Abstract: A new real-time image mosaic method based on gray correlation and feature points was proposed by analyzing and comparing the advantages and disadvantages of different registration algorithms. Image was zoomed adaptive to the required size of need for output. The direction of registration and general location of potential were determined by the method based on feature points. Best match point was determined accurately by the method based on gray correlation. Finally the overlapping images were fused by weighted average. The experimental results show that the new algorithm can mosaic the image correctly and in real-time efficiency.

Key words: image mosaics, image registration, image fusion, gray correlation, feature correlation, panoramic image

摘要： 通过分析比较不同配准算法的优缺点，提出一种基于灰度相关与特征点相结合的实时图像拼接方法。算法首先对图像进行自适应缩放，调整为需要输出的大小，然后用基于特征边缘点方法确定配准方向和大体位置，用基于灰度相关方法准确定位最佳匹配点，最后通过加权平均法对重叠图像进行融合处理。实验结果表明，算法能准确地对图像进行拼接，具有实时性和高效性。

关键词: 图像拼接, 图像配准, 图像融合, 灰度相关, 特征相关, 全景图