摘要
数字体积相关是通过分析具有相关关系的两组三维图像,获得物体变形过程中位移场和应变场的计算方法。了解该算法在实际噪音下的表现,对这种方法的应用具有十分重要的意义。利用爬山法、最小二乘法,结合三次B样条插值,建立了一套数字体积相关计算方法,用两组具有真实噪音的三维X射线CT数据对该算法进行了系统评测。在没有噪音的理想情况下,该算法的位移误差为0.0065voxel;在真实X-CT噪音的情况下,位移误差增加到0.008voxel,应变误差约为100με。结果表明,本方法可对真实物体变形进行三维位移场、应变场分析。
Digital Volume Correlation（DVC）is one calculation method to obtain the displacement field and stain field in an object deformation process through analyzing two groups of 3D correlated images.Understanding the performance of this algorithm under real noise condition is extremely important for its application.A DVC algorithm was developed based on Newton-downhill method,least square method and tricubic B-spline interpolation.The performance evaluation of this DVC algorithm was carried out by using two groups of 3D X-CT images under real noise condition.In the ideal case without noise,the displacement error of this algorithm is 0.0065 voxel;while in the case of X-CT real noise,the displacement error increases to 0.008 voxel.The stain error is about 100με from the calculation of repeat unloaded X-CT data.Above results show that DVC algorithm is suitable for calculation of displacement field and stain field in real object deformation.
出处
《实验力学》
CSCD
北大核心
2014年第5期589-600,共12页
Journal of Experimental Mechanics
基金
国家自然科学基金(51008072)
中央高校基本科研业务费专项资金资助(No.2242014R30014)
