主成分分析的匹配点对提纯方法
发布时间：2019-06-29 07:29
【摘要】：传统的匹配点对提纯算法通常需要寻找小部分点集作为初始输入,再迭代求解出能够满足大多数点对约束要求的最优解,其提纯结果易陷入局部极值,造成正确匹配点对的遗漏。针对这一问题,本文引入了主成分分析思想,将整体点集作为初始输入,逐步剔除误匹配点对,稳健求解,得到更为准确的全局最优解,降低正确匹配点对的遗漏率,达到较好的提纯效果。试验表明,本文方法在一定的原始误匹配率下,能够得到整体最优解,在剔除误匹配点对的同时,能够避免或减少正确匹配点对的遗漏。
[Abstract]:The traditional matching point pair purification algorithm usually needs to find a small part of the point set as the initial input, and then iteratively solve the optimal solution which can meet the constraint requirements of most point pairs. The purification results are easy to fall into local extremum, resulting in the omission of correct matching point pairs. In order to solve this problem, this paper introduces the idea of principal component analysis, takes the whole point set as the initial input, gradually eliminates the mismatched point pair, solves it steadily, obtains the more accurate global optimal solution, reduces the leakage rate of the correct matching point pair, and achieves better purification effect. The experimental results show that the proposed method can obtain the global optimal solution under a certain original mismatching rate, and can avoid or reduce the omission of the correct matching point pair while eliminating the mismatching point pairs.
【作者单位】： 信息工程大学;
【基金】：国家自然科学基金(41401534) 地理信息工程国家重点实验室开放基金(SKLGIE2013-M-3-1)~~
【分类号】：P23
本文编号：2507660
[Abstract]:The traditional matching point pair purification algorithm usually needs to find a small part of the point set as the initial input, and then iteratively solve the optimal solution which can meet the constraint requirements of most point pairs. The purification results are easy to fall into local extremum, resulting in the omission of correct matching point pairs. In order to solve this problem, this paper introduces the idea of principal component analysis, takes the whole point set as the initial input, gradually eliminates the mismatched point pair, solves it steadily, obtains the more accurate global optimal solution, reduces the leakage rate of the correct matching point pair, and achieves better purification effect. The experimental results show that the proposed method can obtain the global optimal solution under a certain original mismatching rate, and can avoid or reduce the omission of the correct matching point pair while eliminating the mismatching point pairs.
【作者单位】： 信息工程大学;
【基金】：国家自然科学基金(41401534) 地理信息工程国家重点实验室开放基金(SKLGIE2013-M-3-1)~~
【分类号】：P23
【相似文献】
相关期刊论文 前6条
1 陈卫东;王锋;刘睿;孔李军;丁伟;董卓莉;;一种基础矩阵的线性估计方法[J];河南科技;2009年07期
2 毛明;;开域场的离散分析原理[J];扬州师院学报(自然科学版);1986年01期
3 吕冀;汪渤;高洪民;周志强;;图像局部特征识别中的多目标分离[J];光子学报;2008年08期
4 甄艳;刘学军;王美珍;;一种估计基础矩阵的新鲁棒算法[J];地理与地理信息科学;2013年02期
5 甄艳;刘学军;王美珍;;一种改进RANSAC的基础矩阵估计方法[J];测绘通报;2014年04期
6 ;[J];;年期
相关会议论文 前3条
1 吴福祥;;双相机自定标中匹配点的获取[A];2009年研究生学术交流会通信与信息技术论文集[C];2009年
2 王亚伟;许廷发;王吉晖;;改进的匹配点提纯算法mRANSAC[A];2013年中国智能自动化学术会议论文集（第三分册）[C];2013年
3 朱振平;李志军;王卫华;陈曾平;;一种基于绝对差加权的相关跟踪算法[A];全国第五届信号和智能信息处理与应用学术会议专刊(第一册)[C];2011年
，本文编号：2507660
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2507660.html
