欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
在图匹配问题中基于松弛迭代的方法能否收敛到全局最优解在很大程度上依赖于初始值的估计,针对这个问题,提出了一种结合亮度序局部特征描述的图匹配算法。该算法首先利用Hessian-Affine方法提取图像的特征点及局部特征区域,以特征点作为图的节点并结合特征点的邻近关系构造结构图;其次,根据亮度序约束关系对局部特征区域进行子区域划分,利用改进的中心对称局部二值模式（CS-LBP）获取局部特征描述;最后,将局部特征描述之间的相似性作为图匹配关系矩阵的初始值,通过松弛迭代的方法获取特征点的准确匹配结果。实验结果表明该算法匹配准确率较高。
英文摘要：
In the graph matching problem,whether the method based on relaxation iteration can converge to a global optimal solution depends on the initial estimate to a great extent. Therefore,a graph matching algorithm combined with brightness order local feature description is presented. Firstly,feature points and local feature regions are extracted by Hessian-Affine. The structural graph is obtained by using each feature point as a node and combining with adjacency relationship of feature points. Secondly,each local feature region is partitioned into sub regions using the constraint of brightness order. Then the improved center-symmetric local binary pattern（ CS-LBP） is used to describe the local feature. Finally,the similarity of local feature description is used to initialize the matching of the graphs,and after relaxation iteration,the exact matching of feature points is achieved. Experimental results showed that the algorithm has high matching accuracy.
同期刊论文项目
同项目期刊论文
Remote sensing inversion of leaf area index based on support vector machine regression in winter whe
期刊信息
位置：