论述了主动、被动和主动被动相结合的3种微波遥感模态反演土壤湿度的方法和特点。为了采用人工神经网络方法反演出土壤湿度，在随机粗糙面双谱散射模型的基础上计算了裸土壤表面的散射系数和发射率，分析了3种不同的微波遥感模态和不同数据组合的等湿度区域分布的特点，从而确定了适合于各个微波遥感模态的输入数据组合。反演结果表明，只要选择适当的人工神经网络输入数据组合，采用3种模态中的任何一种微波遥感模态反演土壤湿度都是可行的，并具有较好的反演精度，结果对于微波遥感反演土壤湿度方案的选取具有指导意义。

The scattering coefficient and emissivity are both very sensitive to the moisture content of bare soil surface, which is the essence of measuring soil moisture by using microwave technology. The methods and characters of three microwave remote sensing modes for soil moisture retrieval are discussed in this paper. Firstly, scattering coefficient and emissivity for bare soil surface are calculated based on the theoretical backscattering model BSM. Then the paper analyzes the distribution of isomoisture regions of different data sets, and presents that soil moisture information can be retrieved by using ANN with appropriate input data set.