A Parallel Implementation of The Rotational Compensator Phase Unwrapping Algorithm

Abstract

Phase unwrapping algorithms are essential in image processing for optical measurements. Many research fields used phase unwrapping algorithms to recover phase data in a 2D phase data map. The rotational compensator phase unwrapping algorithm (RC) has high accuracy compared to other algorithms. However, RC takes a long time to execute. In this paper, parallel processing techniques such as SIMD and multithreading are exploited on multicore processors to accelerate the RC algorithm. Exploiting SIMD instructions inside multiple threads on a machine with Intel Xeon gold 6140 processor accelerates the RC algorithm by 3.11. Moreover, implementing a mathematical approximation of arctangent beside the parallel RC algorithm increases the speedup by 4.36 times over the original algorithm.