摘要
相位法是采用正弦投影光和移相技术的一种非接触测量方法．其中从主值相位（-π，π）恢复绝对相位，即相位展开问题（Phaseunwrapping）是一个不适定问题．国内外的研究者进行了大量的研究工作，至今尚无满意的一般解决方法．本文利用相位法条纹图像的纹理特性，利用Gabor小波变换检测条纹纹理的边界，从而把相位场分割为一致的一些区域，使相位展开分别在一致的区域内进行．试验结果表明，本文的算法在存在真实的相位不连续等情况时仍能取得满意的结果，它比利用极点连线等一类的算法更为可靠．
Unwrapping a 2-D inconsistent principal-value phase fields of PMP,that may contain regions of discontinuous phase, is an ill-posed problem. This paper proposes an algorithm to detect the texture boundaries of deformed fringe image by using multiscale Gabor wavelet transformation. The detected texture boundaries form cut lines and segment the phase fields of PMP to consistent regions. The deformed fringe texture boundaries are detected through three stages. The input image is first processed through a bank of Gabor wavelet filters. Following feature extraction using Gabor wavelets, local competitive and cooperative interactions of these feature are introduced. Interscale interactions help in localizing the line ends and illusory contours. The final stage integrates similar features aiding in texture boundary completion. The approach is demonstrated by application to synthetic and real PMP image of body models.
出处
《计算机学报》
EI
CSCD
北大核心
1998年第12期1096-1102,共7页
Chinese Journal of Computers
基金
国家自然科学基金
