IBR (image-based rendering) is an important technology in VR (virtual reality). Panoramic image stitching used to create virtual environment for many applications is a key technology for IBR, and lots of stitching algorithms are developed in recent years. In this paper, we introduce an algorithm based on SIFT features to stitch panoramic images. At last, some improvements are made that we stitch the two images by using the coordinates of the image after transformation instead of interpolation; in this way complex algorithm and unnecessary calculation are avoided. Through lots of analyses and experiments we find that it is very effective. Because this method is not sensitive to ordering, orientation, scaling and illumination, its stitching precision is much higher than many other methods.
[1]
Lizhuang Ma,et al.
A Fast and Robust Image Stitching Algorithm
,
2006,
2006 6th World Congress on Intelligent Control and Automation.
[2]
G LoweDavid,et al.
Distinctive Image Features from Scale-Invariant Keypoints
,
2004
.
[3]
Richard Szeliski,et al.
Construction of Panoramic Image Mosaics with Global and Local Alignment
,
2001
.
[4]
H. Opower.
Multiple view geometry in computer vision
,
2002
.
[5]
Heung-Yeung Shum,et al.
Panoramic Image Mosaics
,
1998
.
[6]
Richard Szeliski,et al.
Systems and Experiment Paper: Construction of Panoramic Image Mosaics with Global and Local Alignment
,
2000,
International Journal of Computer Vision.