组合模型的图像配准

Image Registration Based on Assembly Model

为解决图像配准中因旋转变换,尺度缩放,光照变化,相机抖动,气流变化以及压缩等影响成像质量,而导致特征点检测数目不足的问题,本文提出了基于组合模型的图像配准方法.该方法采用KAZE与SURF算子联合检测局部区域中线性与非线性特征点;同时为了提高计算效率,采用二进制向量描述符表征特征点,并使用汉明距离计算特征点之间的匹配距离,有效地提高匹配效率;最后,利用随机一致性算法(RANSAC)进一步消除异常点,并根据内点之间的对应关系来计算几何变换模型.实验结果表明:本文有效地解决了因特征点数目不足配准失效的问题,多幅图像的配准实验结果说明了本文方法具有更好的稳定性与鲁棒性,同时运行效率最快.

In order to solve the problem of insufficient number of feature points detected due to rotation transformation, scale scaling, illumination changes, camera shake, airflow changes, and compression in image registration, this paper proposes an image registration method based on a assembly model.This method uses KAZE and SURF operators to jointly detect linear and non-linear feature points in the local area.At the same time, in order to improve the calculation efficiency, binary vector descriptors are used to characterize the feature points, and the Hamming distance is used to calculate the matching distance between the feature points, effectively The matching efficiency is improved;finally, the random consensus algorithm(RANSAC)is used to further eliminate the Outliers and the geometric transformation model is calculated according to the correspondence between the interior points.The experimental results show that: this paper effectively solves the problem of insufficient number of feature points.The registration experiment results of multiple images show that this method has better stability and robustness, and the fastest running efficiency.