Strategies for unwrapping multisensors interferometric side scan sonar phase

This paper concerns the description of a method of 2/spl pi/ bias removal for the interferometric phase. The first part is the description of different noise sources for interferometric side-scan sonar. In fact, these noise sources can generally be classified into two categories: a. environmental noise category (variable celerity, sonar motion, etc.), and, b. interferometric process related noise (decorrelation baseline, decorrelation of wave train, propagation, multipath propagation). Each source of noise is described as a physical phenomena. This study aims at developing a global density of probability function (dpf) including all statistical phenomena corresponding to the 'b' category of noise (interferometric process). The second part checks the statistical properties of phase on real data and uses them as a confidence factor in the 2/spl pi/ bias removal process. The third part deals with the global process using the vernier technique to derive benefit from the interferometric multi arrays and the theoretical properties (vernier efficiency) connected to the specific configuration of the sonar. The main conclusion of this paper concerns the theoretical interest of the global comprehension of the vernier technique (statistical properties checked on experimental data and intrinsic properties of the vernier) to plan strategies to remove the 2/spl pi/ bias and gain of treatment on real data. Based on these results different applications are possible such as the incorporation of noise in simulation process; this allows data collection in more complex virtual experimental scene than in a real scene for example.