Study on side scan sonar image matching based on the integration of SURF and similarity calculation of typical areas

In this paper, a method for side scan sonar image matching is proposed based on SURF and similarity calculating in typical areas. The process is composed of three steps of rough matching, geometric correction and accurate matching. Firstly, rough matching is implemented to get a rough matching relationship of reference image and matching image based on SURF. Then geometric correction is executed with the relationship to make the matching image have a same viewing condition with the reference image. Finally, accurate matching is implemented by similarity calculating of typical areas in two corrected images. This matching algorithm combined the merits of the regional matching and feature matching. Experiment proved that the method has advantages in speed, accuracy, reliability and invariance to geometric differences between two side scan sonar images.