Data-Driven Autocalibration for Swath Sonars

Sidelobes in swath sonar water column imagery can obscure targets of interest and create erroneous bottom detections. Differences between ideal and actual element responses limit the achievable sidelobe level for practical arrays. Therefore, removing or reducing the errors by calibration may reduce the sidelobe level. We examine an autocalibration technique for swath sonars that use data collected while mapping. Our method is based on the generalized interferometric array response (GIAR). The GIAR values are used to find pointlike signals and maximized to estimate the amplitude and phase errors for each element. On simulated data, this reduces the sidelobe level to below -50 dB. In the field, we find the calibration estimate stable over ping groups and that most sidelobes improved by 2 dB or more. This suggest that GIAR calibration can work well when the calibration matches the swath sonar system. However, some sidelobes are also unchanged or degraded by a few decibels. An extended calibration model seems necessary to achieve further improvements in the field.