基于联合稀疏变换学习的工件去噪方法研究
|
摘要
针对视觉测量中给光好坏会直接影响工件检测精度和效率的问题,提出了一种基于联合稀疏变换学习对不同光源下工件图像去噪的方法。该方法首先从噪声图像中提取噪声图像块,通过稀疏编码和稀疏变换更新交替运算对噪声图像块进行内部聚类变换学习;然后计算每个噪声图像块中聚类信号的稀疏水平,并选择最小稀疏水平作为该去噪块的稀疏水平;最后对去噪块进行聚类并用最后一次迭代去噪块的均值估计去噪图像。实验结果表明,提出的方法对不同光源下工件图像的去噪效果和计算速度均优于其他算法,具有较好的去噪性能。
As the quality of light in visual measurement will directly affect the accuracy and efficiency of workpiece detection, to address the problem, an approach is proposed to denoise the workpiece image under different light sources with union-of-transforms learning. The proposed method firstly extracts the noise image patches from noise images, and studies the intra-cluster transform learning of the noise patches by alternated between sparse coding and transform update steps. Next, the transform sparse levels of clustering signal in each noise patch are calculated, and the minimum sparse level is selected as the sparse level of the denoising patches. Finally, denoised patches are clustered and the denoised image is estimated by the mean value of the final iterative denoised patches. The experimental results show that the proposed mothed is superior to the other algorithms in terms of both denoising effect and algorithm speed for workpiece images under different light sources, and gains better denoise performance.
As the quality of light in visual measurement will directly affect the accuracy and efficiency of workpiece detection, to address the problem, an approach is proposed to denoise the workpiece image under different light sources with union-of-transforms learning. The proposed method firstly extracts the noise image patches from noise images, and studies the intra-cluster transform learning of the noise patches by alternated between sparse coding and transform update steps. Next, the transform sparse levels of clustering signal in each noise patch are calculated, and the minimum sparse level is selected as the sparse level of the denoising patches. Finally, denoised patches are clustered and the denoised image is estimated by the mean value of the final iterative denoised patches. The experimental results show that the proposed mothed is superior to the other algorithms in terms of both denoising effect and algorithm speed for workpiece images under different light sources, and gains better denoise performance.
引文
[1]祝严刚,张桂梅.一种改进的非局部均值图像去噪去噪算法[J].计算机工程与应用,2017,53(18):192-198.
[2]杨岳飞,刘辉,谭检平.带噪语音信号小波去噪算法研究[J].计算机工程与应用,2015,51(14):211-214.
[3]郭德全,杨红雨,刘东权,等.基于稀疏性的图像去噪综述[J].计算机应用研究,2012,29(2):406-413.
[4]华志胜,付丽华.基于块分类和字典优化的K-SVD图像去噪研究[J].计算机工程与应用,2017,53(16):187-192.
[5]Mairal J,Elad M,Sapiro G.Sparse representation for color image restoration[J].IEEE Transactions on Image Processing,2008,17(1):53-69.
[6]Tropp J A,Gilbert A C.Signal recovery from random measurement via orthogonal matching pursuit[J].IEEETransactions on Information Theory,2007,53(12):4655-4666.
[7]Kunis S,Rauhut H.Random sampling of sparse trigonometric polynomals,II.Orthogonal matching pursuit versus basis pursuit[J].Foundations of Computational Mathematics,2008,8(6):737-763.
[8]邵婕,孙成禹,唐杰,等.基于字典训练的小波域稀疏表示微地震去噪方法[J].石油地球物理勘探,2016,51(2):254-260.
[9]Ravishankar S,Bresler Y.Learning sparsifying transforms for image processing[C]//Proceedings of IEEE International Conference on Image Processing,2013:681-684.
[10]Ravishankar S,Bresler Y.Learning overcomplete sparsifying transforms for signal processing[C]//Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing,2013:3088-3092.
[11]Pfister L.Tomgraphic reconstruction with adaptive sparsifying transforms[D].Urbana:University of Illinois at Urbana-Chanpanign,2013:17-27.
[12]王念兵,吴秦,许洁.局部联合结构化稀疏表示的单样本人脸识别[J].计算机工程与应用,2018,54(1):204-228.
[13]Shekhar S,Patel V M,Nasrabadi N M,et al.Joint sparse representation for robust multimodal biometrics recognition[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2014,36(1):113-126.
[14]高美凤,王晨.基于联合滤波的聚类稀疏表示图像去噪算法[J].计算机工程与应用,2015,51(24):180-185.
[15]Borhani R,Watt J,Katsaggelos A.Global hard thresholding algorithms for joint sparse image representation and denoising[J].arXiv:1705.09816,2017.
[16]Ning X U,Xiao X,You H,et al.A pansharpening method based on HCT and joint sparse model[J].Acta Geodaetica Et Cartographica Sinica,2016,45(4):434-441.
[17]Wen B,Ravishankar S,Bresler Y.Learning overcomplete sparsifying transforms with block cosparsity[C]//Proceedings of IEEE International Conference on Image Processing,2015:803-807.
[18]Wen B,Ravishankar S,Bresler Y.Structured overcomplete sparsifying transformlearning with convergence guarantees and applications[J].International Journal of Computer Vision,2015,114(2/3):137-167.
[19]Bruckstein A M,Donoho D L,Elad M.From sparse solutions of systems of equations to sparse modeling of signals and images[J].SIAM Review,2009,51(1):34-81.
[20]Elad M,Milanfar P,Rubinstein U R.Analysis versus synthesis in signal priors[J].Inverse Problems,2007,23(3):947-968.
[21]Ravishankar S,Bresler Y.Learning sparsifying transforms[J].IEEE Transactions on Signal Processing,2013,61(5):1072-1086.
[22]Ravishankar S,Bresler Y.Learning doubly sparse transforms for images[J].IEEE Transactions on Image Processing,2013,22(12):4598-4612.
[23]Ravishankar S,Bresler Y.Closed-form solutions within sparsifying transform learning[C]//Proceedings of IEEEInternational Conference on Acoustics,Speech and Signal Processing,2013:5378-5382.
[24]Eiad M,Aharon M.Image denoising via sparse and redundant representations over learned dictionaries[J].IEEE Transactions on Image Processing,2006,15(12):3736-3745.
[25]Aharon M,Elad M,Bruckstein A.K-SVD:An algorithm for designing overcomplete dictionaries for sparse representation[J].IEEE Transactions on Signal Processing,2006,54(1):4311-4322.
[2]杨岳飞,刘辉,谭检平.带噪语音信号小波去噪算法研究[J].计算机工程与应用,2015,51(14):211-214.
[3]郭德全,杨红雨,刘东权,等.基于稀疏性的图像去噪综述[J].计算机应用研究,2012,29(2):406-413.
[4]华志胜,付丽华.基于块分类和字典优化的K-SVD图像去噪研究[J].计算机工程与应用,2017,53(16):187-192.
[5]Mairal J,Elad M,Sapiro G.Sparse representation for color image restoration[J].IEEE Transactions on Image Processing,2008,17(1):53-69.
[6]Tropp J A,Gilbert A C.Signal recovery from random measurement via orthogonal matching pursuit[J].IEEETransactions on Information Theory,2007,53(12):4655-4666.
[7]Kunis S,Rauhut H.Random sampling of sparse trigonometric polynomals,II.Orthogonal matching pursuit versus basis pursuit[J].Foundations of Computational Mathematics,2008,8(6):737-763.
[8]邵婕,孙成禹,唐杰,等.基于字典训练的小波域稀疏表示微地震去噪方法[J].石油地球物理勘探,2016,51(2):254-260.
[9]Ravishankar S,Bresler Y.Learning sparsifying transforms for image processing[C]//Proceedings of IEEE International Conference on Image Processing,2013:681-684.
[10]Ravishankar S,Bresler Y.Learning overcomplete sparsifying transforms for signal processing[C]//Proceedings of IEEE International Conference on Acoustics,Speech and Signal Processing,2013:3088-3092.
[11]Pfister L.Tomgraphic reconstruction with adaptive sparsifying transforms[D].Urbana:University of Illinois at Urbana-Chanpanign,2013:17-27.
[12]王念兵,吴秦,许洁.局部联合结构化稀疏表示的单样本人脸识别[J].计算机工程与应用,2018,54(1):204-228.
[13]Shekhar S,Patel V M,Nasrabadi N M,et al.Joint sparse representation for robust multimodal biometrics recognition[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2014,36(1):113-126.
[14]高美凤,王晨.基于联合滤波的聚类稀疏表示图像去噪算法[J].计算机工程与应用,2015,51(24):180-185.
[15]Borhani R,Watt J,Katsaggelos A.Global hard thresholding algorithms for joint sparse image representation and denoising[J].arXiv:1705.09816,2017.
[16]Ning X U,Xiao X,You H,et al.A pansharpening method based on HCT and joint sparse model[J].Acta Geodaetica Et Cartographica Sinica,2016,45(4):434-441.
[17]Wen B,Ravishankar S,Bresler Y.Learning overcomplete sparsifying transforms with block cosparsity[C]//Proceedings of IEEE International Conference on Image Processing,2015:803-807.
[18]Wen B,Ravishankar S,Bresler Y.Structured overcomplete sparsifying transformlearning with convergence guarantees and applications[J].International Journal of Computer Vision,2015,114(2/3):137-167.
[19]Bruckstein A M,Donoho D L,Elad M.From sparse solutions of systems of equations to sparse modeling of signals and images[J].SIAM Review,2009,51(1):34-81.
[20]Elad M,Milanfar P,Rubinstein U R.Analysis versus synthesis in signal priors[J].Inverse Problems,2007,23(3):947-968.
[21]Ravishankar S,Bresler Y.Learning sparsifying transforms[J].IEEE Transactions on Signal Processing,2013,61(5):1072-1086.
[22]Ravishankar S,Bresler Y.Learning doubly sparse transforms for images[J].IEEE Transactions on Image Processing,2013,22(12):4598-4612.
[23]Ravishankar S,Bresler Y.Closed-form solutions within sparsifying transform learning[C]//Proceedings of IEEEInternational Conference on Acoustics,Speech and Signal Processing,2013:5378-5382.
[24]Eiad M,Aharon M.Image denoising via sparse and redundant representations over learned dictionaries[J].IEEE Transactions on Image Processing,2006,15(12):3736-3745.
[25]Aharon M,Elad M,Bruckstein A.K-SVD:An algorithm for designing overcomplete dictionaries for sparse representation[J].IEEE Transactions on Signal Processing,2006,54(1):4311-4322.