摘要
针对多层显微图像的拼接,提出一种基于3D拓扑图的图像拼接方法.利用基于区域的方法和基于特征的方法,分别对同层和异层图像对准;利用3D拓扑图的修正以消除空白区的影响;根据修正的3D拓扑图和多层图像投影坐标的误差建立多层图像的全局对准模型;利用非线性最优化方法,得到具有全局一致性的多层图像拼接结果.
This paper describes an automatic image mosaicing method for multilayer microscopic images based on 3D topology graph. The method first aligns the images at same layers with area-based alignment method, and aligns the images at different layers with feature-based alignment method, and eliminates the effect of blank areas with 3D topology graph modification. Then it constructs the global image alignment model according to the modified 3D topology graph and the misalignment of multilayer image pairs. The paper gets globally consistent estimation of multilayer image alignment parameters with nonlinear optimization method, and succeeds in achieving automatic mosaicing of large-scale microscopic images.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2006年第7期959-964,共6页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(60272042)
关键词
图像拼接
图像对准
多层显微图像
3D拓扑图
image mosaicing
image alignment
multilayer microscopic images
3D topology graph
