜率探测噪声的Zernike最优复原模式，通过合理剔除波前模式复原中易受噪声影响的模式组合，减少了波前测量噪声引起的波前模式复原误差；采用蒙特卡罗随机试验法，验证了一子孔径成方形排布的波前传感器对于波前斜率噪声的受影响程度，仿真结果显示，复原模式系数误差由0．0212λ下降为0．0048λ。结果表明，在剔除部分模式项后，最优复原模式统计优化对于滤除波前斜率探测噪声有一定的作用，提高了波前探测器的探测精度和复原能力。

Reconstruction modes of Hartmann-Shack wavefront sensor have direct influence on the precision of sensing and reconstruction of incident wavefront. Because reconstruction modes of Zernike polynomials based on Hartmann-Shack wavefront sensor were easy to be influenced by the gradient sensing noise,optimal reconstruction modes according to the wavefront sensing noise were put forward by the correlation analysis and decoupling transform of wavefront reconstruction matrix. Furthermore,the error of wavefront model reconstruction thanks to the wavefornt sensing noise can be reduced by filtering the portion of the optimal modes. By Monte-Carlos' random experiment,the antinoise level of Hartmann-Shack wavefront sensor whose subapertures distributes at a square was verified. As a result,the reconstruction error reduced from 0.0212λ to 0.0048λ after getting rid of the part of optimal reconstruction. The way to analyze the optimal reconstruction modes has an effect on filtering the wavefront gradient sensing and improve the sensing precision of wavefront sensor.