Research On Full-waveform Data Processing Of Airborne Lidar System

Airborne LiDAR system is an active remote sensing system composed of laser scanner,inertial navigation system and other subsystems.It has the characteristics of non-contact and high precision,and can accurately obtain the three-dimensional coordinate information of the target to be measured.It is widely used in civil fields,such as building extraction,forest tree classification,shoreline survey and so on,as well as in many military fields,such as underwater exploration,mine detection and so on.The full-waveform airborne lidar system that has emerged in recent years can obtain complete ground echoes and provide clearer target vertical changes.The core task of measurement by airborne lidar system is to process the full-waveform data.Because the full-waveform data contains multiple echo signals and is affected by environmental noise in the process of data acquisition,it is necessary to preprocess the full-waveform data first.Secondly,the waveform can be decomposed.Aiming at the airborne lidar system,this paper focuses on the full-waveform data processing method of the system.The main contents are as follows:(1)Explain the domestic airborne dual-frequency LiDAR system:Firstly,the composition,scanning method and system applicability of domestic and foreign airborne LiDAR system are studied.Secondly,from the introduction of the RADAR equation,the influence of the airborne lidar parameters and the target backscattering rate on the laser waveform is introduced.Based on this,the characteristics of the full-waveform data echo of the airborne LiDAR are described.Finally,the principle of local positioning of airborne LiDAR system is introduced,and the complete process of laser foot coordinate conversion to WGS84 coordinate system in laser scanner coordinate system is derived.(2)Research on the denoising algorithm of full waveform data filtering:Although the dark current noise can be reduced through aiming at dark current noise,Gauss filter,Butterworth filter,the broadening is changed and the waveform character is damaged greatly;however,the S-G filter does not change the broadening of waveform data and it can reduce the effect of dark current noise effectively.As to the background noise,using the absolute median deviation method to estimate the threshold can not eliminate the background noise of full waveform data which has high echo intensity,while the averaging method can estimate the background noise threshold to eliminate the influence of background noise.To synthesize the experimental results,S-G filter and the averaging method are combined to estimate the background noise and have the denoising algorithm of full waveform data filtering.Finally,it proves that the method selected in this paper is universal through the denoising results of different objects.(3)Research on the decomposition algorithm of full waveform data filtering:For the shortcomings of existing decomposition algorithm of full waveform data filtering on the detection of weak wave and superimposed wave,an improved and strict Gauss detection algorithm is put forward in this paper.The waveform decomposition is completed by the steps of waveform peak detection,model parameter calculation,shape of wave fitting,component elimination of invalid wave and so on.The L-M algorithm is used to optimize the model parameters.Through the analysis of decomposition results of the imitate and actual full wave data,the The method used in this paper can complete the waveform decomposition effectively.