摘要
为了准确地提取出结构光测量中变形条纹的瞬时相位，需对变形条纹图像进行滤波。在OWT—SURE—LET小波去噪的基础上，用冗余小波代替正交小波，用DB8小波代替Sym8小波，利用一个六变量的收缩函数代替四变量的收缩函数，提出了一种改进的SWT—SURE—LET小波滤波算法，该算法速度快、稳健性强。峰值信噪比和相位均方根值提取精度两个指标表明，无论何种类型的变形条纹，无论变形条纹的载频怎么变化，改进算法都是有效的。和其他经典的及最新的去噪算法相比，改进后的算法具有最好的结果和稳健性。
To retrieval the instantaneous phase of the fringe pattern accurately in Fringe Projection Profilometry, the fringe pattern should be denoised. Based on OWT_ SURE_ LET wavelet denoising method, using stationary wavelet substitute for orthogonal wavelet, using DB8 substitute for Sym8, and using a shrinkage function with six parameters substitute for that with four parameters, an improved wavelet de-noising algorithm is proposed. Compared with other de-noising methods, the new method is very fast and robust. Experiments show that this new method is effective in the sense of peak signal-to-noise ratio （PSNR） and root square mean error （RSME） of the instantaneous phase of fringe pattern, no matter what types of the fringe pattern or the carrier frequency in the fringe pattern.
出处
《中国激光》
EI
CAS
CSCD
北大核心
2013年第B12期190-195,共6页
Chinese Journal of Lasers
关键词
信号处理
小波去噪
贝叶斯估计
相位提取
三维测量
signal processing
wavelet de- noising
Bayes estimation phase retrieval
three- dimensional measurement
