Research On Key Techniques Of Digital Speckle Pattern Interferometry Phase Information Process

Digital speckle pattern interferometry(DSPI)uses optical interference principle to measure the mechanical parameters such as stress,strain,defect and vibration mode of the material,which has the advantages of high precision,full field,non-contact and so on,and is widely used in aerospace,biomedical,precision manufacturing and other fields.In the process of DSPI,the phase distribution is related to deformation,defect and displacement of the object surface.Therefore,phase information processing is the key research content of DSPI.The conventional phase information processing methods require high working conditions,which restricts the practical application of DSPI.For the requirements of dynamic measurement,full field wide range and multi-dimensional deformation synchronous detection in aerospace applications,some methods in DSPI are proposed in the aspects of phase extraction,phase map filtering and phase unwrapping.The main innovative work is as follows:1.A single digital speckle image phase extraction method is proposed.First of all the digital speckle image is decomposed adaptively using empirical wavelet transform and variational mode decomposition methods.Then,the fringe information is processed by kernel probability density and histogram of the decomposed component sequentially.Finally,the phase is calculated by Hilbert method and the high-order fuzzy function method.Compared with traditional methods,experimental result show that the proposed method can extract the speckle phase and improve the accuracy of speckle measurement.2.An adaptive sine-cosinse phase filtering method is proposed.In our method,the noise energy is selected as the evaluation standard.By evaluating the change of the noise energy of the phase map before and after the filtering process,the appropriate number of filtering cycles is automatically determined.The filtering effect is enhanced and the noise reduction of the speckle phase map is realized consequently.The experimental results show that the adaptive sine-cosine filtering method can automatically filter out the noise in the speckle phase map.Compared with the traditional filtering method(median filtering method,mean filtering method),the filtered phase map speckle index is smaller and the peak signal to noise ratio is larger.3.An improved variational mode decomposition and DSPI phase de-noising method are proposed.The DSPI phase map is decomposed by the improved variational mode decomposition,the noise threshold of the decomposed component is calculated,and the de-noised components of phase map is reconstructed,and the phase map after noise reduction is obtained.The experimental results show that the peak signal-to-noise ratio(PSNR)of the map after noise reduction is 12.46 d B,which is better than the variational mode decomposition method(peak signal-to-noise is10.66 d B).4.An improved row-by-point method is proposed.Using the speckle index and peak signal-to-noise ratio to sort the different areas of parcel phase,the phase region of different quality is expanded by the row point-by-point method and the least-squares method.The phase after the expansion of the different regions is reconstructed to obtain the true phase distribution.The experimental results show that the proposed method can expand the parcel phase and improve the accuracy of unwrapping.