在分析影响Hartmann-Shack(H-S)传感器质心探测精度因素的基础上，提出结合自平方算法、大津算法、光斑位置区域自适应标定、回归局部阈值法等方法，给出一套适合波前像差测量的质心自适应探测方法，该方法具有准确、自适应性强的特点。经验证，与传统的质心探测方法相比，该方法能准确识别并保留更多的有效光斑信息，提高质心探测的精度，扩大H-S传感器的适用范围。

A new adaptive method is given in this paper to detect the centroid of images caught by Hartmann-Shack (H-S) wavefront sensor after analyzing the reason of causing error. This method is based on the self-square arithmetic, Otsu arithmetic, light spots template matching and special threshold method in different sub-area. It is more accurate and adaptive than general methods. The experimental results show that more available information of light spots are kept through this way, which improves the accuracy of centroid detection and extends the applicability of H-S wavefront sensor.