利用Zernike多项式进行最小二乘拟合来计算矩形元件的离焦及像散,不需要波面数据在圆域内正交，从而避免了波面数据缺失区域的数据插值。通过计算圆形口径数据，与商业软件计算结果比较，差别小于0.005倍波长，证明了该方法计算离焦及像散的可行性。分析了不同长宽比下矩形元件的离焦及像散。结果表明用商业软件计算矩形口径元件像散时，随着长宽比的增加，偏差会明显增大。

Least square method is used to fit power and astigmatism of rectangular pupil wavefront data, which avoids data interpolation of the blank area in circular domain. In the experiments, the maximal difference of power and astigmatism of circular domain data between commercial software and our method is less than 0.005 wavelength, which proves the proposed method’s feasibility. For rectangular pupil wavefront data, the differences of astigmatism increase as the rectangle element’s length-width ratio increases. Therefore the proposed method is more applicable for rectangular pupil wavefront calculation.