一种基于SIFT特征的序列图像拼接算法

Algorithm for Sequence Image Splicing Based on SIFT Features

针对序列图像拼接问题,提出一种基于尺度不变特征(SIFT)的自动拼接方法。该方法首先计算图像SIFT特征向量,作为匹配的依据,其次利用SIFT特征进行图像配准,进一步配准的基础上利用图像交叠处距离差来线性融合边缘,取得较好的平滑的镶嵌效果,最后通过对真实场景序列图像进行拼接,实验验证了该方法的有效性。

For sequence image splicing, propose an automatic splicing method on the basis of fixed dimension. The method first calculates images＇ SIFT eigenvector and makes it the matching basis, then makes an image registration by using SIFT features and makes a linear edge fusion using the distance difference of image overlapping section. The experiment of image sequence stitching in real scene achieves the better smooth mosaic effect and show the method is effective.