Automatic Panorama Recognition and Stitching Based on Graph Structure

DOI

10.2991/cse.2013.58How to use a DOI?

Keywords

feature detection; feature matching; image recognition; image matching; image stitching

Abstract

An automatic panorama recognition and stitching method based on graph structure was proposed to solve the multiple images recognition and matching problem. With multiple unordered images of user input, the method can automatically finding overlapping portion between images, and stitch them. First, the MOPS feature points were detected from the input images, and using kd-tree nearest neighbor search to perform fast feature matching between images. The motion model between any two images can be established by RANSAC algorithm based on the correspondence of feature points, and robust verification by a probabilistic algorithm. The automatic recognition of multiple panoramic images problem can be solved by building undirected connected graph corresponding to image matching relationship. Finally, a recursive algorithm was used to do depth-first traversal of the established undirected connected graphs, and multi-band blending algorithm was used to eliminate stitching seam. The experiments showed that multiple unordered images can be matched and recognized automatically, and panoramas can also be stitched automatically and seamlessly.

Copyright

© 2013, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Yi-Li Zhao
AU  - Yan Xia
PY  - 2013/07
DA  - 2013/07
TI  - Automatic Panorama Recognition and Stitching Based on Graph Structure
BT  - Proceedings of the 2nd International Conference on Advances in Computer Science and Engineering (CSE 2013)
PB  - Atlantis Press
SP  - 261
EP  - 263
SN  - 1951-6851
UR  - https://doi.org/10.2991/cse.2013.58
DO  - 10.2991/cse.2013.58
ID  - Zhao2013/07
ER  -