Sonar image segmentation using an unsupervised hierarchical MRF model

• Published in: IEEE transactions on image processing Vol. 9; no. 7; pp. 1216 - 1231
• Main Authors: Mignotte, M., Collet, C., Perez, P., Bouthemy, P.
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.07.2000; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Acoustic noise; Acoustic waves; Algorithms; Applied sciences; Exact sciences and technology; Image processing; Image segmentation; Information, signal and communications theory; Iterative methods; Magnetorheological fluids; Markov random fields; Mathematical models; Maximum likelihood detection; Maximum likelihood estimation; Noise; Parameter estimation; Reverberation; Segmentation; Signal processing; Sonar; Sonar applications; Studies; Telecommunications and information theory; Performance evaluation; Acoustic reverberation; Energy analysis; Parameter estimation; High resolution; Statistical analysis; Segmentation; Image processing; Prior distribution; Sonar; Iterative method; Markov model; Algorithm; Noisy image; Random field; Experimental result; Least squares method; Ising model; Maximum likelihood; Hierarchical system; Comparative study
• ISSN: 1057-7149; 1941-0042
• DOI: 10.1109/83.847834

This paper is concerned with hierarchical Markov random field (MRP) models and their application to sonar image segmentation. We present an original hierarchical segmentation procedure devoted to images given by a high-resolution sonar. The sonar image is segmented into two kinds of regions: shadow (corresponding to a lack of acoustic reverberation behind each object lying on the sea-bed) and sea-bottom reverberation. The proposed unsupervised scheme takes into account the variety of the laws in the distribution mixture of a sonar image, and it estimates both the parameters of noise distributions and the parameters of the Markovian prior. For the estimation step, we use an iterative technique which combines a maximum likelihood approach (for noise model parameters) with a least-squares method (for MRF-based prior). In order to model more precisely the local and global characteristics of image content at different scales, we introduce a hierarchical model involving a pyramidal label field. It combines coarse-to-fine causal interactions with a spatial neighborhood structure. This new method of segmentation, called the scale causal multigrid (SCM) algorithm, has been successfully applied to real sonar images and seems to be well suited to the segmentation of very noisy images. The experiments reported in this paper demonstrate that the discussed method performs better than other hierarchical schemes for sonar image segmentation.