Abstract: The accurate center location of fringe pattern is vital to measurement accuracy of interferometric method.For complex background fringe pattern，the location of fringe center is seriously affected by noise.The key noises are laser speckle and non-homogeneous background intensity，to suppress these noises pre-processing of Butterworth Low-pass filter and background compensation is utilized，then binary fringe is obtained based on area segmentation.Thinning the binary fringe and take it as the rough center of fringe，gray scale sampling around the rough center is fitted by polynomial，and the center of fringe is calculated by extremum method.At last，center location of a simulated fringe pattern is calculated by this method，the result shows that the location obtained by this method is more accurate than that obtained by directly thinning the fringe.

Key words: fringe pattern, fringe analysis, image segmentation, image thinning

摘要： 干涉条纹中心的定位精度对于干涉法的测量精度至关重要。对于背景复杂的干涉条纹图像，噪声严重影响了条纹中心的定位。激光干涉图像中主要噪声为激光散斑噪声和背景亮度不均匀，采用Butterworth低通滤波和背景补偿对图像预处理以抑制噪声的影响，再进行区域分割得到二值化条纹，细化二值条纹得到初步的条纹中心，在细化条纹的周边区域通过多项式拟合求极值的方法计算条纹中心。最后用该定位方法计算了模拟条纹图像的条纹中心，计算结果表明该方法比直接细化条纹的方法更精确。

关键词: 干涉条纹, 条纹分析, 图像分割, 图像细化