Electronic Speckle Pattern Interferometry Image Processing

Electronic speckle pattern interferometry is a well-known technology for non-contact nondestructive measurement of optical coarse surface. With the advantages of whole-field, real-time, high precision, high sensitivity, and strong anti-interference,it has been widely applied in measuring the object deformation, displacement, vibration, medical diagnosis and so on. In order to collect information, we have to extract the phase diagram from the fringes. However, a lot of noise in the fringe pattern affects the proper extraction of the phase diagram. Therefore, removing the noise from the fringes effectively has always been an important issue in electronic speckle pattern interferometry.In this paper, after analyzing the the existing filtering methods in detail, two innovative filtering algorithms are proposed: improved spin filtering and a combined algorithm based on wave-atom threshold algorithm and improved spin filtering. The filtering results of both computer-simulated fringes and experimental fringe show that the two innovative algorithms can both achieve good performance. The improved spin filtering, based on fuzzy directional window, achieves an outstanding performance by using the directivity of fringe pattern, meanwhile; avoid the increase of algorithm's complexity and instability, which come after the accurate estimating of fringe-orientation. The combined algorithm based on wave-atom thresholding and improved spin filtering, can remove the noise completely while preserve edges very well. One iteration of improved spin filtering after the implement of wave-atom thresholding with a conservative threshold, have a shorter running time than several iterations of improved spin filtering. Furthermore, due to its few parameters and wide range of parameter selection, the combined algorithm is easy to debug. Owing to all the advantages mentioned above, the combined algorithm possesses high practical value.To further verify the effectiveness of the proposed filtering algorithms, we get the denoised phase field after four-step phase shift method and the least-squares unwrapping algorithm were applied on the denoised fringe patterns. The experimental results show that the combined method, based on wave atom threshold algorithm and improved spin filtering algorithm, lead to the most smooth phase field and the biggest phase field peak, which verify the combined method can remove the noise completely while preserve edges very well. After some calculation and implement of integral process, the practical displacement field with current order of magnitude is get, which prove the correctness of the filtering, phase extraction and phase unwraping for ESPI fringe.