摘要
模板匹配是图像配准的重要步骤，已有许多方法被提出，但如何进一步快速而准确地找到匹配点仍是一个问题。由于相似性度量值在每个匹配位置是随机的，并在匹配点取得全局最优值，所以可以使用模拟退火的思想来加速搜索过程。本文对此进行了研究和探讨，并通过实验与原有算法加以比较，取得了满意的效果。
Template match is an important step of image registration,in this field,many approachs have been presented,but how to find match point more quickly and correctly is still a problem. Because the value of similarity measure of two images in each match place is random,and become the global optimal in match point,so can apply improved simulated annealing algorithm to speed-up searching. This paper does some research in that,and compares with original algorithm by experiments.These experiments show this approach can achieve an ideal result.
出处
《小型微型计算机系统》
CSCD
北大核心
1997年第8期32-37,共6页
Journal of Chinese Computer Systems
关键词
模板匹配
模拟退火
相似性度量
计算机图形学
Template match, Simulated annealing, Similarity measure, Match point
