Effect of Speckle Edge Characteristics on DIC Calculation Error

Abstract

Background

In DIC studies, positional parameters and speckle size are commonly used to characterise speckle images. The influence of edge parameters is ignored. This leads to a great difference between the DIC calculation results of simulated and real images. And some contradictory results are also produced.

Objective

The main objective of this paper is to investigate the effect of edge parameters. As well as to give more reasonable parameters to describe the speckle characteristics.

Methods

Firstly, this paper proposes a series of more reasonable parameters to describe the speckle features based on the mathematical expression of the speckle image. Subsequently, the effect of different edge functions on the computational error of DIC is investigated. The effect of different edge functions on pre-filtering is also investigated. Finally, real speckle images are produced using Gaussian and step functions to study the difference between the simulated and real speckle images.

Results

Generally, it is believed that prefiltering can reduce the computational error of DIC, but for Gaussian edges, prefiltering hardly reduces the error, whereas hybrid edges correctly exhibit this phenomenon. Although the Gaussian edge perform well in the simulation, the actual speckle images taken show that the DIC error corresponding to the camera-acquired Gaussian speckle is much larger than that of the step speckle.

Conclusions

The introduction of edge parameters to describe speckle images is necessary for DIC studies. Pre-filtering always reduces the DIC error, but for Gaussian edges this property cannot be demonstrated. The most suitable edges in reality are step edges, not Gaussian edges.