摘要
针对基于SIFT(Scale-invariant feature transform)算法的特征检测与匹配存在错误匹配的情况,采用了随机抽样一致性算法对错配点对进行剔除,通过有针对性地对待所有输入数据,创建一个目标函数,使用迭代算法估计出该目标函数的参数,依据所估计出的参数把所有的数据分为两种,保留满足估计参数的所有点,剔除不满足估计参数的所有点,再在全部的满足估计参数的点中重新估计出函数的参数。实验结果表明,该方法可有效剔除三维模型中的误匹配点。
For the case of mismatching in feature detection and matching based on SIFT algorithm, a method based on RANSAC is proposed in this paper to eliminate false matches. RANSAC treats all input data differently. An objective function is created and its parameters are estimated by iterative algorithm. All the data are divided into two categories based on the estimated parameters and the points that satisfy the estimated parameters are reserved and the points which do not satisfy the estimated parameters are eliminated. And then the function parameters are re-estimated with all the satisfied points. Experiment results show that this method can effectively eliminate the mismatching points in the 3D model.
出处
《系统仿真学报》
CAS
CSCD
北大核心
2017年第11期2644-2648,共5页
Journal of System Simulation
基金
国家自然科学基金(61262070
61462097)
