基于SURF和互信息的图像配准方法研究
Image Registration Based on SURF and Mutual Information
-
摘要: 将SURF算法和互信息结合起来.首先利用SURF算法分别提取两幅图像的特征点，接着计算两幅图像特征点集间的互信息，最后通过改进的同步扰动随机逼近算法（SPSA）获得最佳配准参数，使互信息达到最大值，从而实现图像配准.实验表明该方法计算简单，配准速度快并且对光照和噪声具有较强的鲁棒性.Abstract: This paper,the feature points of two images respectively are extracted by SURF.Then the mutual information between the two image feature points sets is calculated.Finally,the best registration parameters are obtained by the modified SPSA,meanwhile the mutual information gets the maximum.Experimental results indicate that the advantages of the method are rapidness and simple calculation with good robust against light and the noise.
下载:
