摘要
分析了在基于互信息方法的图像配准中,目标函数产生局部极值的原因,提出以模拟退火单纯形法作为优化策略,该方法利用了单纯形法的一种修改后的形式作为模拟退火中随机变化的发生器.此外还添加了新的边界约束处理环节,最后将本方法同Powell法和边界约束采用撞壁法时的模拟退火单纯形法的实验结果做了对比,表明本方法具有更好的精确性和稳定性.
The cause of local maximum of object functions image registration was analyzed based on mutual information, and an optimization strategy by using simulated annealing-simplex method proposed. A modified form of simplex method be the generator of random changes in the annealing process was used; and a new stage to deal with edge conditions was added. The results of our algorithm was compared with that of Powell method and that of simulated annealing-simplex method, using impingement method to be the edge constraint, showing better precision and stability in our method.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2005年第6期67-69,共3页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
总装备部预研基金资助项目(51401020204JW05).
关键词
图像配准
互信息
局部极值
模拟退火-单纯形法
image registration
mutual information
local maximum
simulated annealing-simplex method
