Research On Joint Coding Of Underwater Single Photon Video Transmission Based On Dictionary Learning

Long-distance underwater wireless video transmission with high efficiency and high reliability has important applications.In order to realize long-distance underwater wireless optical communication,a high-sensitivity single-photon detector is used to detect the weak light signal at the receiving end.However,due to the attenuation effect of the channel,the optical quantum effect and the existence of noise photons,the probability of the occurrence of deleted symbols is relatively large,while the probability of the occurrence of erroneous symbols is small in long-distance underwater single-photon communication,and the channel exhibits strong deletion characteristics and high bit errors.Dictionary learning(DL)is a sparse representation algorithm that can be efficiently applied to video and image compression,and the sparse matrix information of each frame of image does not affect each other,and can quickly and effectively restore continuous video streams.Therefore,this article focuses on the needs of underwater single-photon communication system for long-distance video transmission,focusing on underwater channel models and efficient and reliable video data transmission.The main research contents and results are as follows:(1)In view of the complexity of the photon movement in the underwater channel,comprehensively considering the photon emission situation and the characteristics of the detector,based on the underwater single-photon communication system,the method of photon counting in the time slot is used for data demodulation,and the underwater photon spatio-temporal random channel model is established,and then the Monte Carlo method is used to simulate the system.Through the statistics of the spatio-temporal information of the photons arriving at the receiving end,the influence of channel parameters on the optical receiving intensity,channel impulse response and system performance is studied.The simulation results show that the system can realize underwater communication with a theoretical communication distance of about 185 meters.(2)In order to reduce the amount of video transmission data,three video sparse coding schemes based on dictionary learning are designed,and two methods of generating the initial dictionary are proposed,and they are evaluated from the perspective of computer running time and reconstruction quality.Theoretical analysis and experimental results show that using the initial dictionary generated by the random selection method,after setting the number of iterations,number of atoms,sparsity and other dictionary parameters,the video residual-free sparse coding scheme based on dictionary learning can make the quality of video frame reconstruction better.(3)Considering the deletion of underwater channels and the randomness of successful decoding,a channel concatenated coding scheme of Luby Transform(LT)code concatenated Low Density Parity Check(LDPC)code with dual feedback mechanism is proposed.Aiming at the demand for efficient and reliable video transmission in underwater single-photon communication systems,a video joint coding scheme based on dictionary learning and LT-LDPC concatenated codes is proposed.The simulation experiment confirmed the feasibility of the channel concatenated coding scheme and the joint coding scheme;and the system experiment showed that by continuously adjusting the dictionary parameters,the video compression rate(CR)can be adjusted to a lower condition to reconstruct the video frame with higher quality.