.png){kind=icon}
.png){kind=icon}
Return
Automatic Oriented-Region feature point matching algorithm integrating Multi-Contrast levels
.png){kind=icon}
PRE
AI
DOI：10.1016/j.optlastec.2026.115211.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Image feature matching mainly establishes the correspondence between different images by comparing the similarity of key features in the images. It is an important technical basis in fields such as image stitching, object recognition and tracking, and 3D reconstruction. To improve the number and accuracy of matching points, this paper proposes an automatic oriented-region feature point matching algorithm integrating multi-contrast levels based on the SURF algorithm. In feature extraction, the left and right images are firstly processed by adjusting different contrast coefficients , and the SURF algorithm is used to extract feature points from images with different contrast levels. Then, the feature points of the left and right images under different contrast levels are integrated into the left and right point sets respectively. During the matching process, the corresponding matching region in the right image is directionally determined based on the positional relationship between the point to be matched in the left image and the reference point. Within this region, matching is accomplished using Euclidean distance, and mismatches are eliminated by employing the grid-based motion statistics (GMS) method. Experimental results show that the proposed method achieved an average matching accuracy of 76.0% (a 21.5% improvement over the traditional algorithm), with correct match counts increasing by 6- and 12-fold (approximately one order of magnitude). Furthermore, it exhibits good robust and stable performance under varying illumination and blur conditions.
Keywords:
feature matching
SURF algorithm
multi-contrast levels
oriented-region matching
GMS method
Journal
O
IF:
0
Papers:
880
Citations:
0