Sidescan sonar mosaicing improvements using a tracking algorithm combining forward looking sonar images and ship heading oscillations
Résumé
To offer optimal data quality, devices like magnetometers or sidescan sonars (SSS) are traditionally deployed near the seabed, often put on a towfish, linked to the ship by a cable. Unfortunately, this line implies imprecision in the towfish locations. In order to obtain a positioning of the towfish with respect to the vessel, a forward looking sonar (FLS) is used to track the different towfish location during the time. A bespoke experiment, realized for this study in Brest Harbor in an area of about 500x100 meters, is first presented. Then, a tracking algorithm is performed to find the SSS reflector in the FLS images. This algorithm uses a Kalman filter to link ship heading oscillations to FLS images to give SSS positions during the time. At that time, the image sonar mosaicing is carried out in a geographical information system (GIS), with the help of the tracking results and the GPS positions of the ship. The last part of this work is to compare the different SSS mosaics results (with or without layback corrections) to the digital terrain models (DTM) of the same area, in order to estimate the benefits of these corrections.