基于形态学聚类算法图像配准仿真研究

研究图像配准精确度问题。由于两张图片几何关系及量度均有不同,要达到配准效果应有空间一致性。传统的聚类图像配准算法进行图像配准时,配准精度较低,算法复杂度高等不足。为了有效提高图像配准的精确度,提出了一种改进的数学形态学和聚类算法相结合的图像配准方法。算法首先改进的基于空间模式均值聚类对图像进行区域分块,并对分块的位置进行空间聚类,并准确计算出基准图像的最后的配准位置,并采用数学形态学方法对配准后的图像进行边缘处理,最后评估配准图像的质量。仿真结果表明,提出的改进的算法有效的提高了配准精确度,是一种可行性有效的图像配准算法,为图像配准提供了依据。

Image Registration Simulation Based on Clustering Algorithm and Morphological

Research the problem of image registration accuracy.The traditional image registration algorithms are based on clustering images with time,which lowers the registration accuracy and lack of high complexity.In order to improve the accuracy of image registration,an improved clustering based on mathematical morphology and algorithm was a combination of image registration methods.Firstly,the improved algorithm partitioned regional sub-image blocks based on means clustering of spatial patterns,and clustered the spatial,locations of these blocks.The accurate registration locations were computed,and then mathematical morphology was used in the edge treatment after registration.At last,the quality of image registration was assessed.Simulation results show that the improved algorithm improves the accuracy of registration effectively and the feasibility of the method is verified to be an efficient image registration algorithm.