Research On Single Photon Counting Lidar 3D Imaging Technology With A Few Echo Photons

Along with the development of laser illumination,high sensitive single photon detection,high precision timing and high speed data processing technology,the new generation single photon counting three dimensional(3D)imaging was emerged.As a new type of active imaging technology,it has great application value in the fields of biophotonic imaging,weak target detection,remote sensing mapping,etc.It has become a research hotspot in recent years.After using the single photon counting mechanism,while improving the detection sensitivity,the shot noise originating from the light particle,as well as the influence of the ambient background noise and the dark count of the detector,become significant and cannot be ignored.Generally,in order to mitigate the influence of noise photons during the detection,it is necessary to accumulate hundreds or thousands of photons per pixel to form a histogram of arrival photons to obtain an accurate 3D image.However,in conditions of low light illumination or when detecting targets at long distances,the available echo energy and the allowable integration time generally do not satisfy the collection of such large numbers of photons.In these cases,the challenge we face is how to use a small amount of arrival photons to accurately restore the desired target image.Therefore,this dissertation focuses on theoretical and experimental research of single photon counting 3D imaging lidar with a few echo photons.The main research contents and innovations are as follows:1.Time-correlated single-photon counting(TCSPC)3D imaging system based on two different optical path modes(namely,the bi-static mode and mono-static mode)have been designed and constructed.3D imaging experiments of indoor targets and outdoor long-range(?1 km)targets were carried out.A galvanometer control program for grid scanning and an image data acquisition program based on equal time division were designed.The arrival time or time delay(i.e.,time-of-flight,TOF)of each echo photon is recorded.The full width at half maximum(FWHM)of the instrument response function(IRF)of the mono-static and bi-static TCSPC 3D imaging system was measured as 484 ps and 920 ps,respectively.The experimental research on the 3D imaging system based on a 32×32 array single photon detector was carried out,and the system deviation calibration experiment in the array detection process was completed.2.We systematically compared and analyzed five different depth estimation algorithms for single photon counting detection data,including traditional peak method and centroid method,as well as the new proposed cross-correlation method(Xcorr),union of subspace method(UOS),and a Markov chain Monte Carlo method(MCMC),which combined photon counting histogram with the IRF information.The experimental results show that when the average number of photons per pixel is greater than 13,the mean absolute error(MAE)of the depth image acquired based on Xcorr,UOS and MCMC methods becomes smaller and tends to be stable,and about twice better than the result of the peak method and the centroid method.Besides,a data model which is more suitable for expressing the distribution of echo photon histogram is introduced,a two-segment generalized Gaussian model,and it improves the convergence speed of the MCMC method for depth estimation by more than twice.3.A high photon-efficiency imaging algorithm based on spatial correlation was proposed to solve the problem of reflectivity and depth image restoration with a few echo photons.The core content includes establishing a single photon detection physical model,introducing the prior information constraints of target or scene,and noise photon censoring and filtering.Experimental results show that when the number of echo photons is small(11.7 photon per pixel),the depth resolution of the proposed algorithm can reach 4 mm,which is 4 times and 8 times higher than the histogram algorithm and the spatial median filtering algorithm,respectively.At the same time,the three-dimensional imaging experiment of 1 km long-distance target with a few ehco photons was completed.The results show that the proposed algorithm can obtain three-dimensional imaging with an average of about 0.65 photons per pixel,and the required acquisition time was reduced by a factor of 8 compared to median filter method,which greatly improves the photon efficiency.4.A five-step processing algorithm for single photon counting 3D imaging under strong background noise environment was proposed to solve the problems from the photon pipe-up effect and when the prior distance of target is unknown.The core content includes: correcting the photon counting histogram which is distorted by the photon pipe-up effect,locating the coarse position of the target,separating the signal and noise photon pixelwisely,filling the empty pixels using effective echo photons from neighborhood and the three-dimensional image regularization maximum likelihood estimation.Experimental demonstration and analysis were performed at different noise levels.The results show that the proposed five-step processing algorithm can achieve 3D image reconstruction when the signal-to-noise ratio is as low as 0.41.