Research On Full-waveform Processing Method Of GF-7 Laser Altimetry System

With the increasing application of deep space exploration activities and rapid development of earth observation technology,satellite laser altimetry technology,as an important means to obtain vertical structure and three-dimensional information of earth's surface accurately,has been paid more and more attention at home and abroad.On November 3,2019,"Gaofen-7"(GF-7)satellite was successfully launched.It is the first civil laser stereoscopic satellite for surveying and mapping in China,which will be mainly applied in 1:10000 stereoscopic surveying and mapping under the condition of few control points.This dissertation relies on the project of "Laser data processing and feature parameter extraction technology for GF-7 satellite,a major special project of high-resolution earth observation system in China",and focuses on the key technical problems such as the correspondence and law between the received waveforms of laser altimeter and the surface features,the quality evaluation and pre-processing of waveform data,the parameter extraction of waveform features,the accuracy and error analysis of ranging,the positioning accuracy and error analysis of laser foot points,etc.In order to provide technical support for high-precision geometric processing of GF-7satellite data,accurate extraction of surface elevation,height information of ground features and change monitoring,the corresponding processing methods were studied,and a software system suitable for domestic laser altimetry data processing was developed.The main work and achievements are as follows:(1)Based on the principle and system structure of satellite laser altimetry,a precise ranging model is established according to the center of gravity or fitting center of the transmitting and receiving waveforms.In addition to the hardware error,atmospheric delay error and noise error of the laser altimeter itself,the ranging error caused by the characteristics of laser beam divergence angle and surface target(slope and roughness)are simulated and analyzed.By analyzing the relationship between the slope and the ranging error,the elevation control points which can meet the requirements of certain ranging accuracy can be selected.(2)The models of different processes of laser emission,transmission,surface and reception are established and the whole link waveform simulation is carried out.Using high-precision airborne point cloud data to build different terrain references,revealing the corresponding relationship between different terrain,surface features and waveforms,improving the accuracy of waveform interpretation,and providing the basis for laser system parameter design;The relative elevation is extracted by using the high-precision terrain simulation waveform,which provides reference and contrast verification for the subsequent waveform decomposition to extract the ranging information;It can help to relocate the laser footprints and solve the problem of obtaining the elevation control points under the complex surface.(3)An adaptive denoising method of received waveform based on EMD is introduced,and a Hurst index method is proposed to remove high frequency noise.According to the simulation experiment and GF-7 simulated waveforms experiment,compared with other methods such as semi-automatic settings to directly remove the first n IMFs or add threshold function,the EMD-Hurst denoising method has the highest mean value of consistent ratio and the most comprehensive denoising effect.(4)In this dissertation,the decomposition model of combining deconvolution and Gauss decomposition is put forward.Through a large number of simulation experiments,it is shown that the combined decomposition method can significantly improve the extraction of the number of back wave peaks(the additional detection rate of the four combined methods in the simulation of complex forest area is at least15.8%);In simulated GF-7 urban scene waveform experiment,Blind deconvolution and Gauss combination decomposition method with average 3.5 Gaussian components,which can detect more peak information than other methods with average 1.2 or 2.5Gaussian components.(5)Based on QT 5.9.0 platform and C + + language,the data processing module of full waveforms is developed.The satellite laser altimetry system can carry out the pre-processing and Gaussian decomposition of waveforms,extraction of waveform length,peak length and other waveform characteristic parameters,atmospheric delay correction,laser footprints positioning,tide correction,waveform data quality evaluation,etc.