K. Kim, N. Intrator, and N. Neretti (USA)
image registration, image mosaicing, acoustic camera, DIDSON
This paper presents an algorithm for image registration and mosaicing on underwater sonar image sequences character ized by a high noise level, inhomogeneous illumination and low frame rate. Imaging geometry of acoustic cameras is significantly different from that of pinhole cameras. For a planar surface viewed through a pinhole camera undergo ing translational and rotational motion, registration can be obtained via a projective transformation. We show that, un der the same condition, an affine transformation is a good approximation for an acoustic camera. We further propose a novel image fusion method, which maximizes the signal to-noise ratio of the mosaic image. The full procedure in cludes illumination correction, feature based transforma tion estimation, and image fusion for mosaicing.
Important Links:
Go Back
