Nonlocal means variants filtering methods for speckle noise reduction in digital speckle pattern interferometric fringes

摘要

Nonlocal means (NLM) and its variant filtering methods such as nonlocal means-average (NLM-av), nonlocal meanslocalpolynomial regression (NLM-LPR), and nonlocal means-shape adaptive patches (NLM-SAP) for speckle noisereduction in digital speckle pattern interferometric (DSPI) fringes are presented. The performance of these filteringmethods is appraised by several criteria such as peak signal-to-noise ratio (PSNR), the quality index (Q), and the edgepreservation index (EPI). The obtained filtering results corroborate the effectiveness of NLM and its variant filteringmethods for speckle noise reduction in DSPI fringes.