有效去除数字散斑条纹图中的噪声是散斑干涉测量技术中的关键问题。提出了一种基于Goldstein滤波的数字散斑条纹图平滑方法。该方法需要将散斑条纹图中的干涉相位转换为矢量空间中的单位矢量，并进行快速傅里叶变换(FFT), 得到其频谱，然后对频谱进行加权处理，从而抑制噪声的频率成分，再将加权处理后的频谱变换到空间域，计算干涉相位，得到原始散斑条纹图的滤波结果。将该滤波方法运用于四步相移数字散斑干涉条纹图像处理。实验结果表明，该方法在滤除散斑噪声的同时能够有效地保护散斑条纹图的轮廓和细节信息，增强了散斑干涉条纹的对比度。

Getting rid of the noises is the key problem in digital speckle pattern interferometry (DSPI). A method based on Goldstein filtering for digital speckle fringe pattern smoothing is proposed. This algorithm maps the raw speckle interferogram to unit vectors in vector space and transforms them to frequency domain by fast Fourier transform (FFT). Then the weighting function is selected to weight the spectrum. The frequency of the speckle noises can be suppressed. Finally, the processed spectrum is transformed to space domain by inverse FFT and phases of filtering results are calculated. The filtering technique is used for the four-step phase-shifting digital speckle image processing. The experimental result shows that the speckle noise in the speckle fringe patterns is removed efficiently and the information of the fringe contour and structure is well preserved. The proposed method can increase the contrast of DSPI fringes.