基于图像块细粒度的自适应单像素成像算法
|
摘要
设计了一种基于图像块细粒度的自适应单像素成像算法。首先获取目标的低分辨图像,将其分为四等块,根据每块中重要小波系数的个数与分布,将其分为非重要、重要和半重要三类。非重要块不采样;重要块全采样;半重要块进行迭代分块、分类,部分采样。利用小波反变换重建高分辨图并重复上述步骤,直到重建图像达到目标分辨率。实验结果表明,对于背景平滑的医学生物图像,本文方法能够在相同采样率下至少提高重建图像PSNR约2 dB。
Images were divided into four equal-sized blocks by adaptive compressive imaging.The four blocks were classified into two types after getting the number of important discrete wavelet transform(DWT) coefficients,which are the important blocks and the unimportant blocks.The important blocks were all sampled with higher resolution,leading to high sampling rate.To solve these problems,an imaging method is designed based on fine precision classification of image blocks.Coarse image is reconstruted and divided into four equal-size blocks first.Then the four image blocks are classified into three types according to the number and distribution of their important DWT coefficients,which are the unimportant blocks,the important blocks and the partly important blocks.The unimportant blocks are ignored.The important blocks are all sampled with higher resolution.The partly important blocks are divided repeatedly,and only the important parts are sampled.Image with higher resolution is reconstructed through inverse wavelet transform.Steps above are repeated until the reconstructed image gets required resolution.Experimental results show that compared with other three algorithms,the peak signal-to-noise ratio(PSNR) of reconstructed images is improved by at least 2 dB with similar sampling rate for biology images.
Images were divided into four equal-sized blocks by adaptive compressive imaging.The four blocks were classified into two types after getting the number of important discrete wavelet transform(DWT) coefficients,which are the important blocks and the unimportant blocks.The important blocks were all sampled with higher resolution,leading to high sampling rate.To solve these problems,an imaging method is designed based on fine precision classification of image blocks.Coarse image is reconstruted and divided into four equal-size blocks first.Then the four image blocks are classified into three types according to the number and distribution of their important DWT coefficients,which are the unimportant blocks,the important blocks and the partly important blocks.The unimportant blocks are ignored.The important blocks are all sampled with higher resolution.The partly important blocks are divided repeatedly,and only the important parts are sampled.Image with higher resolution is reconstructed through inverse wavelet transform.Steps above are repeated until the reconstructed image gets required resolution.Experimental results show that compared with other three algorithms,the peak signal-to-noise ratio(PSNR) of reconstructed images is improved by at least 2 dB with similar sampling rate for biology images.
引文
[1] Duarte M F,Davenport M A,Takbar D,et al.Single-pixel imaging via compressive sampling[J].IEEE Signal Processing Magazine,2008,25(2):83-91.
[2] Baraniuk R G.Compressive sensing[J].IEEE Signal Processing Magazine,2007,24(4):118-121.
[3] Candès E J,Wakin M B.An introduction to compressive sampling[J].IEEE Signal Processing Magazine,2008,25(2):21-30.
[4] Candès E J.The restricted isometry property and its implications for compressed sensing[J].Comptes Rendus Mathematique,2008,346(9-10):589-592.
[5] Tropp J A,Gilbert A C.Signal recovery from random measurements via orthogonal matching pursuit[J].IEEE Transactions on Information Theory,2007,53(12):4655-4666.
[6] Barranca V J,Kova,et al.Improved compressive sensing of natural scenes using localized random sampling[J].Scientific Reports,2016,6:31976.
[7] Howland G A,Lum D J,Ware M R,et al.Photon counting compressive depth mapping[J].Optics Express,2013,21(20):23822-23837.
[8] Tajahuerce E,Durán V,Clemente P,et al.Image transmi-ssion through dynamic scattering media by single-pixel photodetection[J].Optics Express,2014,22(14):16945-16955.
[9] Durán V,Soldevila F,Irles E,et al.Compressive imaging in scattering media[J].Optics Express,2015,23(11):14424-14433.
[10] Cheng J.Ghost imaging through turbulent atmosphere[J].Optics Express,2009,17(10):7916-7921.
[11] Lv Pei,Zhou Ren-kui,HE Jun-hua,et al.Research on underwater single-pixel imaging system[J].Journal of Optoelectronics·Laser,2011,22(9):1425-1430.吕沛,周仁魁,何俊华,等.水下单像素成像系统研究[J].光电子·激光,2011,22(9):1425-1430.
[12] KONG Xiang-hai, ZHANG Yuan, LIANG Yan-mei. Rese-arch on a new method of medical image sampling based on compressed sensing[J].Journal of Optoelectronics·Laser,2014,25(8):1635-1640.孔祥海,张媛,梁艳梅.基于压缩感知的医学图像采样新方法研究[J].光电子·激光,2014,25(8):1635-1640.
[13] Liang Zhen-yu,Fan Xiang,Cheng Zheng-dong,et al.Rese-arch of high-order thermal ghost imaging for an axial moving target[J].Journal of Optoelectronics·Laser,2017,28(5):547-552.梁振宇,樊祥,程正东,等.轴向运动目标的热光高阶鬼成像研究[J].光电子·激光,2017,28(5):547-552.
[14] Radwell N,Mitchell K J,Gibson G M,et al.Single-pixel infrared and visible microscope[J].Optica,2014,1(5):285-289.
[15] Sun M J, Edgar M P, Gibson G M, et al. Single-pixel three-dimensional imaging with time-based depth resolution[J].Nature Communications,2016,7:12010.
[16] Sun B,Edgar M P,Bowman R,et al.3D computational imaging with single-pixel detectors[J].Science,2013,340(6134):844-847.
[17] Kirmani A,Cola?o A,Wong F N C,et al.Exploiting sparsity in time-of-flight range acquisition using a single time-resolved sensor[J].Optics Express,2011,19(22):21485-21507.
[18] Cola?o A,Kirmani A,Howland G A,et al.Compressive depth map acquisition using a single photon-counting detector:parametric signal processing meets sparsity[C].IEEE Conference on Computer Vision and Pattern Recognition,2012,96-102.
[19] Sun S,Liu W T,Lin H Z,et al.Multi-scale adaptive computational ghost imaging[J].Scientific Reports,2016,6:37013.
[20] Peng Y,Liu Y,Ren W,et al.Guided compressive sensing single-pixel imaging technique based on hierarchical model[J].Journal of Modern Optics,2016,63(7):677-684.
[21] Yu W K,Li M F,Yao X R,et al.Adaptive compressive ghost imaging based on wavelet trees and sparse representation[J].Optics Express,2014,22(6):7133-7144.
[22] Soldevila F, Salvador-Balaguer E, Clemente P, et al.High-resolution adaptive imaging with a single photodiode[J].Scientific Reports,2015,5:14300.
[23] CVG university of granada,computer vision group.M-iscelaneous gray level images[OL].(2014-03-13).[2018-02-01].http://decsai.ugr.es/cvg/dbimagenes.
[2] Baraniuk R G.Compressive sensing[J].IEEE Signal Processing Magazine,2007,24(4):118-121.
[3] Candès E J,Wakin M B.An introduction to compressive sampling[J].IEEE Signal Processing Magazine,2008,25(2):21-30.
[4] Candès E J.The restricted isometry property and its implications for compressed sensing[J].Comptes Rendus Mathematique,2008,346(9-10):589-592.
[5] Tropp J A,Gilbert A C.Signal recovery from random measurements via orthogonal matching pursuit[J].IEEE Transactions on Information Theory,2007,53(12):4655-4666.
[6] Barranca V J,Kova,et al.Improved compressive sensing of natural scenes using localized random sampling[J].Scientific Reports,2016,6:31976.
[7] Howland G A,Lum D J,Ware M R,et al.Photon counting compressive depth mapping[J].Optics Express,2013,21(20):23822-23837.
[8] Tajahuerce E,Durán V,Clemente P,et al.Image transmi-ssion through dynamic scattering media by single-pixel photodetection[J].Optics Express,2014,22(14):16945-16955.
[9] Durán V,Soldevila F,Irles E,et al.Compressive imaging in scattering media[J].Optics Express,2015,23(11):14424-14433.
[10] Cheng J.Ghost imaging through turbulent atmosphere[J].Optics Express,2009,17(10):7916-7921.
[11] Lv Pei,Zhou Ren-kui,HE Jun-hua,et al.Research on underwater single-pixel imaging system[J].Journal of Optoelectronics·Laser,2011,22(9):1425-1430.吕沛,周仁魁,何俊华,等.水下单像素成像系统研究[J].光电子·激光,2011,22(9):1425-1430.
[12] KONG Xiang-hai, ZHANG Yuan, LIANG Yan-mei. Rese-arch on a new method of medical image sampling based on compressed sensing[J].Journal of Optoelectronics·Laser,2014,25(8):1635-1640.孔祥海,张媛,梁艳梅.基于压缩感知的医学图像采样新方法研究[J].光电子·激光,2014,25(8):1635-1640.
[13] Liang Zhen-yu,Fan Xiang,Cheng Zheng-dong,et al.Rese-arch of high-order thermal ghost imaging for an axial moving target[J].Journal of Optoelectronics·Laser,2017,28(5):547-552.梁振宇,樊祥,程正东,等.轴向运动目标的热光高阶鬼成像研究[J].光电子·激光,2017,28(5):547-552.
[14] Radwell N,Mitchell K J,Gibson G M,et al.Single-pixel infrared and visible microscope[J].Optica,2014,1(5):285-289.
[15] Sun M J, Edgar M P, Gibson G M, et al. Single-pixel three-dimensional imaging with time-based depth resolution[J].Nature Communications,2016,7:12010.
[16] Sun B,Edgar M P,Bowman R,et al.3D computational imaging with single-pixel detectors[J].Science,2013,340(6134):844-847.
[17] Kirmani A,Cola?o A,Wong F N C,et al.Exploiting sparsity in time-of-flight range acquisition using a single time-resolved sensor[J].Optics Express,2011,19(22):21485-21507.
[18] Cola?o A,Kirmani A,Howland G A,et al.Compressive depth map acquisition using a single photon-counting detector:parametric signal processing meets sparsity[C].IEEE Conference on Computer Vision and Pattern Recognition,2012,96-102.
[19] Sun S,Liu W T,Lin H Z,et al.Multi-scale adaptive computational ghost imaging[J].Scientific Reports,2016,6:37013.
[20] Peng Y,Liu Y,Ren W,et al.Guided compressive sensing single-pixel imaging technique based on hierarchical model[J].Journal of Modern Optics,2016,63(7):677-684.
[21] Yu W K,Li M F,Yao X R,et al.Adaptive compressive ghost imaging based on wavelet trees and sparse representation[J].Optics Express,2014,22(6):7133-7144.
[22] Soldevila F, Salvador-Balaguer E, Clemente P, et al.High-resolution adaptive imaging with a single photodiode[J].Scientific Reports,2015,5:14300.
[23] CVG university of granada,computer vision group.M-iscelaneous gray level images[OL].(2014-03-13).[2018-02-01].http://decsai.ugr.es/cvg/dbimagenes.