融合合成孔径雷达和光学影像的植被覆盖区土壤湿度反演

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.

Retrieval of soil moisture in vegetation covered area using SAR and optical image

Using synthetic aperture radar ( SAR) satellites can achieve a large area and continuous inversion of soil moisture. Aiming at overcome the difficulty of eliminating vegetation occlusion in this process, this paper proposes a method to fuse SAR microwave data and optical data, use the modified water cloud model to eliminate vegetation effects, solve the pure surface backscattering coefficient, and then substitute it into the bare soil area soil moisture inversion algorithm, and set up a measured site in Wuhan Baoxie area to record the measured data for comparison. The result analysis shows that, after the vegetation impact is eliminated ,in time series the correlation coefficient between the average attenuation of the vertical polarization backscatter coefficient at the measured site and the vegetation coverage reaches 0.83; the correlation between the soil moisture inversion value and the measured value trend increases by an average of 0.08. Moreover, the correlation coefficient between the lifting degree and the vegetation coverage also reached 0.8 or above, indicating that the algorithm has better inversion effect for the greater vegetation coverage. This method will has good reliability in retrieving long-term sequence soil moisture through SAR image and optical image fusion.