
十主题:检测技术及应用 Detection Technology and Application
基于非线性光流方程的数字图像相关非迭代灰度
梯度算法*
刘小勇李荣丽郝兆朋王晖崔高健范依航官岩
(长春工业大学机电工程学院,吉林长春130012)
摘要:在传统的非选代灰度梯度数字图像相关算法中,假设变形前后同一个像素点的灰度值相等。但在实
际应用中,图像的采集会受到环境光照的变化、相机曝光过度/不足、加载引起被测表面图像变化等因素的影响而引起变形前后图像像素点灰度值的变化。针对这一问题,应用非线性光流方程来描述变形前后像素点间的灰度关系,从而提出一种基于非线性光流方程的非迭代灰度梯度算法。应用模拟实验与真实单轴拉伸实验对算法的有效性与计算精度进行验证。结果表明,该算法对图像的灰度变化不敏感,具有较高的测量精度。
关键词:数字图像相关;非线性光流方程;非选代灰度梯度算法
中图分类号:0348
文献标识码:A
D0I:10.19287/j.cnki.10052402.2016.11.005
Non-iterative gray-gradient algorithmbased on nonlinear optical flow
equation indigital imagecorrelation
LIU Xiaoyong,LI Rongli, HAO Zhaopeng,WANG Hui,CUI Gaojian, FAN Yihang,GONG Yan(School of Mechatronic Engineering, Changchun University of Technology, Changchun 130012, CHN)
Abstract : In traditional non iterative gray gradient algorithm of digital image correlation method, the gray inten-
sity of a physical point on object surface is assumed that it does not change before and after deformation. However, in reality, the speckle images are sometimes acquired with illumination lighting variations, o-ver/under exposure, intensity variations during loading and so on. The gray intensity of physical points in speckle images before and after deformation is changed. In this paper, a non iterative gray gradi-ent algorithm based on nonlinear optical flow equation is proposed. The nonlinear equation is employed to build the relationship between the points in the undeformed and deformed images. Both numeral ex-periment and real tensile experiment are conducted to verify the performance. The results show that the proposed algorithm displays the most robust for intensity variations and has high precision.
ee
经过近40年的发展,数字图像相关方法(简称 DIC方法,又称为数字散斑相关方法[2])已成为一种较成熟的非接触全场变形光学计量方法,产泛应用于实验力学及其他相关科学研究与应用领域。在变形测量过程中,首先利用成像设备获得被测表面变形前后的散斑图像,然后应用数字图像相关算法通过跟踪变形前后散斑图中像素点的位置变化而获得位移及应变信息。
数字图像相关方法的计算求解过程通常分为两步,即首先进行快速整像素位移求解(整像素定位),然后在整像系位移的基础上进行亚像素位移求解(3-4)。通常,对整像素位移求解的精度要求不高,只需达到1像素精度即可,且求解算法较简单易于实现。因此,国内外研究人员对DIC的亚像素位移求解算法开展了大量的研究工作,并取得了丰硕的成果。逐渐形成了以Newton-Raphson算法(简称N-R算
*国家自然科学基金项目(51505038);吉林省教育厅"十二五"科学技术研究项目(吉教科合字[2015]第88号)
· 34. 万方数据
到更教车与和 2016年第11期