Remote Sensing Imagery for Soil Characterization: a Wavelet Neural Data Fusion Approach

Technological advances in remote sensing imagery allows to obtain high resolution images, helpful in soil characterizing, monitoring and predicting natural hazards. On the other hand, the different kind of in-service sensors allows inferences on a large frequency band. In spite of this, due to the increasing requirements of industrial and civil entities, the academic research is actually involved in improving the quality of imageries. The aim is to implement automatic tools able to work in real-time applications, above all in order to solve pattern identification problem in remote sensing. Within this framework, our work proposes a data fusion methodology, based on the Multiscale Kalman Filter, in order to improve soil characterization in Ikonos surveys.