Seabed Image texture Clustering Using Parallel Computing

Seabed image clustering is the one of the most important applications in sonar imaging. Due to textured appearance of sonar images, texture analysis methods are a common choice for seabed acoustic images. The subsampled contourlet transform is a new version of the contourlet transform that allows analysis of images in square-size coefficients with various resolution levels and directions. This paper uses subsampled contourlet transform for clustering seabed image texture. Due to complexity of clustering algorithm and the need of some sonar applications to run algorithm faster, this paper presents the implementation of seabed image clustering using parallel computing. In this paper seabed image clustering algorithm is done using multi core programming to improve efficiency. In this paper results are executed on single core and multiple core of CPU on different sets of seabed image texture and the performance analysis shows improvement in terms of speedup and execution time.

Keywords:

Contourlet transform , Sonar , Image Clustering , Parallel computing , Multi core computing