Device Design In Electronic Speckle Pattern Interferometry System And Image Filtering Processing Based On Kernel Regression

Electronic speckle pattern interferometry(ESPI) is an effective combination of optics and applied mechanics, and it is an important method to realize precise measurement. With its unique advantages of high accuracy, high sensitivity, non-contact, real-time and fast, it is widely used in the measurement of three-dimensional shape and deformation of objects. After several decades of development, the device and the optical path of the electronic speckle pattern interference system in the experiment have been improved and perfected continuously. In this paper, a kind of aspheric beam expander is designed, which can be used to realize the asymmetrical beam expanding and has important application in the measurement of the non-symmetrical object.Filtering noise reduction is an important part in data processing of speckle interference image. In the experiment, speckle interference fringe image can be directly obtained by CCD camera. By using digital image processing technique, the phase of deformation can be obtained and the phase information of each point in the fringe pattern can be solved.Also, the phase information of the whole field can be obtained. Thus, the whole field phase is obtained and the deformation measurement is realized. However, due to the external environment, the sensitivity of electronic devices and other factors, there is a lot of noise in the speckle pattern image, which is not conducive to the accurate extraction of phase information, so speckle image de-noising and filtering processing is still the focus of the field of research. In this paper, the kernel regression method is used to reduce the noise image, and the method is applied to the speckle image de-noising, and the ideal results are obtained. The research contents of this paper mainly include the following three aspects:1. The development history and present situation of the electronic speckle pattern interferometry technology is introduced. The basic principle of electronic speckle pattern interferometry and Fourier solution phase is derived. The theoretical conditions for the realization of shearing interference by the dislocation square prism are analyzed.2. The basic principle of the non-spherical beam expanding mirror is introduced, and the structure parameters are obtained by constructing the mathematical model. The simulation experiment was carried out, and the characteristic of the beam spot was analyzed. The results show that the beam expanding mirror can realize the expansion of the aspheric surface, and the elliptical spot can be obtained after the beam expanding.3. The basic principle of kernel regression is introduced. Using this method, simulation of the noise reduction treatment performed on color and gray image and also the speckle image that was recorded in the experiment. The root of mean square error between the image after the noise reduction and the original image is obtained, and the phase of speckle image is solved. The results show that the kernel regression can effectively reduce the image noise, and has a good image smoothing effect.