能量分布识别相位畸变的自适应双域滤波方法
|
摘要
在基于能量分布识别相位畸变方法中,条纹图像的质量是影响畸变识别精度的主要因素之一。为了准确辨别条纹相位畸变,需对条纹图像进行滤波处理,滤除图像中的噪声以及背景信息等。提出了一种基于能量分布的自适应图像滤波方法,通过傅里叶变换将图像转换为能量分布图,根据畸变与未畸变图像能量弥散度差值大小对小波滤波的阈值进行调整,旨在寻求能量弥散度差值最大的阈值作为针对当前图像的滤波方法可接受阈值。通过对单空域、单频域以及双域滤波方法进行实验验证,结果表明单一使用空域或频域滤波处理,畸变图像与未畸变图像能量分布差别在1.4%～3.2%,使用非自适应双域滤波方法进行滤波处理,能量分布差别在7%以上,而采用自适应双域滤波方法进行滤波处理,能量分布差别超过8%。自适应双域滤波方法显著提高了能量的识别精度。
In the method of identifying phase distortion based on energy distribution,the quality of fringe image is one of the main factors affecting the accuracy of distortion recognition.In order to accurately distinguish the phase distortion,the fringe image needs to be filtered to filter out noise and background information in the image.An adaptive image filtering method based on energy distribution is proposed,which converts an image into an energy distribution map by Fourier transform.The threshold of wavelet filtering is adjusted according to the difference between the dispersive and undistorted image energy dispersion.It is intended to seek a threshold with the largest difference in energy dispersion as an acceptable threshold for the filtering method for the current image.Experimental verification is performed on single spatial domain,single frequency domain and dual domain filtering methods.The results show that the energy distribution between the distorted image and the undistorted image is between 1.4 % and 3.2 % in a single spatial or frequency domain filtering process.The non-adaptive dual-domain filtering method is used for filtering,the energy distribution is more than 7 %,while the adaptive dual-domain filtering method is used for filtering,and the energy distribution difference exceeds 8 %.The adaptive dual-domain filtering method significantly improves the energy recognition accuracy.
In the method of identifying phase distortion based on energy distribution,the quality of fringe image is one of the main factors affecting the accuracy of distortion recognition.In order to accurately distinguish the phase distortion,the fringe image needs to be filtered to filter out noise and background information in the image.An adaptive image filtering method based on energy distribution is proposed,which converts an image into an energy distribution map by Fourier transform.The threshold of wavelet filtering is adjusted according to the difference between the dispersive and undistorted image energy dispersion.It is intended to seek a threshold with the largest difference in energy dispersion as an acceptable threshold for the filtering method for the current image.Experimental verification is performed on single spatial domain,single frequency domain and dual domain filtering methods.The results show that the energy distribution between the distorted image and the undistorted image is between 1.4 % and 3.2 % in a single spatial or frequency domain filtering process.The non-adaptive dual-domain filtering method is used for filtering,the energy distribution is more than 7 %,while the adaptive dual-domain filtering method is used for filtering,and the energy distribution difference exceeds 8 %.The adaptive dual-domain filtering method significantly improves the energy recognition accuracy.
引文
[1] SUN Yuchen,GE Baozhen,ZHANG Yimo.Review for the 3D information measuring technology[J].Journal of Optoelectronics · Laser,2004,15(2):248-254.(in Chinese)孙宇臣,葛宝臻,张以谟.物体三维信息测量技术综述[J].光电子·激光,2004,15(2):248-254.
[2] ZHANG Qican,SU Xianyu.Research progress of dynamic three-dimensional shape measurement[J].Laser & Optoelectronics Progress,2013,50(1):1-14.(in Chinese)张启灿,苏显渝.动态三维面形测量的研究进展[J].激光与光电子学进展,2013,50(1):1-14.
[3] Ko S J,Lee Y H.Center weighted median filters and their applications to image enhancement[J].IEEE Transaction on Circuits Systems,1991,38(9):984-993.
[4] DENG Xiuqin,XIONG Yong.Weighted median filter algorithm for image processing[J].Computer Technology and Development,2009,19(3):46-48.(in Chinese)邓秀勤,熊勇.用于图像处理的加权中值滤波算法[J].计算机技术与发展,2009,19(3):46-98.
[5] Srinivasan Y,Liu H C,Hallioua M.Automated phase measuring profilometry of 3-D diffuse objects[J].Appl Opt,1984,23(18):3105-3108.
[6] LI Junfeng,YANG Feng,HUANG Jing.A fast algorithm for improved dimensional bilateral filter[J].Journal of Circuits and Systems,2013,18(1):137-143.(in Chinese)李俊峰,杨丰,黄靖.一种改进的增维型双边滤波器的快速算法[J].电路与系统学报,2013,18(1):137-143.
[7] PEI Zhipeng,LI Xin.Research on adaptive image denoising method based on wavelet transform[D].Huainan:Anhui University of Science and Technology,2018.(in Chinese)裴志鹏,李昕.基于小波变换的自适应图像去噪方法的研究[D].淮南:安徽理工大学,2018.
[8] ZHANG Hongwei.Research and implementation of image denoising method based on MATLAB[J].Journal of Daqing Normal University,2016,36(3):1-4.(in Chinese)张宏伟.基于MATLAB的图像去噪方法的研究与实现[J].大庆师范学院学报,2016,36(3):1-4.
[9] DONG Jian,DENG Guohui,LI Jinwu.Research on image transformation based on two-dimensional Fourier Transform[J].Fujian Computer,2015,31(9):102-103.(in Chinese)董健,邓国辉,李金武.基于二维傅里叶变换实现图像变换的研究[J].福建电脑,2015,31(9):102-103.
[10] Goldberg RR.Fourier transforms[J].Physics Today,1961,14(10):52-52.
[11] CHEN Haiyun,CHEN Cheng,PENG Baojin,et al.Fourier analysis of the fringe signal of a fiber grating Mach-Zehnder interferometer[J].Acta Photonica Sinica,2016,45(9):173-178.(in Chinese)陈海云,陈成,彭保进,等.光纤光栅马赫-曾德干涉仪条纹信号的傅里叶分析[J].光子学报,2016,45(9):173-178.
[12] An Ping.Processing and analysis of regular interference fringe images[D].Hohhot:Inner Mongolia University of Technology,2017.(in Chinese)安平.规则干涉条纹图像的处理与分析[D].呼和浩特:内蒙古工业大学,2017.
[13] Auslander L,Grunbaum F A.The Fourier transform and the discrete Fourier transform[J].Inverse Problems,1989,5(2):149.
[14] REN Xuhu,SONG Shanshan,QI Yaoguang.Application research of fourier transform profilometry based on wavelet analysis[J].Journal of China University of Petroleum:Natural Science,2007,(1):148-153.(in Chinese)任旭虎,宋珊珊,綦耀光.基于小波分析的傅里叶变换轮廓术应用研究[J].中国石油大学学报:自然科学版,2007,(1):148-153.
[15] Bacchelli S,Papi S.Filtered wavelet thresholding methods[J].Journal of Computational and Applied Mathematics,2004,164-165(1):39-52.
[16] Roberge D,Sheng Y.Optical wavelet matched filter[J].Applied Optics,1994,33(23):5287-93.
[17] YANG Haimei.Extraction and analysis of phase information for high-speed flow field laser interferometry[D].Yantai:Yantai University,2017.(in Chinese)杨海梅.高速流场激光干涉检测相位信息的提取与分析[D].烟台:烟台大学,2017.
[18] WEN Jia,ZHAO Junsuo,WANG Cailing,et al.Interference hyperspectral image decomposition based on improved MCA[J].Spectroscopy and Spectral Analysis,2016,36(1):254-258.(in Chinese)温佳,赵军锁,王彩玲,等.基于改进MCA的干涉高光谱图像分解[J].光谱学与光谱分析,2016,36(1):254-258.
[2] ZHANG Qican,SU Xianyu.Research progress of dynamic three-dimensional shape measurement[J].Laser & Optoelectronics Progress,2013,50(1):1-14.(in Chinese)张启灿,苏显渝.动态三维面形测量的研究进展[J].激光与光电子学进展,2013,50(1):1-14.
[3] Ko S J,Lee Y H.Center weighted median filters and their applications to image enhancement[J].IEEE Transaction on Circuits Systems,1991,38(9):984-993.
[4] DENG Xiuqin,XIONG Yong.Weighted median filter algorithm for image processing[J].Computer Technology and Development,2009,19(3):46-48.(in Chinese)邓秀勤,熊勇.用于图像处理的加权中值滤波算法[J].计算机技术与发展,2009,19(3):46-98.
[5] Srinivasan Y,Liu H C,Hallioua M.Automated phase measuring profilometry of 3-D diffuse objects[J].Appl Opt,1984,23(18):3105-3108.
[6] LI Junfeng,YANG Feng,HUANG Jing.A fast algorithm for improved dimensional bilateral filter[J].Journal of Circuits and Systems,2013,18(1):137-143.(in Chinese)李俊峰,杨丰,黄靖.一种改进的增维型双边滤波器的快速算法[J].电路与系统学报,2013,18(1):137-143.
[7] PEI Zhipeng,LI Xin.Research on adaptive image denoising method based on wavelet transform[D].Huainan:Anhui University of Science and Technology,2018.(in Chinese)裴志鹏,李昕.基于小波变换的自适应图像去噪方法的研究[D].淮南:安徽理工大学,2018.
[8] ZHANG Hongwei.Research and implementation of image denoising method based on MATLAB[J].Journal of Daqing Normal University,2016,36(3):1-4.(in Chinese)张宏伟.基于MATLAB的图像去噪方法的研究与实现[J].大庆师范学院学报,2016,36(3):1-4.
[9] DONG Jian,DENG Guohui,LI Jinwu.Research on image transformation based on two-dimensional Fourier Transform[J].Fujian Computer,2015,31(9):102-103.(in Chinese)董健,邓国辉,李金武.基于二维傅里叶变换实现图像变换的研究[J].福建电脑,2015,31(9):102-103.
[10] Goldberg RR.Fourier transforms[J].Physics Today,1961,14(10):52-52.
[11] CHEN Haiyun,CHEN Cheng,PENG Baojin,et al.Fourier analysis of the fringe signal of a fiber grating Mach-Zehnder interferometer[J].Acta Photonica Sinica,2016,45(9):173-178.(in Chinese)陈海云,陈成,彭保进,等.光纤光栅马赫-曾德干涉仪条纹信号的傅里叶分析[J].光子学报,2016,45(9):173-178.
[12] An Ping.Processing and analysis of regular interference fringe images[D].Hohhot:Inner Mongolia University of Technology,2017.(in Chinese)安平.规则干涉条纹图像的处理与分析[D].呼和浩特:内蒙古工业大学,2017.
[13] Auslander L,Grunbaum F A.The Fourier transform and the discrete Fourier transform[J].Inverse Problems,1989,5(2):149.
[14] REN Xuhu,SONG Shanshan,QI Yaoguang.Application research of fourier transform profilometry based on wavelet analysis[J].Journal of China University of Petroleum:Natural Science,2007,(1):148-153.(in Chinese)任旭虎,宋珊珊,綦耀光.基于小波分析的傅里叶变换轮廓术应用研究[J].中国石油大学学报:自然科学版,2007,(1):148-153.
[15] Bacchelli S,Papi S.Filtered wavelet thresholding methods[J].Journal of Computational and Applied Mathematics,2004,164-165(1):39-52.
[16] Roberge D,Sheng Y.Optical wavelet matched filter[J].Applied Optics,1994,33(23):5287-93.
[17] YANG Haimei.Extraction and analysis of phase information for high-speed flow field laser interferometry[D].Yantai:Yantai University,2017.(in Chinese)杨海梅.高速流场激光干涉检测相位信息的提取与分析[D].烟台:烟台大学,2017.
[18] WEN Jia,ZHAO Junsuo,WANG Cailing,et al.Interference hyperspectral image decomposition based on improved MCA[J].Spectroscopy and Spectral Analysis,2016,36(1):254-258.(in Chinese)温佳,赵军锁,王彩玲,等.基于改进MCA的干涉高光谱图像分解[J].光谱学与光谱分析,2016,36(1):254-258.