【摘要】 基于全球导航卫星系统反射信号（Global Navigation Satellite System-Reflection,GNSS-R）的土壤湿度监测弥补了传统测量方法的不足，是近年来遥感领域研究的热点。针对土壤粗糙度及植被含水量影响反演精度的问题，研究了利用辐射计数据辅助提升精度的方法。提出了一种基于非线性自回归模型的神经网络（NARX）的GNSS-R和辐射计数据融合的土壤湿度反演模型，通过信号处理的一般流程，进行现场实验，验证了该方法。结果表明，在测试集上所提出的反演方法相比于传统的GNSS-R方法，相关系数提高了77%，均方根误差下降了78%，与辐射计方法相比，相关系数提高了47%，均方根误差下降了68%，证明了该方法可以实现对固定区域土壤湿度的长期连续观测。更多还原

【Abstract】 Soil moisture monitoring based on Global Navigation Satellite System-Reflection(GNSS-R) makes up for the shortcomings of traditional measurement methods, and has become a hot research topic in the field of remote sensing in recent years. In view of the influence of soil roughness and vegetation moisture content on the inversion accuracy, we studied the method of improving the accuracy by using radiometer data and proposed the soil moisture inversion model based on GNSS-R and radiometer data fusion with the neural network of nonlinear autoregressive model(NARX). The method is verified by field experiments through the general flow of signal processing. The results show that, compared with traditional GNSS-R method,the correlation coefficient of the proposed inversion method is increased by 77%, the root mean square error is decreased by 78%, and the correlation coefficient is increased by 47% compared with the radiometer method,the root mean square error decreased by 68%. We proves that the proposed method can achieve long-term continuous observation of soil moisture in a fixed area.更多还原