Study On Data Reconstruction Algorithm Based On DSP For Single Pixel Detection

According to the acousto-optic correlator structure adjustment in the difficult problem,the author of this paper proposes a linear array image signal imaging system based on compressed sensing,and Based on DSP,this paper mainly studies the acoustic and optical signals collected by the imaging system.In this thesis,five aspects of the study of the acoustic and optical signal reconstruction based on DSP are developed:(1)Selection of system architecture.First of all,the problem of structural adjustment of time integral acousto-optic correlator is discussed,and adopted a new linear array image signal imaging system based on compressed sensing,then the acquisition of sound and light signals of the system are analyzed based on the DSP compressed sensing reconstruction research.(2)Selection of DSP operating system.The study adopted the TMS320C6678 of TI,the DSP supports different system to run,Therefore,the experiment was carried out mainly on the two systems without operating system and Linux operating system,and the experimental results were compared and analyzed.(3)Selection of measurement matrix.Different measurement matrix for signal reconstruction effect is different,have been conducted on several measurement matrix experimental comparison and analysis,choose a measurement matrix that has the generality,the measurement matrix with good effect as the study adopts measurement matrix.(4)Selection of algorithm.In this study,the real time and portability of sound and light signal reconstruction are considered,Two representative classical algorithms are reconstructed and analyzed.A new algorithm for improving the efficiency of sound and light signal reconstruction is selected.(5)The reconstruction of sound and light signals is simulated.After completion of the above screening,M sampling is conducted for the sound and light signals of length N.,in the case of sampling rate(M/N),the reconstruction experiment was carried out with 0.25,0.5 and 0.75.Through the experimental data,it isfound that the sampling rate of 0.5 is a better choice.Different experimental baffles can obtain the reconstructed images with SNR over 25 dB under the sampling rate of 0.5.