Study On Filtering Method Of Electronic Speckle Interference Image

In recent years,electronic speckle interference technology has gradually developed into a new type of non-contact measurement technology.The technology has lots of advantages,such as real-time,non-contact and whole-field,thus has been used widely measurement of vibrations,displacements,strains and medical diagnoses.Electron speckle interference technology can produce the speckle interference fringe image,where the object surface deformation information is hidden in.However,due to the electronic speckle interference fringe images accompanied by strong particle noise,interference fringes contrast and resolution will be relatively low,affecting the results of electronic speckle interference technology,and ultimately affect the accuracy of measurement.Therefore,this paper proposed an improved which is based on SVM-PCNN filtering method for the above-mentioned problems.According to the characteristics of multi-scale decomposition and local directionization,the advantages of SVM in small sample,nonlinear learning,high-dimensional model classification and so on are combined with the synchronous release and asynchronous oscillation characteristics of PCNN neural network,so as to achieve the purpose of filtering.The experimental results show that the algorithm has the advantages of high contrast and clear fringes.The main work of this paper includes the following aspects:1.The detailed introduction of the basic principles of electronic speckle interference technology and the development process of interference fringe image filtering technology.2.The introduction of the principle of several filtering methods as airspace filtering,frequency domain filtering and other fringe filtering methods,and the analysis of their applicable scenarios and the advantages and disadvantages of them.For their own shortcomings,we proposed an improved which is based on SVM-PCNN filtering method.3.Doing the stimulation experiments on the experimental object are carried out by using MATLAB,and discuss the SNR and the speckle index of the filtered image and other evaluation index,and then analyze the effect of fringe images which are been refined.Based on this,this paper can confirm objectively that the algorithm the author proposed can achieve better filtering effect,and which is a feasible image filtering method.