Terrain Correction Methods For Multi-dimensional SAR Data Applied To Forest Above Ground Biomass Estimation

As an important forest parameter,the forest above around biomass(AGB)plays a vital role in the global carbon cycle and climate change.Therefore,the accurate estimations of large-scale or global scale forest AGB are of great importance for human production and life.At present,among the many remote sensing technologies,multi-dimensional synthetic aperture radar(SAR)is one of the most promising technologies that can estimate forest AGB of large area.Compared with the single-frequency,single-polarization,single-baseline SAR,multi-dimensional SAR technology combined with different observation modes,such as multi-frequency,multipolarization and multi-baseline,can obtain more characteristic parameters associated with forest AGB on physical meaning.Therefore,the estimation accuracy of forest AGB can be improved based on multi-dimensional SAR.However,its application is greatly limited by the terrain.Owing to the characteristics of side-looking illumination by SAR sensors,terrain undulations seriously affect the quality of multi-dimensional SAR data.In recent decades,most methods proposed for radiometric terrain correction involved the backscattering intensity values in SAR data.These methods are not fully applicable to multi-dimensional SAR data.Aiming at this problem,this paper studied the terrain correction method for polarimetric SAR(Pol SAR)data and Interferometric SAR(InSAR)coherence,and the combined estimation approach of forest AGB based on multi-dimensional SAR data by integrating the terrain correction methods proposed above.The research contents include the following three aspects:(1)Three-step terrain correction method for PolSAR dataIn this paper,we propose a three-step semi-empirical radiometric terrain correction approach for PolSAR data.The three steps of terrain effects correction are: polarisation orientation angle(POA),effective scattering area(ESA),and angular variation effect(AVE)corrections.We propose a novel method to determine adaptively the “n” value in the third step by minimising the correlation coefficient between corrected backscattering coefficients and the local incidence angle;we then constructed the correction coefficients matrix and used it to correct PolSAR matrix data.PALSAR-2 HBQ(L-band,quad-polarisation)data were used to verify the proposed method.After three-step correction,differences between front and back slopes were significantly reduced.Our results indicate that POA,ESA,and AVE corrections are indispensable steps to producing PolSAR data.In the POA correction step,horizontal–vertical(HV)polarisation was maximally influenced by the POA shift.The max correction effect of the POA correction was greater than 1 dB for HV polarisation at an intermediate shift angle(±20°).ESA correction and AVE correction can achieve about 2.5 dB and 0.5 dB of correction effect.And based on Li DAR-derived forest AGB data,we analysed the relationship between forest AGB and backscattering coefficient;the correlation was improved following the terrain correction.HV polarisation had the best correlation with forest AGB(R = 0.81)and the correlation improved by approximately 0.3 compared to the uncorrected data.(2)Terrain correction approach for InSAR coherence based on algebraic difference methodBase on the theory of interference decorrelation and the idea of algebraic difference,we developed a terrain correction approach for coherence image of InSAR data.Firstly,based on the simplified InSAR decorrelation model(SINC model),we derive the SINC differential coherence model that can be used for terrain correction.Then,the model was evaluated by spaceborne(TanDEM/TerraSAR-X)and air-borne(CASMSAR)InSAR data.The experimental results showed that the proposed method can effectively remove the terrain effect in InSAR coherence image.And the proposed method can effectively improve the display effect of InSAR interference land use image(ILU)and enhance the ability of interpretation of InSAR data.In addition,the correlation of coherence and forest AGB was improved following the terrain correction.The correlation of corrected data improved by about 0.09 compared to the uncorrected data.(3)Combined estimation approach of forest AGB based on X-band In SAR and P-band PolSAR dataBased on the terrain correction method of PolSAR and InSAR coherence,we studied the combined estimation approach of forest AGB based on the domestic air-borne CASMSAR multidimensional SAR data.Firstly,the high resolution DSM data of the experimental area was acquired from the X-InSAR data.Then,the terrain correction of P-PolSAR data and X-InSAR coherence was completed.Finally,forest AGB was estimated based on the characteristics of multi-dimensional SAR that after the terrain correction.The experimental results showed that the accuracy of forest AGB was improved following the terrain correction.Compared with the uncorrected data,the estimation accuracy based on corrected P-PolSAR,corrected X-InSAR coherence and combination of corrected P-PolSAR and X-InSAR was improved by 0.9%,5.6% and 3.6%,respectively.Moreover,the combined multi-dimensional SAR features can obtain higher estimation accuracy than the single-dimensional SAR features.Compared to only using P-PolSAR features and X-InSAR coherence feature,the accuracy of combined estimation approach was improved by 6.4% and 5.1%,respectively.In summary,aiming at the terrain problem of multi-dimensional SAR data applied to forest AGB estimation,we developed the terrain correction methods for PolSAR data and InSAR coherence in this paper.The proposed methods were verified by space-borne and air-borne PolSAR/InSAR data.Experiment results showed that the terrain effect of multi-dimensional SAR data can be effectively removed by proposed method.And the accuracy of forest AGB estimation can be improved following the terrain correction.In this paper,the related research results can provide new ideas and technical support for the accurate mapping of large-scale forest AGB.