崔宗会,缪泓*,熊宸,张明,胡文欣.基于拟合求解数字图像相关亚像素位移算法[J].实验力学,2016,31(6):769~778 |
基于拟合求解数字图像相关亚像素位移算法 |
A New Sub-pixel Displacement Algorithm in Digital Image Correlation Based on Fitting Method |
投稿时间:2016-05-20 修订日期:2016-06-27 |
DOI:10.7520/1001-4888-16-096 |
中文关键词: 数字图像相关方法 时间序列 移动最小二乘算法 |
英文关键词:digital image correlation(DIC) temporal sequence moving least square algorithm |
基金项目:国家自然科学基金项目(No.11472267, No.11372182); 国家重点基础研究发展计划(No.2012CB937504) |
|
摘要点击次数: 1271 |
全文下载次数: 658 |
中文摘要: |
数字图像相关方法是一种非接触式的变形场测量方法。虽然目前该算法能取得较高的计算精度,但是通常存在计算量较大、计算时间较长的问题。为了提高算法的计算效率,本文提出了一种基于移动最小二乘算法拟合整像素位移求解亚像素位移的数字图像相关方法。因为采用移动最小二乘算法有效利用了局部特征,所以可以采用相对简单的多项式拟合求解复杂的周期型余弦变形,同时计算结果具有较高的精度。计算结果的平均误差为0.0238pixel,标准误差为0.0791pixel,与常用的空间相关亚像素位移计算结果精度相当。实验结果验证了本文算法的正确性和有效性。通过与等计算精度的NR算法对比分析,结果表明本文算法在亚像素位移计算阶段具有更高的计算效率。计算效率至少提升了58.1倍,最多提升了437.8倍。 |
英文摘要: |
Digital image correlation (DIC) method is a non-contact method for measuring the deformation field. Although this method can achieve higher accuracy, but there usually are shortcomings such as large calculation amount and longer computation time. In order to improve the computation efficiency, based on moving least square algorithm, a new DIC method is proposed in this paper that fitting integer pixel displacement to calculate sub pixel displacement. Local feature is effectively utilized due to using moving least square algorithm, thus relatively simple polynomial fitting can be used to solve complex periodic cosine deformation with higher precision. The average error of calculation result is 0.0238 pixels, the standard error is 0.0791 pixels, which is equivalent to usual spatial correlation sub-pixel displacement calculation results. Experimental results verify the correctness and effectiveness of this method. This method has a higher computational efficiency than NR algorithm in solving sub pixel displacement calculation. The computational efficiency has been increased by at least 58.1 times, up to 437.8 times. |
查看全文 下载PDF阅读器 |
关闭 |