Study On The Method Of Extracting Information Of Electronic Speckle Pattern Interferometry Based On The Fourier Transform And Its Application

The Electronic Speckle Pattern Interferometry (ESPI) is a non-destructive, whole-field optical measurement technique, which is applied widely in optical rough surface displacement measurement, strain analysis, non-destructive technology (NDT) and vibration measurement. Therefore, the key of ESPI application is to extract information of ESPI correctly. In this paper, we make a research on the method of extracting information of ESPI based on the Fourier Transform (FT).The primary coverage of this paper includes four parts: (1) A method based on FT is proposed to estimate the orientation of ESPI, which is compared with the Gradient based Method and Accumulate Differences Method. (2) We proposed a method to calculate the frequency of ESPI based on FT. As the basic information of ESPI, orientation and frequency are not only the emphasis of the paper, but also can direct fringe image processing. In particular, the direction affects the accuracy of the fringe analysis critically. (3) The angular filter and the density filter are derived by the angle and density of the speckle fringes. And then the noise reduction achieved by multiplying the FT of the image by the two filters. We utilize orientation coherence measure to adapt the angular bandwidth and characteristic of the directional filter. Density filter is a simple 2D band-pass Butterworth filter, which used to reduce noise. Combining these two filters can suppress noise effectively. The proposed filters have better result when adding root filtering and threshold. (4) We experimentally compare the proposed approach to other filtering approaches based on FT in literature via application to a variety of test cases, which are obtained by computer simulation or experiment. The experimental results show that our technique performs favorably and has high practical value.Three advantages are mentioned about extraction information of ESPI based on FT. Firstly, the algorithm keep simple and applicative, and is proved effective for most of the patterns, especially for dense poor quality patterns. Secondly, it's convenient to make use of the algorithm since there's only two variable: window size and the threshold. Thirdly it can estimate the intrinsic properties of the patterns such as the local ridge orientation and local ridge frequency.