CAJ | 학술논문

针对极化干涉SAR植被高度反演中RVOG模型未考虑地形影响，且三阶段算法受到地面相位估计误差和纯体相干性估计误差影响，提出了一种植被高度反演思路，采用考虑地形因素的S-RVOG模型作为反演模型校正地形影响，同时引入PD相干最优算法用于改善三阶段算法中直线拟合地表相位估计和纯体相干性估计精度。为验证算法的有效性，首先采用欧空局提供的PolSARpro软件模拟了不同地形坡度水平的PolInSAR数据进行仿真试验，然后采用德国宇航局提供的E-SAR机载全极化SAR数据进行真实植被场景测试，并进行了定性和定量分析。结果表明，本文方法对于不同坡度水平数据，均能有效改善传统RVOG反演模型中地形影响和三阶段算法自身误差影响，反演精度更高。
침대겁화간섭SAR식피고도반연중RVOG모형미고필지형영향，차삼계단산법수도지면상위고계오차화순체상간성고계오차영향，제출료일충식피고도반연사로，채용고필지형인소적S-RVOG모형작위반연모형교정지형영향，동시인입PD상간최우산법용우개선삼계단산법중직선의합지표상위고계화순체상간성고계정도。위험증산법적유효성，수선채용구공국제공적PolSARpro연건모의료불동지형파도수평적PolInSAR수거진행방진시험，연후채용덕국우항국제공적E-SAR궤재전겁화SAR수거진행진실식피장경측시，병진행료정성화정량분석。결과표명，본문방법대우불동파도수평수거，균능유효개선전통RVOG반연모형중지형영향화삼계단산법자신오차영향，반연정도경고。
The widely used random volume over ground (RVOG)model in forest height inversion with polarimetric SAR interferometric (Pol InSAR)technology do not consider the effect of terrain slope.Besides, the classical and widely appl ied three-stage inversion method is affected by the errors about estimation of underlying topography phase and pure volume coherence.In order to correct these two types of errors,a forest height inversion strategy is proposed.On the one hand,a slope RVOG (S-RVOG)model is adopted in this paper which incorporates the range terrain slope factor to correct the terrain distortion;on the other hand,the phase diversity (PD)coherence optimization is introduced to improve accuracy of underlying topography phase estimation in l ine fit process and pure volume coherence estimation.In order to verify efficiency of the method,a simulation experiment is carried out by using Pol InSAR data in different range slope levels simulated by PolSARpro software provided by ESA.Then,E-SAR ai rborne ful l polarization data provided by DLR are uti l ized to test real scenario.Final ly,it is conducted that some qual itative and quanti-tative analyses.The results show that for the different data in different range slope levels,the inversion method promises to correct effect due to terrain distortion in RVOG model and compensate error effect in traditional three-stage inversion method.Therefore,the proposed method provides much more accurate estimation of forest parameters.