Research On Key Techniques Of Photon-counting Lidar

Lidar(Light detection and ranging)has the advantages of high angular resolution,high ranging precision and strong anti-interference capability and is widely used in military and civilian fields.Based on the Geiger-mode avalanche photo diode(Gm-APD),the photoncounting lidar improves the detection sensitivity of the echo-pulse to the quantum limit and breaks through the limitation of the traditional lidar in long ranging and weak echo-signal detection.It's one of the main developing direction in lidar field.Because lidar based on GmAPD array has the problem of crosstalk among pixels and high cost,researchers usually use the single-point detector to establish the photon-counting lidar system combined with the scanning system.However,the scope of applications and detection performance of the Gm-APD 3D imaging lidar are limited by the slowdown of imaging speed caused by scanning and the falsealarm response caused by the background noise and the range walk error of Gm-APD detectors.Based on this,this doctoral dissertation aims at improving the detection performance of the Gm-APD 3D imaging lidar.Based on the lidar equation and Poisson response model,the doctoral dissertation studies on the key techniques of improving scanning efficiency,restraint of range walk error and reducing the false-alarm response.The main contents of this dissertation are as follows:To solve the problem of slow imaging speed of single point imaging lidar,a multiresolution adaptive scanning method is proposed to reduce the useless scanning points to improve the scanning efficiency.An algorithm based on the spatial correlation is developed to determine the scanning resolution of different zones.Based on this,the distribution of the scanning points is relevant to the changes of depth value in the scanning field of view.Then the limited scanning amount is distributed near the target profile as much as possible and the useless scanning points in depth continuous zone are reduced.Finally under the premise of integrity of the target profile,the amount of scanning points decreases and the imaging is speeded up.The experimental results show that the multi-resolution adaptive scanning method reduces 85%scanning time of the traditional method which uses a fixed scanning step and ensure the imaging precision meanwhile.The proposed method well solves the contradiction between the imaging speed and imaging precision.To deal with the range walk error from Gm-APD detectors,a restraint method for range walk error based on the special correalation is introduced.Compared with the existing methods,the proposed method doesn't need to model and calibrate the system in advance and isn't limited by the specified parameters of the prior modeling.Therefore the proposed method has more advantages in convenience and accuracy.The proposed method isolates 10% of energy from the echo-pulse to detect individually and calculates the range walk error distribution matrix combined with the depth image.Then the noises in the range walk error distribution matrix are reduced combined with the intensity image based on the relationship between the range walk error and the intensity of the echo-pulse.Finally,the depth image is corrected by the range walk error distribution matrix and the range walk error is restrained.The experimental results demonstrate the feasibility and the effectiveness of the proposed method and show that the proposed method reduces the range walk error from 35 mm to 5mm.To reduce the false-alarm probability and improve the ranging precision of the Gm-APD3 D imaging lidar,a reconstruction method based on the dual-detector structure is proposed.Different from conventional dual-detector structure,the proposed method divides the echopulse into two parts in an unequal splitting ratio and the two parts are detected respectively by two Gm-APDs to obtain different photon-counting histograms.Then the distribution of signal photoelectrons in time domain is reconstructed to obtain the accurate time-of-flight based on the Poisson response model and the known splitting ratio.The proposed method uses the dualdetector structure to reduce the false-alarm probability and exploits extra information from the unequal splitting ratio to reconstruct the distribution of signal photoelectrons in time domain to reduce the range walk error.The theoretical analysis and numerical calculation show that the best splitting ratio is 4:6.The experimental results show that the proposed method reduced the standard deviation from 14.8mm to 3.6mm and efficiently improves the ranging precision of the Gm-APD 3D imaging lidar.