基于特征点匹配的全景图像拼接技术研究

摘要

提出了一种基于特征点匹配的全景图像拼接算法，首先提取各图像中的SIFT特征，通过特征点匹配完成两幅图像的配准；再根据图像配准结果计算出图像间的变换矩阵；最后采用渐入渐出加权平均的融合方法对两幅图像进行无缝拼接。实验表明，该算法具有匹配精度高、鲁棒性强等特点，可以快速而自动地生成全景图像。%In this paper,an efficient panoramic image mosaic algorithm based on SIFT is proposed.Firstly,SIFT feature points are extracted from a sequence of images captured by ordinary camera.Then,images are registered by matching SIFT feature points.And then the transform matrix can be calculated by registration results.Finally,the weighted average method is used to achieve image mosaic.Experiments indicate that the proposed algorithm has high matching accuracy,robustness,and other characteristics,can show the cylindrical panoramic image automatically and effectively.