In the development of future sonar systems, computer aided classification (CAC) becomes increasingly important. One element in a Multi-Beam/Multi-Aspect sidescan sonar CAC system is the segmentation and image fusion of the multidimensional data. Recent developments have been made to prevent the effects of aspect dependent backscattering strength (target strength) of objects lying on the seafloor. Because the target strength of an object varies with aspect angle, the sonar echo and consequently its representation in the sidescan image is rather random. To make sure maximum echo strength is obtained, overlapping bottom areas are insonified by temporal successive pings giving the echoes of the target under different aspect angles. The resulting images of successive pings are fused to one sidescan image. The existing image fusion algorithms now require an operator to set threshold values distinguishing target and shadow zones from bottom reverberation zones. The authors propose an unsupervised method for segmentation and optimal fusion of multidimensional sidescan sonar images for Multi-Beam/Multi-Aspect sidescan sonars.
[1]
Ramin Samadani,et al.
A finite mixtures algorithm for finding proportions in SAR images
,
1995,
IEEE Trans. Image Process..
[2]
I. Miller.
Probability, Random Variables, and Stochastic Processes
,
1966
.
[3]
Abdelwaheb Marzouki,et al.
Estimation of generalized mixtures and its application in image segmentation
,
1997,
IEEE Trans. Image Process..
[4]
Jun Zhang,et al.
Maximum-likelihood parameter estimation for unsupervised stochastic model-based image segmentation
,
1994,
IEEE Trans. Image Process..
[5]
Donald Geman,et al.
Stochastic Relaxation, Gibbs Distributions, and the Bayesian Restoration of Images
,
1984,
IEEE Transactions on Pattern Analysis and Machine Intelligence.