Journal of Infrared and Millimeter Waves, Volume. 42, Issue 1, 61(2023)

Multiblock compressed sensing imaging in real time

Hu LI1,3,4, Xue-Feng LIU1,4、*, Xu-Ri YAO2,5、**, Fan LIU1,4, Shen-Cheng DOU1,4, Tai HU3,4, and Guang-Jie ZHAI1,4
Author Affiliations
  • 1Key Laboratory of Electronics and Information Technology for Space System,National Space Science Center,Chinese Academy of Sciences,Beijing 100190,China
  • 2Center for Quantum Information Sciences and Key Laboratory of Advanced Optoelectronic Quantum Architecture and Measurements(MOE),School of Physics,Beijing Institute of Technology,Beijing 100081,China
  • 3Laboratory of Satellite Mission Operation,National Space Science Center,Chinese Academy of Sciences,Beijing 100190,China
  • 4University of Chinese Academy of Sciences,Beijing 100049,China
  • 5Beijing Academy of Quantum Information Sciences,Beijing 100081,China
  • show less
    Figures & Tables(13)
    Block compressed sensing(BCS)architecture with a graphics processing unit(GPU)acceleration imaging system
    Inverse(a,c,e)and CS(b,d,f)reconstruction results with subsampling rate = 0.3,the image sizes are(a),(b)32 × 32 pixels,(c),(d)64 × 64 pixels,and(e),(f)128 × 128 pixels
    Unit of projection vectors derived from a compressive element block. A part of the coding pattern on the SLM is divided into four identical parallelly measuring blocks. One measurement entry,which corresponds to a measurement operation and an observed value,is reshaped into a vector according to the vertical orientation
    Blocki is the observed value vector of the i-th compressed block used to recover the i-th original image block,Mj indicates the j-th measurement corresponding to j-th coding pattern,Blocki and Mj exactly indicate the observed value of i-th compressive block and j-th measurement,here,n=N×NC×C and m≤C×C
    Peak signal-to-noise ratio(PSNR)and reconstruction time with different block sizes and different under-sampling rates
    Block-compressive reconstruction procedure with GPU acceleration
    Comparison of experimental results from different low-resolution images with different compression ratios[12],a-1)–a-9)shows the digital chart,b-1)–b-9)is the film,c-1–c-9)is the toy,a,b,c-1),a,b,c-4)and a,b,c-7)are the low-resolution sampling images with 64 × 64 pixels,high-resolution MBCS reconstruction results with 128 × 128 pixels and the traditional block CS results,respectively,further,a,b,c-2),a,b,c-5)and a,b,c-8)are the low-resolution sampling images with 32 × 32 pixels,high-resolution MBCS reconstruction results with 128 × 128 pixels and the traditional block CS results,respectively,also,a,b,c-3),a,b,c-6)and a,b,c-9)are the low-resolution sampling images with 16 × 16 pixels,high-resolution MBCS reconstruction results with 128 × 128 pixels and the traditional block CS results,respectively
    Reconstruction time for the 128 × 128 scene by the MBCS algorithm using CPU and with GPU acceleration for different block sizes
    Reconstruction time for the 256 × 256 scene by the MBCS algorithm using CPU and with GPU acceleration for different block sizes
    Reconstruction time for the 512 × 512 scene by the MBCS algorithm using CPU and with GPU acceleration for different block sizes
    • Table 1. Comparison of the quality between traditional Block CS and MBCS

      View table
      View in Article

      Table 1. Comparison of the quality between traditional Block CS and MBCS

      Image TypeCompression RatioTraditional Block CSMBCS
      PSNRFSIMPSNRFSIM
      Digital Chart2x212.990.5415.630.85
      4x412.670.7614.740.84
      8x816.450.7916.70.81
      Film2x220.230.9233.890.99
      4x424.850.9329.40.96
      8x818.090.8923.260.91
      Toy2x216.770.6140.310.99
      4x430.440.9437.480.97
      8x834.550.9535.010.97
    • Table 2. Comparison of the reconstruction time between Matlab–CPU and GPU

      View table
      View in Article

      Table 2. Comparison of the reconstruction time between Matlab–CPU and GPU

      Image SizeCPU-MatlabGPUSpeedup
      32×320.39s0.05 s7.79
      64×647.86s0.03 s216
      128×12838.12s0.140 7 s270
    • Table 3. Comparsion of MBCS reconstruction times between the CPU algorithm and GPU acceleration for 128 × 128, 256 × 256, and 512 × 512 scenes. The first column lists the size of high-resolution images, HR stands for high resolution. The second column is the block size used to reconstruction, the third column shows the number of blocks in block reconstruction, the fourth column lists the time to recover one HR image in Matlab, the fifth column lists the time to recover one HR image by the MBCS algorithm with GPU acceleration, the sixth column lists the average time to recover each block of HR image, and it is equal to corresponding value in column “GPU (s)” divided by the corresponding value in column “Blks Cnt”

      View table
      View in Article

      Table 3. Comparsion of MBCS reconstruction times between the CPU algorithm and GPU acceleration for 128 × 128, 256 × 256, and 512 × 512 scenes. The first column lists the size of high-resolution images, HR stands for high resolution. The second column is the block size used to reconstruction, the third column shows the number of blocks in block reconstruction, the fourth column lists the time to recover one HR image in Matlab, the fifth column lists the time to recover one HR image by the MBCS algorithm with GPU acceleration, the sixth column lists the average time to recover each block of HR image, and it is equal to corresponding value in column “GPU (s)” divided by the corresponding value in column “Blks Cnt”

      HR img SizeBlk SizeBlks CntCPU(s)GPU(s)AVG/blk(s/blk)
      128×1282×240965.54261.4210.0638
      4×410241.3847.40670.0462
      8×82560.739.911360.0387
      16×16642.592.42480.0378
      32×32161.920.61560.0384
      64×64421.020.15550.0388
      256×2562×216384553642.820.0392
      4×44096113158.720.0387
      8×8102448.1739.42210.0384
      16×1625613.669.59570.0374
      32×32648.762.39230.0373
      64×641638.250.60470.0377
      128×1284119.920.16170.0404
      512×5122×265536998.982604.890.0397
      4×416384282.13636.3930.0388
      8×84096102.49158.6970.0387
      16×16102455.4238.3460.0374
      32×3225645.139.68510.0378
      64×6464106.642.43760.0380
      128×12816125.160.63570.0397
      256×256471.350.22390.0559
    Tools

    Get Citation

    Copy Citation Text

    Hu LI, Xue-Feng LIU, Xu-Ri YAO, Fan LIU, Shen-Cheng DOU, Tai HU, Guang-Jie ZHAI. Multiblock compressed sensing imaging in real time[J]. Journal of Infrared and Millimeter Waves, 2023, 42(1): 61

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Research Articles

    Received: Oct. 14, 2021

    Accepted: --

    Published Online: Feb. 23, 2023

    The Author Email: LIU Xue-Feng (liuxuefeng@nssc.ac.cn), YAO Xu-Ri (yaoxuri@bit.edu.cn)

    DOI:10.11972/j.issn.1001-9014.2023.01.009

    Topics