基于改进的遗传算法的图像匹配研究  

Research of Image Matching Based on Improved Genetic Algorithm

作　　者：袁张露[1] 许建龙[1] 曾龙龙[1] 

机构地区：[1]浙江理工大学信息电子学院,杭州310018

出　　处：《浙江理工大学学报（自然科学版）》2011年第6期910-914,共5页Journal of Zhejiang Sci-Tech University(Natural Sciences)

摘　　要：匹配算法的计算量由搜索位置数与相关算法的计算量之积来决定,因此,为了减少总的计算量,需要改进匹配算法和减少搜索位置。提出一种改进的遗传算法的匹配方法,设计了染色体的表示方法、新的适应度函数以及新的遗传操作算子。为了加快算法的收敛速度,对初始种群的选取和遗传算子操作概率的选取提出了新的方法。实验结果表明,该方法更有效地完成了特征图像之间的匹配。The whole computational complexity of match algorithm is decided by the product of computational complexity and the amounts of searching locations relevance algorithm. Therefore, to Improve matching algorithm and reduce the search location can cut down the computational amounts. The authors propose an improved genetic algorithm, a designed the chromosome representation, the new fitness function, and new genetic operators. In order to speed up the convergence rate, the authors propose a new method on the selection of initial population and gentie operators selection operation. Experiment shows that the approach is more efficient to complete the feature match between images.

关 键 词：图像匹配 遗传算法 模板 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]