Monte Carlo Simulation And Denoising Of White Light Interference Data On The Dimple Fracture Surface

As the characteristic information of the dimple fracture surface micro-topography has important value for the study of the fracture mechanism,people pay more and more attention to the measurement and quantitative analysis of the surface micro-topography.In the scanning white-light interferometry measurement,poor physical condition on the dimple fracture surface leads to extremely low optical reflectivity and its nonuniform distribution.As result,the interference signal is extremely weak and correspondingly highlights the severity of influences from the noises on the measurement of the surface micro-topography.Because of the photoelectric devices,photoelectric conversion process and environmental factors,the white-light interference signals collected in the experiment contain a variety of complex random noises,resulting in a large reconstruction error.Hence,the pre-processing of the interference signal appears to be particularly important.Firstly,this paper designs a joint control system for PZT scanning and CCD image acquisition using MFC programming technique.Secondly,by analyzing the statistical characteristic parameters,the Monte Carlo method is used to simulate the dimple fracture surface white light interference detection signal,and a priori model for two-dimensional interference data is obtained.Thirdly,this paper proposes a new adaptive wavelet threshold denoising algorithm based on bayesian estimation.In combination with the Laplacian noise model,we extract the signal variance with noise ?y2 from the measured signal.We find the signal variance without noise ?x2 according to the gray probability density distribution function of the priori model getting from Monte Carlo simulation.Then,the noise variance ?2 is obtained.Based on these,we calculate the wavelet soft threshold T,by which the original series of two-dimensional white-light interference data can be denoised.Meanwhile,we use the improved spatial frequency domain algorithm to reconstruct the three-dimensional micro-topography of the dimple fracture surface,and then we use the two-dimensional profile method to analyze and compare with the original micro-topography date of the dimple fracture surface.The experimental results show that the joint control system designed in this paper can realize the automatic acquisition of the interference images and significantly improve the collection efficiency and quality of the data.The fracture interference simulation data obtained by Monte Carlo simulation are in accordance with theLaplace distribution,at the location parameter ? of 118.84,the scale parameter ? is 1.219.The denoising method of adaptive wavelet threshold based on bayesian estimation not only improves the visual effect of the series of two-dimensional white light interference image,but also achieves 48.4891 dB increasing in the peak signal-noise ratio,retaining more details and edge information of the white-light interference image.