Generation an ideal DEM by fusion shape from shading and interferometry bathymetries for seafloor remote sensing

In sonar imaging for seafloor remote sensing, research activities are more and more oriented on the use of data fusion approaches. Nowadays, it is well established that using sonar images, the Digital Elevation Maps (DEMs), can be generated by exploiting either the amplitude information or the phase information of the acoustic signal. In this paper, the main interest consists on the generation of a complete Digital Elevation Map (DEM) by the use of a data fusion approach of two existing DEMs issued from two different techniques. The aim of the proposed approach is to elaborate a general interpretation system that coherently links works on data selection and fusion leading to improve DEMs generation and to exploit it in the seafloor remote sensing applications (particularly for the inhomogeneous scenes with a variety terrain). In this paper, shape from shading and the interferometry techniques are considered. Then, the manner of the DEMs fusion proposed, has been based on fuzzy logic and some fuzzy propositions, which defined using experts a priori knowledge source. This promising idea enables information to be managed through the consideration of the imprecision and ambiguity information and the benefit provide by the injection of the a priori knowledge in the decision taken system.