Research On Geiger-APD Array Lidar Range Image Stepwise Reconstruction Algorithm

Geiger-mode APD(Avalanche Photodiodes)array lidar imaging has the advantages of high sensitivity,high accuracy,and high integration,which makes the detection of long-range targets in the field of lidar one of the current research hotspots.This paper mainly focuses on the difference in imaging properties between the traditional range profile and Geiger-mode APD array lidar range profile,and conducts a corresponding range image reconstruction method study,so as to achieve the purpose of improving the target reduction degree and image quality.Firstly,this paper investigates the current research status of GM-APD array laser imaging at home and abroad,and determines the main research content through its analysis.At the same time,the working principle of GM-APD and the system noise model is analyzed and studied,and the working process of the radar system is simulated for subsequent algorithm verification.Finally,an objective evaluation index suitable for the distance profile of the system is selected for evaluating the performance of the algorithm.Secondly,this paper tests the effect of traditional filtering methods on reconstructed images,and uses improved bilateral filtering and adaptive Kalman filtering to process the original data.According to the advantages and disadvantages of each method,a combination of bilateral filtering and adaptive gradual filtering is proposed.The stepwise restoration method of eliminating Kalman filtering and neighborhood gradient weighted operation,and introducing the gray morphology to smooth the remaining distance anomalies and the missing information in the reconstructed distance profile.Evaluation indicators show that the improved signal-to-noise ratio is 25.40,the root-mean-square error value has decreased by 38%,the structural similarity has increased by 27%,a large amount of l ost information has been restored,and the image performance of the simulated distance profile has been greatly improved.Finally,in order to verify the reconstruction ability of the above algorithm and the adaptability to the system,use a fixed frame number of strong laser real echo data to verify,improve the signal-to-noise ratio value is 4.13;the root mean square error value compared to the noise-containing distance The image has dropped by 40%;the structural similarity has increased by 27% compared to the noisy distance image,and the quality of the reconstructed distance image has been improved.Using a fixed frame number of weak laser real echo data for verification,the image quality evaluation index shows an improved signal-to-noise ratio of 3.44;the root-mean-square error value is reduced by 28% compared to the noisy distance image;the structural similarity is increased by 40%.Some of the missing information in the range profile has been successfully reconstructed,and the target edge information has been well protected while filtering the anomalous distance.The performance has been greatly improved compared to the original image.