A DNN-Based Inversion Algorithm of Soil Moisture

With the rapid development of computer science, deep learning methods have achieved applications in many fields. Due to the time-consuming and laborious measurement of soil moisture during ground scattering characterization experiments, a deep learning (DL)-based inversion algorithm of soil moisture was proposed in this paper. First, the Oh model was utilized to generate the training samples. Then, a deep neural network (DNN) was constructed to establish the relationship between soil moisture and the back scattering coefficients. Finally, the inversion values were compared with the measured data. The results demonstrated that the proposed DNN model outperforms other models in terms of accuracy.