预测搜索算法在图像相关中的应用  

An improvement of digital speckle correlation method using the predictive search algorithm

作　　者：王昊[1] 马志峰[1] 

机构地区：[1]北京理工大学信息与电子学院,北京100081

出　　处：《光学技术》2013年第3期251-255,共5页Optical Technique

摘　　要：数字散斑相关方法在亚像素测量过程中运算量大,在对实时性要求较高的系统中该方法的应用受到了限制。提出了一种原理简单、搜索速度快、精度高的预测搜索算法,介绍了预测搜索算法的原理,并给出预测位移的公式。通过实验验证了对算法的精度、效率和鲁棒性,说明对数字散斑相关方法进行优化后,既不损失亚像素位移的计算精度又提高了图像处理的效率。The discrepancy of precision and rate is still an obvious drawback of the Digital Speckle Correlation Method （DSCM）, using a single camera, though it has been widely used for measuring the full-field displacement and surface strain field. What limits the use of DSCM in real-time system is its tremendous computational complexity. The predictive search algorithm presented can effectively estimate the magnitude and direction of displacement vectors and decrease the amount of calculation in a sense. In the way, the predictive search algorithm sharply speeds up the efficiency of DSCM without any loss of measurement accuracy. The capability of proposed search algorithm is verified by experiments. The results show that the algorithm can improve the efficiency of correlation method and also satisfy the need of measuring in- plane displacements in accuracy.

关 键 词：机器视觉 预测搜索算法 数字散斑相关方法 数字图像相关 亚像素 

分 类 号：TN911.74[电子电信—通信与信息系统] V556.5[电子电信—信息与通信工程]