摘要
为了提高实时宽场景图像拼接应用中的图像配准精度和配准速度,提出了一种利用相机标定信息与相位相关技术相结合的图像配准方法。利用相机间的相对位置参数计算待拼接图像间的旋转和缩放模型,利用相位相关技术求解图像间的平移转换参数。通过理论推导对缩放模型的误差分布进行分析。实验结果证明,该算法理论推导正确,在场景深度范围较小情况下,图像拼接准确快速。
An image matching algorithm based on calibration information of cameras and phase correlation technology is de- signed to improve the matching accuracy and speed in wide scene images mosaic application. Relative position parameters between cameras are used to deduce the rotation amt the scaling model, and the phase correlation is selected to calculate the translation parameters. The error distribution of scaling model is analyzed by theory deducing. The experiment results proves the exactness of theory deducing and the validity in image mosaic application under limited effective scene depth instance.
出处
《微型机与应用》
2012年第18期38-41,共4页
Microcomputer & Its Applications
基金
中国人民大学科学研究基金(中央高校基本科研业务费专项资金资助)(10XNF039)
关键词
图像配准
图像拼接
相位相关
image matching
image mosaic
phase correlation
