The Method Of Underlying Topography Estimation Over Forest Areas Based On Polarimetric SAR Interferometry

Underlying topography estimation over forest areas is the basic data for vegetation resource monitoring and the focus of forest mapping and forestry science research.Therefore,the underlying topography estimation over forest areas plays a vital role in many fields.With the development of remote sensing technology,the research on forest areas is becoming more and more extensive.Nowadays,optical remote sensing can accurately obtain the elevation data at the top of the vegetation canopy,but it is difficult to penetrate the vegetation canopy and trunk to obtain the information of the sub-canopy topography.So how to obtain the real terrain that does not contain forest trees is a problem that needs to be solved.Some of the bands in Synthetic Aperture Radar(SAR)systems have a certain ability to penetrate the vegetation cover.Therefore,the SAR system has an irreplaceable role in the underlying topography estimation over forest areas and tree height inversion.Polarized Synthetic Aperture Radar Interferometry(PollnSAR)has the characteristics of polarization SAR measurement that sensitive to the shape and direction of scatterers,and combined with interferometric SAR measurement to retrieve terrain changes using phase information.PolInSAR can obtain the information of vegetation parameters through backscattering information and phase information,which is difficult to obtain by other remote sensing methods,so PolInSAR provides a good technical guarantee for the inversion of forest terrain.Based on the detailed research of PolInSAR theory,aiming at the problem that the remote sensing method is difficult to penetrate the vegetation canopy to obtain a large-scale and high-precision understory of the forest.After focusing on the backscattering model,the method of underlying topography estimation over forest areas based on the X-band data PolInSAR is proposed.Aiming at the problem that P-band data is difficult to apply to the existing backscattering model,it is proposed to solve the model based on the existing backscattering model.In general,this paper uses different PolInSAR methods to invert the underlying topography estimation over forest areas according to the characteristics of different bands of data,focusing on solving the different method of underlying topography estimation applied by different bands to invert the understory of the forest as the main research content.Specific work and innovation are as follows:(1)A method for inversion of underlying topography estimation of PolInSAR based on Random-Volume-over-Ground Model(RVoG)is constructed.The surface phase of the underlying topography is obtained by the three-stage algorithm of RVoG model,and use this surface phase to achieve underlying topography estimation based on X-band.In-depth study of the principle of RVoG,clear the process of the three-stage algorithm.The RVoG model three-stage algorithm is proposed for the high-inversion of stand,but this paper starts from the surface phase of the underlying topography,complex-plane straight line fitting and surface phase estimation for complex coherence coefficients using a three-stage algorithm.It solves the problem that it is difficult to obtain accurate surface phase due to vegetation canopy occlusion,and proposes to use the obtained surface phase to realize the inversion of the underlying topography.In addition,for the problem of the missing elevation value of the Digital Elevation Model(DEM)under the forest,the missing elevation values are filled by Kriging interpolation.Because the X-band data satisfies the applicable conditions of the RVoG model,the method is validated by X-band data.The experimental results show that the proposed method can correctly estimate the underlying topography of the X-band data,The estimated under-forest DEM has the same trend as the corresponding digital surface model(DSM),and the difference between the two is basically consistent with the tree height.(2)The sub-aperture decomposition technique is used to increase the single baseline data observation information,and the sub-optical complex coherence obtained by subaperture decomposition is used to model.The new model is based on the RVoG model,and the sub-view complex coherence is used to replace the complex coherence coefficients in different polarization states in the RVoG model.It solves the problem that the P-band data penetration is high,which makes it heterogeneous in the forest height range,and the heterogeneity makes the P-band data not satisfy the RVoG model.Then,using the interference coherence of the sub-view image,the surface phase is obtained in the complex plane,and then the inversion of the understory is realized by phase unwrapping,phase-high conversion and geocoding.In addition,the problem of orbital error caused by the heavy-track acquisition method for P-band data is studied,and the polynomial fitting method is used to remove.The P-band data covering the Danling County of Sichuan Province was selected for experiments.The experimental results show that the new method can estimate the underlying topography based on the P-band.The DEM under the forest is basically consistent with the topographic trend of the real DSM,and the difference between the two is consistent with the height of the tree.