条纹投影三维重建中相位提取的抗噪能力研究  

Research on the anti-noise ability of phase extraction in 3D reconstruction of fringe projection

作　　者：刘锦涛 徐鹏 张雯 单硕 李泽 王建华 LIU Jintao;XU Peng;ZHANG Wen;SHAN Shuo;LI Ze;WANG Jianhua(School of Information and Control Engineering,Qingdao University of Technology,Qingdao 266525,China)

机构地区：[1]青岛理工大学信息与控制工程学院,青岛266525

出　　处：《青岛理工大学学报》2024年第5期142-153,共12页Journal of Qingdao University of Technology

基　　金：山东省自然科学基金资助项目(ZR2021MF024)。

摘　　要：对于静态漫反射物体的三维重建,不可避免的噪声干扰是影响条纹投影三维重建的重要因素之一。介绍了相移轮廓术、傅里叶变换轮廓术与小波变换轮廓术,并对其抗噪能力进行了分析和对比。研究发现,相移法在非噪声和噪声干扰条纹图像的三维重建中的相位误差最小。条纹图像不含噪声干扰时,傅里叶变换法与小波变换法的相位误差接近,均大于相移法的相位误差。条纹图像含噪声干扰时,基于直接最大脊的二维小波变换轮廓术的均方根误差(RMSE)比基于直接最大脊的一维小波变换轮廓术和傅里叶变换轮廓术的RMSE降低了45.88%和58.41%。基于评价函数的小波脊提取算法的相位误差要远小于基于最大脊的小波脊提取算法的相位误差。进一步证实了采用Fan小波的抗噪能力要优于采用Morlet小波的,为条纹投影轮廓术(FPP)的不同场景的相位算法选择提供了参考。For the three-dimensional reconstruction of static diffuse reflection objects,unavoidable noise interference is one of the important factors affecting fringe projection three-dimensional reconstruction.Phase-shifting profilometry,Fourier transform profilometry and wavelet transform profilometry are introduced in this study,and their anti-noise capabilities are analyzed and compared.The research found that the phase error of phase-shifting method is the smallest in the three-dimensional reconstruction of fringe images with and without noise interference.When the fringe image does not contain noise interference,the phase error of Fourier transform method and wavelet transform method are similar,both greater than the phase error of phase-shifting method.When the fringe image contains noise interference,the root mean square error(RMSE)of the two-dimensional wavelet transform profilometry based on the direct maximum ridge is reduced by 45.88%and 58.41%respectively compared with the RMSE of the one-dimensional wavelet transform profilometry based on the direct maximum ridge and the RMSE of Fourier transform profilometry.The phase error of the wavelet ridge extraction algorithm based on the evaluation function is much smaller than that of the wavelet ridge extraction algorithm based on the maximum ridge.It further confirms that the noise resistance of using Fan wavelet is superior to using Morlet wavelet.This provides a reference for the selection of phase algorithms for different scenarios of fringe projection profilometry(FPP).

关 键 词：相移 傅里叶变换 小波变换 评价函数 抗噪能力 

分 类 号：O439[机械工程—光学工程]