| Forests are an important part of terrestrial ecosystems and research on carbon cycling and global climate change requires a large number of high-precision forest parameters worldwide.Due to the complexity of the spatial structure of forest ecosystems and the wide geographical distribution of forest resources,forest resource surveys and forest ecological monitoring require specialized remote sensing techniques to provide qualitative and quantitative data urgently.The spaceborne laser altimeter is a new type of active remote sensing system with high operating orbit,continuous observation throughout the day,global coverage,strong anti-interference ability and good vegetation penetration,etc.Besides it can also achieve high-resolution vertical information of the forest.At present,there is no specialized altimeter satellite for forest monitoring in the world.In order to draw on the strengths of laser altimeter in forest survey,this paper focuses on the design and application requirements of the laser altimetry payload for the satellite project which has been approved to monitor the terrestrial ecosystems in the country.According to the application requirements,a semi-analytical vegetation model based on the radiative transfer model and a simplified satellite laser altimetry forest echo model were established.The influence of the system parameters,surface and vegetation parameters on the received waveform and the inversion of the vegetation were systematically analyzed.The data processing method of received waveform of the forest and the inversion method of the canopy height are improved.The research of the inversion and verification of the forest canopy height,using the data of GLAS,are carried out by applying the methods proposed by this paper.The thesis carried out theoretical and technical research on the theoretical modeling and system parameter optimization,data processing and inversion of vegetation parameters of the space borne laser altimeter for vegetation target.The main work of the thesis includes the following cotents:(1)Based on the classical theory of the spaceborne laser altimetry system,a simplified analytical model of the forest target waveform is established,and the decomposition conditions of the vegetation waveform are further derived.Based on this,the laser divergence angle suitable for forest target observation is analyzed,and the result indicates that the reduction in beam size will reduce the negative effect of the slope on the waveform and improve the usability of the data measured by the laser altimeter.When the orbit of the satellite is about 500 km,the optimal laser divergence angle is 60 ?rad,i.e.the diameter of the laser beam is 30 m.The decomposition constraints of the three typical laser spot spatial energy distributions are derived.It is proved that the slope threshold of the fundamental mode Gaussian spot is higher,which is beneficial to the detection of forest targets.According to the sampling accuracy requirements of the forestry survey,3 beams can meet the survey sampling accuracy.The laser altimetry payload scheme is adopted in the demonstration of terrestrial ecosystem carbon monitoring satellites.(2)Based on the geometric parameters and optical parameters of forest vegetation targets,a discrete three-dimensional model of vegetation targets was established.Based on the radiation transfer theory,a semi-analytical model of laser altimeter for forest vegetation targets was constructed and the waveform simulator based on the model was developed.By analyzing a large number of real data,the statistical rules of forest tree height distribution and canopy distribution were obtained,and an unbiased forest spatial distribution model was constructed.The simulated waveform was calculated and correlated with the measured results of GLAS.The average coefficient of determination R2 reaches 0.91.The unbiased spatial distribution model of forest based on tree height distribution and canopy distribution is verified and the correctness of forest waveform simulator is proved.(3)The waveform simulator for the forest vegetation target based on the semi-analytical model was used to quantitatively analyze the influence of system parameters,terrain parameters and vegetation canopy parameters' effect on the waveform.A large number of simulation results show that the surface slope,roughness and canopy height jointly affect the length of the waveform.The differences between the waveform length and the broadening of the pulse increase from 0.15 m to 0.75 m and form 1ns to 5ns respectively,when the slope varies from 0° to 30°;the absolute value of the change caused by roughness is on the order of 10 ns,the difference in variation is less than 5 ns,corresponding to a height of 0.75 m.Therefore,the accurate decomposition of surface waveforms and the pulse broadening can comprehensively reflect the influence of surface slope and roughness,and can be used for vegetation height inversion correction.(4)A self-adaptive filtering based on the characteristics of waveform from forest target is proposed.Combined with the waveform data processing algorithm of Gaussian iterative decomposition,the physical interpretability of forest waveform is improved,and the probability of mislocalization of decomposed wavelets is reduced.At the same time as the superimposed pulses are decomposed correctly,the pseudo-waves generated by the over-fitting are suppressed.Among the 829 data points decomposed by the algorithm proposed by this paper and NASA,the number of data that has a correlation over 0.95 is 97.46% and 94.33%,respectively,and the data with R2>0.90 account for 99.76% and 97.95%,respectively.Compared with the classical Gaussian inflection point method,the algorithm proposed by this paper can decompose more sub-pulses while processing the laser pulses for vegetation.The comparison also indicates that the decomposition can obtain more vertical information of the vegetation.(5)The classical canopy height inversion physical geometry model is improved,and a new inversion method of maximum canopy height and average canopy height is proposed,which can overcome the limitation that applying statistical analysis requires a large amount of measured data for regression analysis.Moreover,the canopy height can be well estimated according to the characteristics of the waveform and without the DEM data.The US Teton Forest Park area and the Lushan Forest Park area in Henan Province,China were selected as the demonstration area.Using the GLAS data and the algorithm proposed by this paper to inverse the maximum canopy height,the RMSE of the results are 2.81 m and 2.98,respectively and the R2 are 0.67 and 0.61,respectively,which indicates that the inversion results of the two regions are both better than the existing methods. |
PDF Full Text Request