Neural Network and Fuzzy Logic for an Improved Soil Moisture Estimation

In the last two decades, various remote sensing techniques have been evaluated and proven to be a valuable source of information for different hydrological applications. Particularly, microwave remote sensing had been frequently used as alternative to traditional methods for estimating spatial soil moisture based on the large contrast between the dielectric properties of wet and dry soil. However, soil moisture response to microwave system from ground surface is mostly influenced by a variety of parameters such as land cover, vegetation density, and soil texture; which make the retrieval process more complex. In such conditions, non-parametric tools such as neural networks and fuzzy logic systems could have more potential in retrieving soil moisture from microwave sensors compared to traditional classification techniques. In this research, we have used Synthetic Aperture Radar data acquired by RADARSAT-1 satellite to retrieve the surface soil moisture along with vegetation-related information (vegetation optical depth and Normalized Difference Vegetation Index). The soil moisture data measured by Electronically Scanned Thinned Array Radiometer during the SGP97 campaign were used as truth data in the training and the validation processes. The performance of neural networks and fuzzy logic algorithms has been investigated by varying several parameters related to their structure and training processes. The preliminary results showed that for neural networks, the variation of the number of hidden layers and the number of neurons in each layer has no significant effect on classification accuracy. Concerning the fuzzy logic algorithm, the preliminary results showed that the cluster radius selection have a significant effect on classification accuracy. Further, the prediction made by neural network was found to be more accurate than fuzzy logic in several runs of the model, but the prediction made by fuzzy logic was more stable in nature.