Theoretical Modeling And Information Acquisition Method Of High-Flux Photon-Counting Lidar

With the progress of technology,the performance of detectors with single-photon sensitivity is getting better.Compared with traditional linear mode detectors,single-photon detectors have the advantages of high sensitivity,high integration,and low power consumption,which has given birth to the research upsurge of photon-counting lidar.However,due to the working mechanism,traditional photon-counting lidar is usually limited to the low-flux condition.In low-flux conditions,it requires lots of pulse accumulations and thus severely limits the data acquisition efficiency.In recent years,the study of photon-counting lidar under high-flux conditions is a growing interest for the field of research.Therefore,this paper studies the mechanism,characteristics,and information acquisition methods of high-flux photon-counting lidar.This paper has important theoretical and practical significance for exploring the universal mechanism,detection method,and expanding its application scenario of photon-counting lidar between‘low to high flux’.(1)Forward recursive model for photon-counting lidar:To solve the output function of photon-counting lidar,a‘result oriented’expression of photon detection probability is proposed.Combined with the conclusion of free-running converging to the steady-state,and overlapping domain elimination technology,a set of general,accurate,and efficient recursive model of detection probability with dead time distortion effect is established and extended to the special case of long dead time.Compared with the existing model,the calculation efficiency is improved at least one order of magnitude.On this basis,this paper discusses the process of the detection probability converging to the steady-state under the free-running mode,puts forward the algorithms of steady-state response and sub-steady response,and discusses the error of the classical short dead time approximation model.The error is due to its only considering the‘arrival processes’of photons.(2)Influence of flux controlling on the performance of photon-counting lidar:Based on the forward model,to study how the system performance changes with the flux level,this paper considered the counting signal and echo signal of the system under different fluxes,dead time,and considered the free-running operation mode.The results show that when the number of single pulse echo photons is greater than 1 or the number of background noise photons in a single detection cycle is greater than 2,appropriate attenuation of incident light can change all performance in a favorable direction,and shorter dead time requires less attenuation.In the working condition that the noise is much higher than the signal intensity,a‘2/B’optimal attenuation criterion is proposed in this paper.Verifications show that the root means square error between this criterion and the theoretical optimal method is reduced RMSE at least by 50%compared with the best method at present.(3)Full waveform recovery method for high-flux photon-counting lidar:To cancel the dead time distortion,range bias and for better parameter estimation,this paper proposed a fast analytical method without prior knowledge to recover the original waveform from distorted data under high-flux conditions,which assisted by the moving average de-noise operation guided by laser pulse width.On this basis,combined with the maximum cross-correlation assisted centroid algorithm,the high-precision extraction of range under high-flux conditions is realized.The results show that under the condition that the dead time occupies more time bins,our method is more efficient,and the full waveform recovery method in this paper improves the waveform recovery efficiency by at least one order of magnitude compared with the representative method and reduces the average relative error of waveform recovery by more than 80%.The range extraction method in this paper reduces the walking range error by 94%on average.(4)Parameters extraction for high-flux photon-counting lidar:Facing the challenge of high-flux distortion,current parameter estimation methods require solving complex optimization problems and rarely focus on the combination of broadening and distortion.To extract the parameters directly,based on the fast waveform recovery method mentioned above,this paper adopts a cross-correlation assisted centroid algorithm and Fourier analysis method to realize‘signal intensity’,‘noise intensity’,and‘depth’.Then,under the assumption of the conical divergent beam,a joint parameter extraction method of‘conventional parameters’and‘tilt angle’is established for high-flux broadening and distortion conditions.The results show that the parameter extraction method under the assumption of an ideal light spot improves the solution accuracy of signal and noise parameters by about 20%compared with the maximum likelihood estimation method,obtains the depth accuracy of the same order of magnitude,and improves the calculation efficiency by at least 20 times.The parameters of the above four dimensions are effectively extracted under the conical divergent beam hypothesis.Among them,the average absolute error of the tilt angle extracted under different plane types ranges from2.56 degrees to 10.27 degrees.(5)High-flux single-photon lidar fast detection using pulse sequence:To cut down the number of accumulated pulses and accelerate the ranging speed.This paper utilized the effect of dead time and concluded the optimal pulse interval criteria of‘≈dead time+pulse width’.Considering the cross-correlation characteristics,threshold,and cluster characteristics of photon signals,we proposed a FARE(Few Accumulations Range Extraction)method.The results show that the‘optimal interval+FARE’method is more suitable for high-flux than the conventional coding method and the classical single pulse detection method,especially under the condition of strong background noise.When the number of echo photons is equal to 9 and the noise rate is less than 5×10~6 cps,our method achieves the ideal result of‘one transmission,one detection’.This paper accomplished the high-flux forward modeling,detection performance analysis using attenuation in high-flux conditions,the full waveform fast recovery method,parameters extraction method,and pulse coding fast detection method for photon-counting lidar.The above studies have important theoretical and practical significance in filling the theoretical and methodological gap between‘low to high-flux’regimes,exploring the unique advantages of the system,and expanding the scene adaptability.