Spatial Resolution of Compressed Ultrafast Photography System
Ultrafast optical imaging technology can acquire image signals of ultrafast scenes in the picosecond or even femtosecond scale.As one of the major devices for ultrafast imaging,streak camera with picosecond-scale time resolution and one-dimensional spatial resolution is commonly used for ultrafast diagnosis of Inertial Confinement Fusion(ICF).Combining streak camera with compressed sensing theory,compressed ultrafast photography has been proposed,which is capable of two-dimensional imaging of ultrafast dynamic scenes at 100 billion frames per second.It can achieve up to hundreds of frames per acquisition without specialized modulated illumination,which has a performance far beyond traditional imaging technologies.The main components of compressed ultrafast photography system include streak camera,Digital Micromirror Device(DMD),and optical path system.The DMD is employed to encode the two-dimensional image signal,whereafter the encoded signal passes through the photocathode,scanning electric field and fluorescent screen of the streak camera successively,and finally arrives at the CCD.Deflected by the scanning electric field in the perpendicular direction,the encoded images of different moments shift and overlap on the fluorescent screen,forming the sampled signal captured by the CCD.Utilizing the reconstruction algorithm,frames of two-dimensional image signals can be reconstructed from the sampled signal.In this paper,a Generalized Alternating Projection(GAP)algorithm based on the PnP framework is used to reconstruct the original images.Based on the image prior of gradient sparsity,the Total Variation(TV)denoise operator is combined with the GAP algorithm for image denoising and meanwhile remaining image detail.The algorithm decomposes the solving operation of the reconstructed images into two sub-processes of alternating projection,and continuously approaches the solution that satisfies all the constraints in the alternating iteration operation.In order to analyze the spatial resolution and imaging quality of the compressed ultrafast photography system,this paper carries out imaging simulation experiments based on MATLAB software and measures the spatial resolution of the system.A mathematical model of image signal acquisition is established according to the working principle of compressed ultrafast photography,with which the sampled signals are obtained.The GAP-TV reconstruction algorithm is employed to iteratively solve and reconstruct the original images from the sampled signal.The simulation results show that the reconstructed images can achieve a dynamic spatial resolution of 10 lp/mm under the frame number of 8.Due to the temporal compressed sampling,the dynamic spatial resolution decreases by 5 lp/mm compared to static imaging.The factors affecting the imaging quality and spatial resolution are researched and analyzed,and the imaging simulation is performed using different DMD coding sampling rates.Higher image quality can be obtained using coding sampling rates between 30%and 50%,and highest spatial resolution of 10 lp/mm can be reached with the coding sampling rates from 40%to 50%when simulating imaging with 8 frames of images.Both large frame number and high coding sampling rate can cause signal aliasing,and thereby lead to the decrease of image quality.Therefore,in order to further analyze the relationship between coding sampling rate and imaging quality,imaging quality simulations were performed with data sets of 16,24 and 32 frames,respectively.It is demonstrated that when the system images dynamic scenes of more frames,better imaging quality can be obtained with lower coding sampling rates due to the reduction of signal aliasing.In order to acquire the optimal coding sampling rate for the highest imaging quality,sets of image data are used for simulation,and the optimal coding sampling rate fitted curve is plotted based on the simulation results.Comparative tests verify that the coding sampling rates from the fitted curve result in better imaging quality,which helps the system to achieve the best spatial resolution.
Ultrafast diagnosisCompressed ultrafast photographyCompressed sensingSpatial resolutionStreak camera
NETL
NSTL
万方数据
