Abstract

In the paper, the approach to seafloor characterisation is presented. The multibeam sonars, besides their well verified and widely used applications like high resolution bathymetry and underwater object detection and imaging, are also the promising tool in seafloor characterization and classification, having several advantages over conventional single beam echosounders. The proposed approach relies on the combined, concurrent use of several techniques of multibeam sonar data processing. The first one is based on constructing the grey-level sonar images of seabed using the backscattering strength calculated for the echoes received in the consecutive beams. Then, the set of parameters describing the local region of sonar image is calculated. These include both the first and the second order statistics of the grey level, and the texture analysis. The second technique utilises the 3D model of the seabed surface, which is constructed as a set of (x, y, z) points using the detected bottom range for each beam in the multibeam system seafloor imaging procedure. For the local region of seabed surface, the descriptors like rms height and autocorrelation radius are calculated. The third technique assumes the use of a set of parameters of the multibeam echo envelope, similarly as in single beam classification. The parameters include echo energy and its statistics, as well as the set of echo shape descriptors, and are calculated for each beam allowing the estimation of their dependence on seafloor incident angle. Then, for selected parameters, the characteristic features quantitatively describing their angular dependence, like slope, or range, are calculated. Finally, the features obtained by these 3 techniques have been combined together and used in seabed supervised classification procedure based on standard classifiers. The proposed method has been tested using multibeam data records acquired from several bottom types in the Gulf of Gdańsk region. The obtained results show that application of the proposed combined approach improves the classification performance in comparison with those of using only the one scheme of seafloor multibeam data processing.

Author (1)