摘要
数字图像相关方法广泛应用于多种光测力学场景,具有全场非接触等优点。数字图像相关算法的性能对光测力系统的测量精度起着决定性作用,高解算精度和效率一直是数字图像相关算法设计努力的方向对现有亚像素算法的比较分析具有重要意义。在分析数字图像相关算法原理和4种常用亚像素算法的基础上,通过数值模拟的散斑图对4种亚像素算法的计算精度和计算效率进行了分析比较,并通过实际采集的散斑图对4种亚像素算法的性能进行了比较验证。结果显示,牛顿—拉普森迭代法的解算精度和稳定性最高,在仿真实验中的解算精度达到0.001 pixel,在实际平移实验中的解算精度达到0.04 pixel;同样条件下,基于相关系数的梯度法的计算效率最高。
Digital image correlation method is widely used in a variety of applications optical measurement mechanics, the performance of digital image correlation algorithm is crucial to the measurement accuracy of optical measurement system, and high algorithm accuracy and efficiency is always the pursuit of digital image correlation algorithm design. It is of meaningful to make a thorough analysis on the existing sub-pixel algorithms. On the basis of analyzing the principle of digital image correlation algorithm and four commonly used sub-pixel algorithms, the calculation accuracy and efficiency of four sub-pixel algorithms are analyzed and compared through the numerically simulated speckle pattern, and the performances of the four sub-pixel algorithms are compared and verified through the speckle patterns collected in practice. The results show that the Newton-Rapson iterative method has the highest solution accuracy and stability, the solution accuracy reaches 0.001 pixels in the simulation experiment and 0.04 pixels in the actual translation experiment. Under the same conditions, the gradient method based on correlation coefficient has the highest solution efficiency.
作者
李开平
蔡萍
Li Kaiping;Cai Ping(Department of Instrument Science and Engineering,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2020年第8期180-187,共8页
Chinese Journal of Scientific Instrument
基金
国家重点研发计划(2016YFF0200700)项目资助
关键词
数字图像相关
亚像素
牛顿—拉普森迭代法
梯度法
曲面拟合法
digital image correlation
subpixel
newton-rapson iterative method
gradient method
curved surface fitting method
