Improved Search Algorithm of Digital Speckle Pattern Based on PSO and IC-GN

Qiang Chen,Liang Hong,Youtian Qu,Zhixin Tie,Dengwen Wang

doi:10.3390/photonics9030167

Qiang Chen, Liang Hong + Show 3 more

Open Access

https://doi.org/10.3390/photonics9030167

Copy DOI

Abstract

Digital speckle correlation method has not only been widely used in a variety of photometric mechanical scenarios, but also integrated with multiple disciplines. In the future, it will even be inextricably linked to the Internet of Things, autonomous driving, deep learning and other fields. For a given hardware condition, it is of great significance to improve the efficiency of integer-pixel search and increase the accuracy and efficiency of the sub-pixel algorithm. In this paper, we propose an improved digital speckle correlation method, which consists of an integer-pixel search algorithm and a sub-pixel search algorithm. With respect to the integer-pixel search, aiming to address the two problems of uniqueness of maximum value and parameter setting of PSO-W algorithm, the algorithm PSO-1 is proposed, and the results of comparison experiments show that it has higher search efficiency. In terms of sub-pixels, based on IC-GN algorithm with the highest accuracy at present, the IV-ICGN algorithm is proposed, and the simulation experiment results show that the proposed algorithm has higher accuracy and higher efficiency than the comparison algorithm.

Highlights

Digital image correlation method (DICM) uses the speckle images of an object before and after deformation to obtain the displacement information of the object by matching the positions of the most similar subsets of these scattered images according to the correlation function
High precision and high efficiency have always been what digital speckle correlation method (DSCM) strives for and pursues; this paper proposes an improved DSCM
Algorithm, which is composed of an efficient particle swarm optimization (PSO)-1 integer-pixel search algorithm and a higher-precision IV-ICGN sub-pixel search algorithm

Summary

Introduction

Digital image correlation method (DICM) uses the speckle images of an object before and after deformation to obtain the displacement information of the object by matching the positions of the most similar subsets of these scattered images according to the correlation function. Jiang et al [31] proposed a cross-correlation algorithm based on fast Fourier transform, which improved the searching efficiency and robustness without sacrificing accuracy. The GA-based algorithm [29] is fast, it is easy to converge prematurely and fall into a local optimal solution; only when the number of iterations is large enough, can high accuracy be achieved. PSO-based algorithm [30] has the advantages of good memory, fast convergence, strong global search ability and high accuracy [34,35]; Wu et al [36] improved on the standard PSO algorithm to achieve the purpose of fast integer-pixel search. He did not consider the case of multiple maxima To address these two points, an efficient integer-pixel search algorithm PSO-1 based on PSO is proposed in this paper, which has a faster solving speed than the algorithm proposed by Wu

Objectives

Findings

Discussion

Conclusion