TY - GEN
T1 - Fast robust image feature matching algorithm improvement and optimization
AU - Chen, Peiyu
AU - Li, Ying
AU - Gong, Guanghong
N1 - Publisher Copyright:
© 2018 Association for Computing Machinery.
PY - 2018/8/27
Y1 - 2018/8/27
N2 - This paper quantitatively analyzes different types of image changes according to the characteristics of each algorithm, and put forward different optimal algorithms for different types of pictures. Firstly, four classical matching algorithms are selected and compared for scale, photometric and rotational robustness. In order to solve the limitation of the robustness of single algorithm, three improved algorithms are proposed. Based on the combination of SURF and ORB algorithms and one or more feature point screening, the improved algorithm is used to improve accuracy. Secondly, the improved algorithm is tested by using images with multiple types of changes at the same time. It is concluded that the improved algorithm has strong robustness and can effectively improve image matching accuracy. Finally, the simulation result shows that the selection of the optimal algorithm according to the features of the picture maximizes the advantages of different algorithms to meet the quantity of matching points and the matching accuracy.
AB - This paper quantitatively analyzes different types of image changes according to the characteristics of each algorithm, and put forward different optimal algorithms for different types of pictures. Firstly, four classical matching algorithms are selected and compared for scale, photometric and rotational robustness. In order to solve the limitation of the robustness of single algorithm, three improved algorithms are proposed. Based on the combination of SURF and ORB algorithms and one or more feature point screening, the improved algorithm is used to improve accuracy. Secondly, the improved algorithm is tested by using images with multiple types of changes at the same time. It is concluded that the improved algorithm has strong robustness and can effectively improve image matching accuracy. Finally, the simulation result shows that the selection of the optimal algorithm according to the features of the picture maximizes the advantages of different algorithms to meet the quantity of matching points and the matching accuracy.
KW - Feature points matching
KW - ORB
KW - Robustness
KW - SURF
UR - https://www.scopus.com/pages/publications/85058664456
U2 - 10.1145/3271553.3271585
DO - 10.1145/3271553.3271585
M3 - 会议稿件
AN - SCOPUS:85058664456
T3 - ACM International Conference Proceeding Series
BT - Proceedings of the 2nd International Conference on Vision, Image and Signal Processing, ICVISP 2018
PB - Association for Computing Machinery
T2 - 2nd International Conference on Vision, Image and Signal Processing, ICVISP 2018
Y2 - 27 August 2018 through 29 August 2018
ER -