为了提高光斑的质心探测精度，要对采集的光斑图在质心位置确定之前进行抑噪预处理，采用5种常用的图像处理方法对光斑图进行抑噪处理。仿真和实验结果表明，两种非线性滤波算法效果最好，极大地提高了质心探测精度，为波前相位探测中的数据处理提供了有益的参考。

In the adaptive optical systems,the precision of Hartmann-Shack wavefront sensor relies on the centroid detecting accuracy,but the noise in the light spot image lead to centroid detecting error. So the noise should be inhibited before estimating centroid. According to analyzing five kinds of common filtering methods by computer simulation and experiment,the two nonlinear filtering algorithms are showed to be the optimal selection. The study will offer valuable reference in enhancement of the precision of wavefront measurement.