Inversion Study Of Vegetation Height Considering Multi-featured All-polarization SAR

With the rapid development of remote sensing technology,polarization interference technology,which combines the advantages of polarization and interference,is playing an increasingly important role in forest resource survey.Especially,it plays an irreplaceable role in quantitative inversion of forest height with its all-weather and multi-polarization advantages.At present,the study of forest height inversion model based on polarization interference has become one of the core problems of polarization interference forest remote sensing research,and it is of great significance to carry out research on it.On the basis of summarizing the latest theoretical achievements of polarization interference,a series of tree height inversion algorithms based on single-band and multi-band fully polarization data are deeply studied and improved in this paper.The validity of the algorithms is verified by using simulated fully polarization SAR data respectively.The main contents are as follows:(1)In this paper,an improved algorithm combining amplitude and phase is proposed.Compared with the traditional method,the improved algorithm has better robustness and higher inversion accuracy.Based on the analysis of the traditional polarization interference inversion tree height algorithm,the simulation experiments were carried out with the DEM phase difference method and the complex coherence amplitude method as the main line,including the use of three-stage algorithm to obtain the ground phase in the DEM phase difference method and the discussion and analysis of the coherent amplitude compensation phase difference method.(2)This paper calculates the inversion accuracy of each frequency band under the specific vegetation scene provided by the experiment and then carries out the overall advantage of joint inversion of vegetation height.The final results show that the accuracy of the joint multi-band algorithm is to a large extent better than that of the single band in the quantitative inversion of forest height.Considering that different frequencies of SAR bands have different penetration abilities,the multi-polarization and multi-band characteristics of full-polarization SAR are used to jointly invert the vegetation height,and C and X-band HV/VH polarization modes with shorter wavelengths are selected from the prior knowledge as the main information source carriers to obtain the scattering center of vegetation canopy.HH/VV polarization modes of P and L wavelengths with longer wavelengths were selected as the main information carrier to obtain the surface scattering center.Finally,on the basis of making full use of the limited singlescene C-band full-polarization data obtained,a method based on backscattering intensity was proposed to study the inversion of vegetation height,and random forest classification experiment was conducted on the full-polarization data,which verified the multi-feature advantage of full-polarization SAR data as a whole.(3)In this paper,the extinction coefficient is set as a constant value,and the effective vertical wave number is tried to replace the extinction coefficient in the two-dimensional table lookup model.The method is applied to the three-stage inversion algorithm and combined with the two table lookup models to obtain better inversion effect.Currently,various polarization optimization algorithms aim at maximal separation of ground scattering centers,which enables the three-stage vegetation height inversion algorithm based on the RVOG model,namely the Random Volume Over Ground(RVOG)model,to obtain more accurate surface scattering centers and near-canopy scattering centers.The improvement of the algorithm still requires a high-precision lookup table model.In principle,the extinction coefficient of RVOG model does not depend on polarization and has great influence on the whole vegetation multi-layer structure inversion effect,and the effective vertical distribution of scatterers does not change with polarization mode.However,the traditional lookup table model sacrificed the stability of extinction coefficient to some extent.