无波前传感自适应波前校正系统的图形处理器加速
|
摘要
基于随机并行梯度下降(SPGD)算法的无波前传感自适应波前校正系统的收敛速度较慢,很难满足无线光相干通信系统对实时性的要求。介绍了SPGD算法的并行化处理,利用图形处理器(GPU)并行计算提高校正系统的收敛速度。选用CCD相机采集的实时光斑形心的周围400个像素平均灰度值作为系统性能指标;利用GPU多线程运算,对性能指标求解过程和变形镜控制电压向量更新过程进行加速处理。室内实验和外场相干光实验结果表明,斯特列尔比达到了0.8以上,时间加速比最大达到了8.6,GPU加速的波前校正系统在提升收敛速度的同时保证了校正效果。
As for the wave-front sensorless adaptive wave-front correction system based on the stochastic parallel gradient descent(SPGD)algorithm,its convergence speed is too slow to satisfy the real-time requirement of a wireless optical coherent communication system.The parallel processing base on the SPGD algorithm is introduced and the graphics processing unit(GPU)parallel computing is used to improve the convergence speed of the correction system.The average gray value of the surrounding 400 pixels centered on the centroid of the real-time spot detected by CCD camera is employed as the value of system performance index.GPU multithreading operation is used to accelerate the solving process of the performance index and the updating process of the deformable mirror control voltage vector.The results from the indoor experiments and the external coherent light experiments show that the Strehl ratio is larger than 0.8 and the maximum time acceleration ratio is up to 8.6.Moreover,the convergence speed of the GPU accelerated wave-front correction system is improved and simultaneously the correction effect is ensured.
As for the wave-front sensorless adaptive wave-front correction system based on the stochastic parallel gradient descent(SPGD)algorithm,its convergence speed is too slow to satisfy the real-time requirement of a wireless optical coherent communication system.The parallel processing base on the SPGD algorithm is introduced and the graphics processing unit(GPU)parallel computing is used to improve the convergence speed of the correction system.The average gray value of the surrounding 400 pixels centered on the centroid of the real-time spot detected by CCD camera is employed as the value of system performance index.GPU multithreading operation is used to accelerate the solving process of the performance index and the updating process of the deformable mirror control voltage vector.The results from the indoor experiments and the external coherent light experiments show that the Strehl ratio is larger than 0.8 and the maximum time acceleration ratio is up to 8.6.Moreover,the convergence speed of the GPU accelerated wave-front correction system is improved and simultaneously the correction effect is ensured.
引文
[1]Wu H L,Li X Y,Yan H X,et al.Study of channel characteristics and its phase compensation technology for atmospheric optical communication[J].Optical Communication Technology,2008,32(7):58-61.吴晗玲,李新阳,严海星,等.大气光通信信道特性及相位补偿技术研究[J].光通信技术,2008,32(7):58-61.
[2]Ellerbroek B L,Rhoadarmer T A.Adaptive wavefront control algorithms for closed loop adaptive optics[J].Mathematical and Computer Modelling,2001,33(1/2/3):145-158.
[3]Booth M J.Wavefront sensorless adaptive optics for large aberrations[J].Optics Letters,2007,32(1):5-7.
[4]Yang H Z,Cai D M,Chen B,et al.Analysis of adaptive optics techniques without a wave-front sensor and its application in atmospheric laser communications[J].Chinese Journal of Lasers,2008,35(5):680-684.杨慧珍,蔡冬梅,陈波,等.无波前传感自适应光学技术及其在大气光通信中的应用[J].中国激光,2008,35(5):680-684.
[5]Ding X W,Chen J X,Wang M.Optimizing parallel join of column-stores on heterogeneous computing platform[J].Computer Engineering and Applications,2017,53(5):73-80.丁祥武,陈金鑫,王梅.异构计算平台上列存储系统的并行连接优化策略[J].计算机工程与应用,2017,53(5):73-80.
[6]Xiao F Z,Mccreath E,Webers C.Fast on-line statistical learning on a GPGPU[C]//Australasian Symposium on Parallel&Distributed,2011:35-42.
[7]Reguera-Salgado J,Calvino-Cancela M,MartinHerrero J.GPU geocorrection for airborne pushbroom imagers[J].IEEE Transactions on Geoscience and Remote Sensing,2012,50(11):4409-4419.
[8]Wu Z H,Tang Q Y,Wang Z J,et al.Numerical simulation of adaptive optical imaging on GPUs[J].Journal of Atmospheric and Environmental Optics,2014,9(3):237-243.吴振华,唐秋艳,王中杰,等.自适应光学数值仿真成像在GPU上的实现[J].大气与环境光学学报,2014,9(3):237-243.
[9]Quan J C,Wang P,Wang H W.Orthorectification of optical aerial images by GPU acceleration[J].Optics and Precision Engineering,2016,24(11):2863-2871.全吉成,王平,王宏伟.计算机图形处理器加速的光学航空影像正射校正[J].光学精密工程,2016,24(11):2863-2871.
[10]Oyama Y,Nomura A,Sato I,et al.Predicting statistics of asynchronous SGD parameters for a large-scale distributed deep learning system on GPUsupercomputers[C]//IEEE International Conference on Big Data,2016:66-75.
[11]Li X K,He Y T.Performance of stochastic parallel gradient descent algorithm in coherent combination[J].Acta Optica Sinica,2016,36(12):1201001.李兴珂,何云涛.相干合成中的随机并行梯度下降算法性能研究[J].光学学报,2016,36(12):1201001.
[12]Vorontsov M A,Carhart G W,Cohen M,et al.Adaptive optics based on analog parallel stochastic optimization:analysis and experimental demonstration[J].Journal of the Optical Society of America A,2000,17(8):1440-1453.
[13]Han L Q,Wang Z B.Fiber coupling efficiency and Strehl ratio for space optical communication based on adaptive optics correction[J].Infrared and Laser Engineering,2013,42(1):125-129.韩立强,王志斌.自适应光学校正下空间光通信的光纤耦合效率及斯特列尔比[J].红外与激光工程,2013,42(1):125-129.
[14]Du Z J,Gao T X,Tang X Y.The usage of graphics processing unit in real-time imaging of optical coherence tomography[J].Acta Laser Biology Sinica,2017,26(2):97-103.杜竹君,高天欣,唐晓英.图形处理器在实时光学相干断层成像中的应用[J].激光生物学报,2017,26(2):97-103.
[15]Guan H C,Li C R,Zhou M,et al.Research on fast fusion of 3Dimaging payload based on CUDA[J].Remote Sensing Technology and Application,2017,32(3):443-448.关宏灿,李传荣,周梅,等.基于CUDA的三维成像载荷快速融合处理方法研究[J].遥感技术与应用,2017,32(3):443-448.
[16]Song Q K,Ma L,Cao J K,et al.Image denoising based on wavelet transform and mean filtering[J].Journal of Natural Science of Heilongjiang University,2016,33(4):555-560.宋清昆,马丽,曹建坤,等.基于小波变换和均值滤波的图像去噪[J].黑龙江大学自然科学学报,2016,33(4):555-560.
[17]Ke X Z,Lei S C,Yang P S.Beam coaxial alignment detection in atmospheric laser communication[J].Chinese Journal of Lasers,2016,43(6):0606003.柯熙政,雷思琛,杨沛松.大气激光通信光束同轴对准检测方法[J].中国激光,2016,43(6):0606003.
[18]Han C,Bai B X,Yang H M,et al.Study fourquadrant detector in the free space laser communication[J].Chinese Journal of Lasers,2009,36(8):2030-2034.韩成,白宝兴,杨华民,等.自由空间激光通信四象限探测器性能研究[J].中国激光,2009,36(8):2030-2034.
[19]van den Bos A.Aberration and the Strehl ratio[J].Journal of the Optical Society of America A,2000,17(2):356-358.
[2]Ellerbroek B L,Rhoadarmer T A.Adaptive wavefront control algorithms for closed loop adaptive optics[J].Mathematical and Computer Modelling,2001,33(1/2/3):145-158.
[3]Booth M J.Wavefront sensorless adaptive optics for large aberrations[J].Optics Letters,2007,32(1):5-7.
[4]Yang H Z,Cai D M,Chen B,et al.Analysis of adaptive optics techniques without a wave-front sensor and its application in atmospheric laser communications[J].Chinese Journal of Lasers,2008,35(5):680-684.杨慧珍,蔡冬梅,陈波,等.无波前传感自适应光学技术及其在大气光通信中的应用[J].中国激光,2008,35(5):680-684.
[5]Ding X W,Chen J X,Wang M.Optimizing parallel join of column-stores on heterogeneous computing platform[J].Computer Engineering and Applications,2017,53(5):73-80.丁祥武,陈金鑫,王梅.异构计算平台上列存储系统的并行连接优化策略[J].计算机工程与应用,2017,53(5):73-80.
[6]Xiao F Z,Mccreath E,Webers C.Fast on-line statistical learning on a GPGPU[C]//Australasian Symposium on Parallel&Distributed,2011:35-42.
[7]Reguera-Salgado J,Calvino-Cancela M,MartinHerrero J.GPU geocorrection for airborne pushbroom imagers[J].IEEE Transactions on Geoscience and Remote Sensing,2012,50(11):4409-4419.
[8]Wu Z H,Tang Q Y,Wang Z J,et al.Numerical simulation of adaptive optical imaging on GPUs[J].Journal of Atmospheric and Environmental Optics,2014,9(3):237-243.吴振华,唐秋艳,王中杰,等.自适应光学数值仿真成像在GPU上的实现[J].大气与环境光学学报,2014,9(3):237-243.
[9]Quan J C,Wang P,Wang H W.Orthorectification of optical aerial images by GPU acceleration[J].Optics and Precision Engineering,2016,24(11):2863-2871.全吉成,王平,王宏伟.计算机图形处理器加速的光学航空影像正射校正[J].光学精密工程,2016,24(11):2863-2871.
[10]Oyama Y,Nomura A,Sato I,et al.Predicting statistics of asynchronous SGD parameters for a large-scale distributed deep learning system on GPUsupercomputers[C]//IEEE International Conference on Big Data,2016:66-75.
[11]Li X K,He Y T.Performance of stochastic parallel gradient descent algorithm in coherent combination[J].Acta Optica Sinica,2016,36(12):1201001.李兴珂,何云涛.相干合成中的随机并行梯度下降算法性能研究[J].光学学报,2016,36(12):1201001.
[12]Vorontsov M A,Carhart G W,Cohen M,et al.Adaptive optics based on analog parallel stochastic optimization:analysis and experimental demonstration[J].Journal of the Optical Society of America A,2000,17(8):1440-1453.
[13]Han L Q,Wang Z B.Fiber coupling efficiency and Strehl ratio for space optical communication based on adaptive optics correction[J].Infrared and Laser Engineering,2013,42(1):125-129.韩立强,王志斌.自适应光学校正下空间光通信的光纤耦合效率及斯特列尔比[J].红外与激光工程,2013,42(1):125-129.
[14]Du Z J,Gao T X,Tang X Y.The usage of graphics processing unit in real-time imaging of optical coherence tomography[J].Acta Laser Biology Sinica,2017,26(2):97-103.杜竹君,高天欣,唐晓英.图形处理器在实时光学相干断层成像中的应用[J].激光生物学报,2017,26(2):97-103.
[15]Guan H C,Li C R,Zhou M,et al.Research on fast fusion of 3Dimaging payload based on CUDA[J].Remote Sensing Technology and Application,2017,32(3):443-448.关宏灿,李传荣,周梅,等.基于CUDA的三维成像载荷快速融合处理方法研究[J].遥感技术与应用,2017,32(3):443-448.
[16]Song Q K,Ma L,Cao J K,et al.Image denoising based on wavelet transform and mean filtering[J].Journal of Natural Science of Heilongjiang University,2016,33(4):555-560.宋清昆,马丽,曹建坤,等.基于小波变换和均值滤波的图像去噪[J].黑龙江大学自然科学学报,2016,33(4):555-560.
[17]Ke X Z,Lei S C,Yang P S.Beam coaxial alignment detection in atmospheric laser communication[J].Chinese Journal of Lasers,2016,43(6):0606003.柯熙政,雷思琛,杨沛松.大气激光通信光束同轴对准检测方法[J].中国激光,2016,43(6):0606003.
[18]Han C,Bai B X,Yang H M,et al.Study fourquadrant detector in the free space laser communication[J].Chinese Journal of Lasers,2009,36(8):2030-2034.韩成,白宝兴,杨华民,等.自由空间激光通信四象限探测器性能研究[J].中国激光,2009,36(8):2030-2034.
[19]van den Bos A.Aberration and the Strehl ratio[J].Journal of the Optical Society of America A,2000,17(2):356-358.