Research On Multi-Echo Target Extraction And Single Photon Detection Of Streak Tube Imaging Lidar

With the introduction and practice of concepts such as digital city,digital earth,and smart city,people's demand for three-dimensional surveying and mapping has increased significantly.Streak tube imaging lidar has the advantages of high data rate,high data accuracy,and large scanning dynamic range.However,the current research on streak tube imaging lidar is still limited,and there is a lack of relevant research on deblur of streak image and multi-target echo information extraction of streak image.Besides,compared to single-photon array lidar and Geiger-mode avalanche photodiode array lidar,streak tube imaging lidar cannot achieve single-photon detection,which limits the detection range,data rate,and surveying efficiency of the system.Therefore,this article focuses on deblur of streak image,multi-target echo information extraction of streak image,single-photon detection of streak tube imaging lidar and accompanying noise problems,aiming to improve the data quality of streak tube imaging lidar.In the streak image,there existes dispersion which will reduces the spatial resolution of the system and interfere with the extraction of multi-echo targets.Therefore,deblur of streak image is needed.Although there is no research on deblur of streak image,there are many deblur research of passive image.So,this article draws on the passive image deblur method to study the streak image deblur.First,according to the characteristics of streak tube imaging lidar,the point spread function of the receiving system is measured.Then,the Wiener deconvolution filtering method and the Richardson-lucy iterative method are selected to deblur the streak image.The results show that the spatial resolution after Wiener deconvolution filtering method and Richardson-Lucy iterative method are 0.9mrad and 1 mrad,respectively,which is doubled and 80% higher than 1.8 mrad before deblur.Finally,the influence of the parameters in the two algorithms on the deblur effect is discussed.At present,the research on the information extraction of streak image mainly focuses on the single-target echo condition,and there is a lack of research on the multi-target echo condition.In this paper,waveform decomposition method of the traditional full-waveform sampling lidar is used to study the multi-target echo extraction of the streak image.Firstly,analyze and summarize the three steps of the multi-echo extraction method of traditional full-waveform sampling lidar: data preprocessing,Gaussian components and parameters estimation,and nonlinear fitting parameter optimization.Then,the direct Gaussian decomposition method and the wavelet transform Gaussian decomposition method are proposed,and the two methods are used to extract the target from the echo in four situations:single-plane target,double-plane target,small target and transparent target.Through experiments and analysis,it is found that the idea of Gaussian decomposition is also applicable to the extraction of multi-target echo information in the streak image,and the resolution limits of the two algorithms are both about0.2 m.After that,the data extraction effects of the parameters in the two algorithms are discussed.Finally,Richardson-Lucy iterative method is used to deblur the streak image of step target.The results show that Richardson-Lucy iteration method can effectively eliminate undesired multi-targets.In order to realize the single-photon detection of streak tube imaging lidar,this paper uses a higher gain microchannel plate and stripe tube,and the dispersion of the receiving system is smaller,resulting in a smaller size of a single photon spot.Through these means,this experimental device detects the smallest indivisible echo,a single photon,which proves that the receiving system has single-photon detection capability.Using the two-dimensional centroid algorithm,the electron dispersion of a single photon spot can be effectively eliminated.After accumulating many single-photon streak image of the same target,a count map can be obtained to restore the echo signal.Using the count value in the count map,the spatial resolution of the signle-photon mode is calculated to be 0.9 mrad,which is22% higher than the 1.1 mrad in the linear mode.There are a lot of noises in the single-photon streak tube imaging lidar point cloud.The sources of these noises include detector dark count,background noise,and optical dispersion noise.In order to eliminate these noises,this paper draws on the idea of spatial noise filtering of single-photon array lidar and proposes a spatial elongation noise filtering method.In order to better evaluate the effect of noise filtering,this paper proposes an evaluation parameter that considers the number of lost signals,the number of residual noise,and the minimum average distance between residual noise and the signal.After that,the simulation point cloud is used to quantitatively evaluate the noise filtering effect,and the best working parameters of each method are found;then the single-photon stripe array lidar point cloud data obtained before is filtered and processed for qualitative evaluation.The results show that compared with the spatial noise filtering method,the spatial extension noise filtering method preserves the signal details while reducing the false alarm rate by 23%.