A surface weaving approach to multibeam depth estimation

The performance of multibeam sonar systems is often limited by external interference, sidelobes, and poor seafloor reflectivity. Under such conditions, the acoustic return of the seafloor is still present but overshadowed by the interfering signal. This can lead to several well-known error conditions, including omega effects, isolated depth outliers, and tunneling. The surface weaving approach addresses this problem by exploiting 2D spatial statistical models of the seafloor, along with statistical models for the sensor and interference sources rather than estimate each sonar beam independently at each ping interval, as per current sonar depth estimation systems, surface weaving simultaneously estimates depths across a 2D grid that encompasses the entire beam set over a sequence of several pings. The proposed method, based on surface reconstruction techniques already in use for computer vision applications, has the potential to significantly reduce sonar artifacts