Seabed characterization using normalized backscatter data by best estimated grazing angles

Today´s multibeam echo-sounders (MBES) provide detailed bathymetric information that allows local slope corrections to be made to the concurrently collected acoustic backscatter imagery, thus improving the seabed characterization possibilities. Despite the availability of these data, local-slope corrections are not typically applied in the backscatter-processing scheme. We describe an approach that compensates backscatter for local slope by taking into account the vessel yaw, pitch and roll angles when determining the beam incident angle at the sea floor, and using all the available closest neighbor soundings within a certain radius. The normal of the 3D local curvature at each beam location is then calculated by a weighted least squares fit to the surrounding surface. The proposed method can use all the available geometric information from the co-registered depths and backscatter strengths either from a single survey line, or multiple survey lines, if overlapping lines exist. With the actual grazing angles computed, the transmitting and receiving beam patterns can be estimated to process and analyze the survey data for the extraction of seafloor property information in regions of complex topographic relief. As an example of this research, a backscatter image created from the multibeam data collected using a Simrad EM1000 sonar system on the mid-outer continental shelf off the New Jersey Margin, was normalized by the true grazing angles. When normalized, a better correlation between the independently measured grain size and backscatter was found, than with non-normalized backscatter data. With backscatter vs. grazing angle curve from a Gulf of Mexico multibeam survey, backscatter model parameter optimization was pursued in order to characterize the seabed by their spatial variations.