Abstract:

The correlation fringes obtained by traditional methods in digital speckle pattern interferometry are se- verely contaminated by speckle noise, which results in a low signal-to-noise ratio in fringes. To overcome this di- sadvantage, a novel approach to the generation of correlation fringes based on cross-entropy is presented by taking into account the concept of cross-entropy and the basic principles of fringes representation for the state change in digital speckle pattern interferometry. The quantitative results of theoretical deduction and practical applications in- dicate that the correlation fringes obtained by the proposed method are essentially coincident with the subtraction fringes, and that, due to the integration of adaptive homomorphic filtering into the generation of cross-entropy frin- ges, the proposed approach may result in higher fringe contrast, smaller speckle index and better quality of wrapped phase map.

Key words: image processing, speckle pattern interferometry, correlation fringe, cross-entropy