Denoising for variable density ESPI fringes in nondestructive testing by an adaptive multiscale morphological filter...

Area of Science:

• Optical Metrology
• Materials Science
• Image Processing

Background:

• Electronic Speckle Pattern Interferometry (ESPI) is crucial for nondestructive testing.
• Analyzing ESPI images with variable fringe density presents significant noise reduction challenges.
• Accurate defect characterization relies on preserving fringe information.

Purpose of the Study:

• To introduce a novel adaptive multiscale morphological filter for ESPI images.
• To improve noise reduction and edge preservation in variable density fringe analysis.
• To enhance the reliability of defect detection in nondestructive testing applications.

Main Methods:

• Image segmentation based on local mean to classify fringe densities.
• Adaptive design of structural elements tailored to different fringe density levels.
• Iterative determination of optimal structural element size for morphological filtering.
• Application of morphological open-closing filtering and edge smoothing via averaging.

Main Results:

• The proposed filter demonstrated superior performance compared to discrete cosine, wavelet, Lee, and nonlocal mean filters.
• Significant noise reduction was achieved while effectively preserving critical fringe edge details.
• The method proved effective on both simulated and experimentally obtained ESPI fringe patterns.

Conclusions:

• The adaptive multiscale morphological filter offers a robust solution for noise reduction in ESPI images with variable fringe densities.
• This technique facilitates a more quantitative correlation between fringe patterns and material defects.
• The method has broad applicability in fields like defect detection, thermal analysis, and heterogeneous material mechanics.