A Novel Approach to Image Moment Computation for Sonar Image Registration in Underwater Vehicle Autonomy

Articles in Press

Document Type : Original Research Article

Author

Department of Electrical and Computer Engineering, Graduate University of Advanced Technology, Kerman, Iran

Abstract

Sonar image registration poses challenges due to its unique characteristics and real-time application demands. With lower resolution and higher noise levels compared to optical images, sonar imagery lacks distinct features, complicating traditional detection methods. Computational complexity further escalates with higher-order moments and larger image sizes. In real-time contexts like underwater vehicle autonomy, rapid and efficient moment computation is vital for timely decision-making and navigation. Overcoming these obstacles requires innovative hardware structures and parallel processing techniques. In our previous study, we introduced a systolic array and pipeline structure for high-order moment computation in grayscale images. Expanding on this, our current paper introduces the Parallel and Comparator-based Structure (PCS) to accelerate moment calculation for sonar image registration. PCS employs parallel computation and a comparator for efficiency. Through strategic adder relocation and integration of power cores, compressors, and adder units, PCS achieves streamlined computation with fixed latency. Experimental results demonstrate significant speed enhancements, positioning PCS as a promising solution for real-time underwater image processing.

Keywords

Main Subjects

International Journal Of Coastal, Offshore And Environmental Engineering(ijcoe)

Articles in Press, Accepted Manuscript
Available Online from 02 March 2024

Files

Share

How to cite

Statistics