加权相位一致性改进的图动态精确配准算法

目的为了解决当前图像配准算法匹配精度较低的问题。方法提出加权相位一致性耦合改进的图变换匹配的精准动态图像配准算法。首先,基于SIFT机制,检测图像中的关键点;并嵌入加权因子,定义相位一致性特征,对关键点完成提纯,消除误配点与稳定性不佳的特征点;随后,设计角度距离,替代相邻特征,改进图变换匹配技术,形成精准匹配关系集;再利用初始匹配特征点与精准匹配特征点间的映射关系,对其完成修正;最后,利用改进的图变换匹配算法处理修正后的匹配关系,进一步提高匹配精度。结果仿真结果显示,与当前图像配准技术相比,改进的算法拥有更强的鲁棒性与更高的匹配精度。结论改进的算法能够提高图像在几何变换程度较大时的匹配精度。

Precise Dynamic Image Registration Algorithm Based on Weighted Phase Coherence Coupled Modified Image Transformation Matching

In order to solve the defects such as low matching accuracy induced by lots of wrong matching when facing the big geometric transformation in current image registration algorithm, the precise dynamic image registration algorithm based on weighted phase coherence coupled modified image transformation matching was proposed in this paper. Firstly, the key points of image were detected by SIFT mechanism, then the weighted factor was embedded to define the phase consistency for purification of the key points to eliminate the wrong features as well as instable features points. The angular distance was designed to substitute the neighbor features for improving the image transformation matching technology. Then the mapping relationship between initial matching feature points and accuracy matching feature points was used to correct the initial matching. Finally, the corrected matching relationship was treated by the improved image transformation matching technology. The simulation results showed that this algorithm had stronger robustness and higher matching accuracy. In conclusion, the improved algorithm could enhance the matching accuracy of images at relatively large geometric transformation degree.