【摘要】 利用合成孔径雷达(synthetic aperture radar, SAR)卫星可对土壤湿度实现大面积、时间连续的反演。针对此过程中难以消除植被遮挡的问题,提出融合SAR微波数据和光学数据,利用修正水云模型消除植被影响,求解纯粹地表后向散射系数,代入裸土区土壤湿度反演算法进行土壤湿度反演,并在武汉豹澥区域设置实测站点记录实测数据进行比对。结果分析表明,经过植被消除后,实测站点垂直极化后向散射系数平均衰减量与植被覆盖度之间时间序列上相关系数达到0.83;土壤湿度反演值与实测值趋势相关性平均提升0.08,且提升程度与植被覆盖度相关系数也达到0.8以上,表明本算法针对植被覆盖度越大区域反演效果越好。该方法通过SAR影像和光学影像融合反演长时间序列土壤湿度具有良好可信度。更多还原

【Abstract】 Based on synthetic aperture radar(SAR) satellites, large-area and continuous time inversion of soil moisture can be achieved. While it is difficult to eliminate the problem of vegetation coverage in soil moisture inversion process, the modified water cloud model was proposed to eliminate the influence of vegetation coverage based on fusing SAR microwave data and optical data. The obtained pure backscattering coefficient was substituted into the soil moisture inversion algorithm of bare soil area. Meanwhile, the inverted soil moisture were compared with the ground-truth measurements in Wuhan Baoxie area. The results show that the correlation coefficient between the average attenuation of vertical polarization backscattering coefficient and vegetation coverage is 0.83 after removing the effect of vegetation coverage. The correlation of trend between the inverted soil moisture and ground-truth measurements is increased by 0.08 on average. The correlation coefficient between the increasing degree and vegetation coverage reaches at 0.8, which indicating that the larger vegetation coverage of the experimental area, the better inversion effect of the algorithm. This method illustrates that the long term sequence of soil moisture inversion through the fusion of SAR images and optical images is of great credibility.更多还原