Abstract: To improve the precision and effective spectrum width of wavefront power spectrum density(PSD) at mid-spatial scale, double linear interpolation and six-sample-wide interpolation are proposed to patch invalid points and reduce aliasing noise, respectively. Double linear interpolation reduces discrete discontinuities and retains spectral structure of the valid areas, without adding visible middle and high spatial frequency errors existing in zero-filled method. Six-sample-wide interpolation effectively separates the signal from the aliasing noise and thus increases the effective spectrum upper limit of PSD from half of Nyquist frequency to Nyquist frequency. The validity of double linear interpolation and six-sample-wide interpolation is verified by computer simulation and experiment.